DDOG · Full Research · 2026.10.21
AI 的用量涨了，Datadog 收到了多少
照着 TEAM 的研究路径走一遍 Datadog：公司基础、thesis、逐条证据、产品、反向思考、框架打分、估值。原文在前，分析在后。
95份材料
12场管理层发言
868条逐字原文
$150–170买点 (现价 $258)
总体判断
这份是什么：不是 primer，也不是 initial review。它照着我们做 TEAM 的整条路走一遍 DDOG：先把公司讲清楚，再写出市场在定价的 thesis 和我自己的假设，用 12 场管理层发言和 Knowledge Base 里的专家、卖方材料逐条验证，再去理解产品实际怎么用、怎么收钱，最后反过来想怎么会错，落到估值和买点。每条引文都用脚本对原文逐字核对过，补成完整段落，带文件名和行号。
Executive Summary
一句话：Datadog 是好公司，这轮加速也是真的，而且比 TEAM 实在得多；但加速的钱主要来自"传统企业更多应用上云、更多产品整合到 Datadog"，不是来自"agent 让遥测数据暴涨"。数据量确实在暴涨，可单价同时在被大客户、竞品和 Datadog 自己往下压。现价 $258 已经按"不扣 SBC 的自由现金流 5 年复合增长约 29%"定价，翻倍需要的条件看不到。不买，放观察名单，买点 $150–170。
加速是真的。剔除最大客户后，其余业务已经连续 5 个季度加速；非 AI 客户的收入增速从一年前的 18% 到了 high-20s，占业务约 85%。渠道商看到的数字基本对得上 (Scicom Q1 +31%、Q2 +33%；一家 SI Q1 +33%、Q2 约 +35%)，只有一位渠道专家看到的是 mid-to-high teens。驱动主要是上云 (CFO 说传统企业只有 25–30% 的负载上了云并现代化)、多产品 (用 8 个以上产品的客户从 13% 升到 20%) 和销售扩张。
"量涨价跌"已经在发生，而且管理层亲口承认了机制。CEO 在 2025 年 11 月说得很直白：客户用量涨，续约时拿更大承诺换更低单价，所以"收入先掉，之后再靠用量涨"。最大客户 (市场普遍认为是 OpenAI) 正是这样：2025 年 Q3 提前续约换更好价格，2026 年 Q2 再续九位数合同，但 Q3 起减量，把不需要实时的指标存储挪了出去；Palo Alto 随后披露，一家大模型客户从"现有供应商"迁到 Chronosphere，带来九位数 ARR。与此同时，Datadog 自己推出的 Infinite Cardinality (多发数据"不额外收费")、Federated Logs (日志可以存在 ClickHouse/Databricks)、BYOC (日志存在客户自己的云) 都在主动给单位数据降价，毛利率从 81.7% 降到 79.6%。
AI 新产品还没进收入。MCP 工具调用比 2025 年 Q4 涨了 22 倍，但 CEO 自己说 agent 带来的 trace "是告诉我们世界往哪走，还不是今天业务的驱动"；Bits AI 的收入"你们在数字里还看不到"，正改成按 AI credits 收费 (约 $1/credit，一次自动调查约 6.5 credits)。渠道端 Bits 付费率不到 10%，一家大客户算过，人工处理一张工单的全成本约 $50，低于 Bits 的成本，ROI 算不过来。
估值已经走在前面。市值约 $972 亿，EV/2026 收入约 21 倍。过去 12 个月 FCF $10.7 亿，SBC $8.2 亿，扣掉 SBC 只剩 $2.5 亿，约等于收入的 6%、市值的 0.3%。把 SBC 当成本 (你的默认口径)，基本情形 (5 年收入复合 20%) 只值约 $114；按每股 FCF、把 SBC 当稀释看，基本情形约 $182，乐观情形约 $317。现价连后一种口径的基本情形都高出 40%。
逐条结论
| Thesis 环节 | 结论 | 证据强度 | 一句话 |
|---|---|---|---|
| E01 非 AI 加速是结构性的 | 大部分成立 | 强 | 公司数字连续 5 季加速，渠道多数对得上；但 CEO 一年前自己说"上云没有大幅加速，只是不再逆风"，有周期成分 |
| E02 AI 原生客户分散、留得住 | 一半成立 | 中 | 最大客户之外很分散 (31 家 >$1M、8 家 >$10M)；但最大客户恰好走了"压价 + 迁出"的路，超大厂客户留下的理由是"赶时间" |
| E03 AI/agent 让计费单位变多 | 方向成立，还没进收入主体 | 中 | 应用数 +50%、遥测 4–5 倍都有实例；但预算锚定云支出的约 10%，不是锚定数据量 |
| E04 单价通缩 | 存在，公司主动推动 | 强 | 管理层承认续约降价、主动帮客户少发数据；Grafana 便宜约 20%；毛利率 -2pt |
| E05 新 AI 产品能收钱 | 还没有 | 中 | Bits 收入不可见、改定价；渠道付费率 <10%；安全 ARR 刚过 $1 亿 (约收入 2–3%) |
| E06 预期 | 高 | 强 | 卖方目标价 $260–340，买方要 30%+；Q3 指引 +28–29%，全年指引隐含 Q4 只有 +25% |
| E07 股东经济学 | 差 | 强 | SBC 约收入 20%，不回购，稀释股数两年 +5%；同股不同权，内部人 7% 股份控制 70% 投票权 |
对照你的两条标准
- 好公司：是。用量计费、GRR 在 mid-to-high 90s、竞品 (Elastic、Snowflake) 自己承认很少从 Datadog 手里抢到客户，管理层执行力强，指引一向保守。比 TEAM 好的地方在于：AI 没有威胁它的计费单位 (按主机、数据量、span 收费，不按人头)。
- 两年一倍：不满足。翻倍是 $517，对应市值约 $1,940 亿，即使 2027–28 年每年增长 25–30%，也需要 2028 年 26–28 倍 EV/Sales。我们 mapping 里的赢家 (CRWD、NET、SNOW) 现在也只有 17–35 倍。
- 按你的规则：放观察名单等价格。买点 $150–170 (按每股 FCF 口径的基本情形，对应 12–15% 年化回报)，和 primer 给的区间一致。
Follow-up
| 时间 | 盯什么 | 什么结果说明什么 |
|---|---|---|
| 2026-11-04 Q3 业绩 (盘前) | 收入 vs 指引 $1.135–1.145B；非 AI 增速；最大客户减量幅度 | 收入 >$1.18B (同比 +33%) 说明剔除最大客户的业务还在加速，减量被消化；落在 $1.16–1.17B 说明只是常规 beat；非 AI 增速从 high-20s 掉下来，就是 Bernstein 说的"走平"开始 |
| 同上 | Q4 指引和 FY26 上调幅度 | 全年指引现在隐含 Q4 约 $1.19B (+25%)。Q4 指引如果在 +27% 以上，说明最大客户的拖累比 Jefferies 担心的小 |
| 同上 | 毛利率、RPO | 毛利率再降到 79% 以下，说明折扣和自我降价在压单价；RPO 连续第二个季度环比不涨，说明大单续约在换更低价 |
| 2027-02 Q4 业绩 | FY27 首次指引；AI credits 第一次单独披露 | 首次指引一般比实际低 5–8 个点。指引 ≥20% 且 Bits/credits 给出数字，才说明 AI 产品开始变现 |
| 持续 | Chronosphere/ClickHouse 有没有拿下第二家大 AI 客户 | Palo Alto 9 月已提到一家约 $2,000 万 ACV 的 AI 推理客户。如果 Anthropic 等前十大 AI 客户里再有一家迁出，E02 就不成立 |
| 可以约的专家 | 年账单 $500 万以上的客户的平台工程负责人 | 最缺的一手数据：2026 年数据量涨了几倍、Datadog 账单涨了几成、续约单价降了多少。这三个数能直接回答"量涨 vs 价跌" |
证据限制：管理层材料占证据的大头，关键数字 (最大客户减量多少、单价降多少) 管理层都拒绝给；专家里真正有 Datadog 内部视角的只有两位前员工，其中印度前员工 ("Jason") 的说法偏乐观且可能涉及内部信息；Funda 和 TMTB 是二手转述，看不到底稿；Q1、Q2 2025 两场业绩会原文没取到。
And then as you know, in most of our large customers, we have variability relating to the commit, so take that into consideration. Yeah. The last thing I will say because I know so it's on people's minds, is if you back out our largest customer from our growth, you get pretty much the same growth rate, as the rest of the business has been accelerating very steadily. Actually we've seen, I think now five quarters of continuous acceleration from the rest of the business. And we feel very good about the – what we see in the market. Yeah. No, I appreciate the thought.
Yes. Yes. I mean, look, this is -- we're not getting to the detail of any specific customer that this is the motion is always the same. Like customers grow they commit to more, they get better prices. So you see again, talking about customers in general, you see growth of usage, drops in revenue as customers renew and get higher commit and a better price and then unit growth after that for those customers. That's the motion that we've had. We have about 30,000 customers so far.
So, there's this whole new set of problems that need to be handled by observability. So, we see all that. This – today, the part of the business growing the fastest is the part that has to do with the build-out at the infrastructure and the application layer. We also see an explosion of traffic in all of the AI surfaces like – so the agents and the number of traces we get from that and things like that. But I think it's more of a – it's showing us where the world is going, rather than already today being the driving part of the business. Bits AI has been a fairly momentous set of product introductions. That portfolio of capabilities has expanded very swiftly. I know you've had new product leadership introduced into the organization.
So, that's been the weight of it. But it isn't the case all the time, and it also isn't 100% or 0%. Many customers are doing a combination. And what we said all along, which is that from our largest customer, is that we have commitment contracts, and that customer was growing rapidly and spending above the commitment. But we know that in certain customers, there'll be sort of a tradeoff between. And what we said in our last earnings call was that, that customer renewed with us, staying with Datadog, extending the contract. And the usage there has moved around relative to what they're doing. Sometimes it makes sense to hook up a database that doesn't have to be or a metric store, doesn't have to be real-time with Datadog and put the other work that has to be real time. So, we had some volatility. And what we decided to do, given the investor focus on that, is to de-risk our guidance by putting in our guidance only the commitment can't go below the commitment. But we've always said that's a fringe case. For the most part, our business is not about outsourcing the hyperscalers' observability. We're not a concentrated company.
公司基础
它是做什么的
一句话：Datadog 卖的是"线上系统的体检仪 + 行车记录仪"。一家公司的网站、App、后台服务跑在几百上千台云服务器和容器上，任何一处慢了、坏了，工程师要第一时间知道哪里坏、为什么坏。Datadog 在每台机器上装一个小程序 (agent)，把三类数据持续收上来：
- 指标 (metrics)：CPU、内存、请求数、错误率这类随时间变化的数字。对应产品"基础设施监控"，按主机收费，是最早的产品，也是大多数客户的入口。
- 调用链 (traces)：一次用户请求经过了哪些服务、每一步花了多少毫秒。对应 APM (应用性能监控)，按主机 + span (调用链里的每一步) 收费。
- 日志 (logs)：程序运行时写下的文字记录，比如错误堆栈。按摄入 GB 和索引条数收费，日志业务 ARR 已超过 $10 亿。
这三类数据放在同一个平台上、可以互相跳转，是 Datadog 相对 Splunk (日志起家)、New Relic (APM 起家)、Grafana (开源看板) 最大的优势。在这个基础上，它一层层往外卖：用户端体验监控 (RUM)、合成测试、数据库监控、云成本、安全 (Cloud SIEM，复用同一份日志)、值班和事故管理 (对标 PagerDuty)，现在一共 26 个产品，5 个 ARR 超过 $1 亿，另有 3 个在 $5,000 万到 $1 亿之间；总 ARR 超过 $40 亿。
谁在用：约 33,400 家客户，4,720 家年付超过 $10 万；财富 500 强里约一半是客户，但每家平均年合同只有约 $40 万，中位数不到 $50 万，所以管理层一直说渗透还很浅。约 85% 的业务来自非 AI 公司 (银行、零售、SaaS 等)，约 15% 来自 AI 原生公司 (750 家，包括所有前十大 AI 公司)。
怎么赚钱：几乎全部按用量收费。大客户签 1–3 年的承诺合同 (take-or-pay，用不满也要付)，承诺额可以在产品之间挪用；用得越多，承诺越大，单价越低。这个设计让 Datadog 不受"agent 替代人，席位变少"的冲击 (TEAM 的核心风险)，但也让它的收入直接跟着客户的数据量和单价走。
怎么收钱：计费单位表
2026-10-21 从 datadoghq.com/pricing 页面抓取的标价 (年度承诺价)。最后一列是我对"AI 用量增长时这个计费单位会不会跟着涨"的判断。大客户实际成交价远低于标价。
| 产品 | 计费单位 | 标价 | 随 AI 增长? | 为什么 |
|---|---|---|---|---|
| Infrastructure Monitoring Pro / Enterprise | 每主机每月 | Pro $15 (按年) / $18 (按需)；Enterprise $23 / $27 | 部分 | 按主机数收费。AI 推理/训练集群、更多微服务会增加主机，但 agent 调用次数本身不增加主机数。 |
| 额外容器 | 每容器每小时 (超出每主机配额部分) | $0.002/容器小时 按需；预付 $1/容器/月 | 部分 | AI 工作负载多跑在 K8s 上，容器数量随部署规模增长。 |
| APM (含 Infra) | 每 APM 主机每月 + 摄入 span (GB) + 索引 span (百万条) | APM $31 / Pro $35 / Enterprise $40 每主机(按年；按需 $36/$42/$48)；每主机含 150GB 摄入 span，超出 $0.10/GB；索引 span $1.70/百万 (15 天，按需 $2.55) | 部分 | 主机费不随 AI 变化。agent 让每个请求的调用链更长、请求更多，摄入和索引的 span 随之增加。 |
| Log Management - Ingest | 每 GB 摄入或扫描 | $0.10/GB (按年与按需相同) | 是 | Agent 和 LLM 应用产生大量日志(prompt、工具调用记录)，日志量随调用量增长。不过客户会用 Pipelines 预先过滤，或分流到 S3/BYOC。 |
| Log Management - Standard Indexing (15 天) | 每百万条日志事件 | $1.70/百万 (按需 $2.55) | 部分 | 索引是客户自己选的，Zendesk、Viber 都把大部分日志排除在索引外。AI 带来的日志多数可能进入 Flex 或归档层。 |
| Flex Logs | 每百万条事件存储/月 + 计算实例小时 | Flex Storage $0.05/百万 (按需 $0.075)；Flex Logs Starter $0.60/百万 (按需 $0.90)；Flex Compute 需询价 | 是 | 按存储的事件条数收费，量大价低。是 AI 时代日志量膨胀时 Datadog 的承接层，单价低于标准索引。 |
| Log Forwarding to Custom Destinations | 每 GB 每目的地 | $0.25/GB | 部分 | 随日志量增长，只在客户把日志转发到外部 SIEM/数据湖时收费。 |
| RUM Measure / RUM Investigate / Session Replay | 每 1,000 个用户会话 | Measure $0.15/千会话 (全量；按需 $0.22)；Investigate $3/千会话 (筛选保留；按需 $4.50)；Session Replay $2.50/千 (按需 $3.60) | 否 | 按真实人类用户的浏览器或 App 会话计费，与 AI 用量无直接关系。如果 agent 替人操作网页，会话可能反而减少。 |
| Agent Observability (原 LLM Observability) | 每 LLM span (一次模型调用) | Free：每月 4 万 LLM span；Pro $160/月 含 10 万 LLM span (按年；M2M $200，按需 $240)，超出 $3.50/万 LLM span；延长保留 30/60/90 天 $6.50/$8/$9 每万 span | 是 | 只按调用模型的 span 计费，工具和检索 span 不收费。agent 调用模型越多，账单越高，是最直接的 AI 用量计费项。 |
| AI Credits (Bits Chat / Bits Investigation / Bits Code / Agent Builder 等) | AI credit (按用量) | $500 每月 500 credits (按年或按月承诺，即约 $1/credit)；按需 $1.30/credit；参考消耗：Bits Investigation 一次自主调查约 6.5 credits(推算约 $6.5-8.5/次)，Bits Code 修复约 5，Agent Builder 每次运行约 1，Chat 每条消息约 0.5，DB 优化约 0.3；Bits Testing 11/30 前免费 | 是 | 纯用量计费：AI 调查、修复、对话做得越多，消耗的 credits 越多。每月重置，不结转，超出部分自动按按需价计费。 |
| Cloud SIEM | 每 GB 分析数据/月 (12 个月保留) | $5/GB (按年)；按需 $7.50；每分析 1TB 附送 100 次 Bits Security Analyst 调查 (每月至少 50 次) | 部分 | 按送去分析的安全日志量计费。AI agent 的审计日志(例如 Claude Compliance API 集成)会增加数据量，AI 调查有免费配额。 |
| AI Guard | 每 100 万扫描 token | $0.80/百万输入 token (按年)；按需 $1.20；限量可用 (Limited Availability) | 是 | 对 agent 的每次对话、工具调用、MCP 调用做内联检查，按 token 计费，和 agent 流量线性相关。 |
| GPU Monitoring | 每个活跃 GPU 设备每月 | $20/设备 (按年)；M2M $24；按需 $29；只对 5 分钟窗口内有活动的 GPU 计费 | 是 | 随客户 GPU 集群规模增长，闲置 GPU 不收费。新实验室(neolab)和 AI 原生客户是主要使用者。 |
| Datadog MCP Server | 无单独计费单位 (定价页只列在 AI 类目下，没有价格) | 定价页未标价；文档列出合理使用限制：每月 10 万次工具调用、每 10 秒 50 次突发 | 部分 | MCP 本身不直接收费(根据定价页未标价和文档推断，非官方明确表述)。但 agent 通过 MCP 查询的是付费数据(索引日志、span、指标)，用量越大，越推动客户保留和索引更多数据，或撞到限额后升级套餐。 |
| On-Call / Incident Management / Incident Response 套装 | 每席位每月 | On-Call $20 (按需 $29)；Incident Mgmt $30 (按需 $43.20)；套装 $40 (按需 $58) | 否 | 按人头计费。AI 替代一部分值班和调查工作后，席位数可能持平甚至减少，是少数可能被 AI 侵蚀的计费单位。 |
| Infinite Cardinality Metrics (自定义指标新计费选项) | 每个指标名 (随数据量)，不按基数 | DASH 2026 发布、已 GA；具体单价未在本次抓取中取得 | 是 | Datadog 称它专为 agent 查询设计；按指标名和数据量计费，不按标签组合数，降低高基数的惩罚。 |
读这张表的要点：真正和 AI 用量线性挂钩的是新产品 (Agent Observability 按模型调用收费、AI Guard 按 token、GPU 按卡、AI credits 按次)，它们现在加起来只占收入的低个位数。占收入大头的老产品 (主机、日志、span) 只是"部分"随 AI 增长：agent 调用工具确实会让更多应用、更多容器跑起来，但调用次数本身不增加主机数；日志和 span 会随调用量涨，可客户能过滤、采样、分层存储，Datadog 自己也提供了这些工具。少数会被 AI 侵蚀的是按人头收费的 On-Call 和事故管理。
历史、股价和季度数字
历史：2010 年由 Olivier Pomel (现 CEO) 和 Alexis Lê-Quôc (现 CTO) 在纽约创办，最早只做云上的基础设施监控，赶上了 AWS 普及；2019 年上市。2021 年疫情期间，云原生客户 (当时约占业务 40%) 大量上云，收入增长 70%+，股价冲到约 $190 (周收盘)。2022–23 年客户集体做云成本优化，增速掉到 25% 左右，股价最低 $64。2024–25 年增速稳定在 25–28%。
这一年发生了什么：2025 年 11 月到 2026 年 4 月，股价从 $191 跌到 $105 (日内最低 $98)，原因有三个：市场担心最大客户 (OpenAI) 减量或迁往 Chronosphere；2026 年 2 月给的全年指引只有 +18–20%；整个软件板块被"AI 取代 SaaS"杀估值。随后 Q1 收入 +32%、非 AI 加速到 mid-20s，全年指引一次上调到 +25–27%，股价一周涨了 42%；Q2 收入 +36%，但 Q3 指引 +28–29% 低于买方约 31% 的期待，股价跌了 13%，之后又涨到 10 月初的 $293。现在 $258，比 4 月低点涨了约 2.5 倍。
周收盘价，来源 IBKR，2021-10-25 至 2026-10-20 (最后一个点为 10 月 20 日收盘 $258.51)。52 周区间 $98.01–$300.68 是日内价。
季度数字
来自公司新闻稿和业绩会原文。"非 AI 客户增速"是管理层口径的"剔除 AI 原生客户后的收入增速"；AI 原生占收入在 2025 年 Q4 起不再披露。
| 季度 | 收入 | 同比 | 非 AI 客户增速 | 毛利率 | FCF (率) | SBC | NRR (TTM) | AI 原生占收入 |
|---|---|---|---|---|---|---|---|---|
| Q1 2024 | $611.3M | 27% | — | 83% | $186.7M (31%) | $135.0M | mid-110%s | — |
| Q2 2024 | $645.3M | 27% | — | 82% | $143.8M (22%) | $134.7M | mid-110%s | — |
| Q3 2024 | $690.0M | 26% | — | 81.1% | $203.6M (30%) | $142.2M | mid-110%s | ~6% |
| Q4 2024 | $737.7M | 25% | — | 81.7% | $241.0M (33%) | $158.5M | high-110%s | — |
| Q1 2025 | $761.6M | 25% | 19% | 80.3% | $244.4M (32%) | $164.3M | high-110%s | — |
| Q2 2025 | $826.8M | 28% | 18% | 80.9% | $165.4M (20%) | $180.5M | about 120% | 11% |
| Q3 2025 | $885.7M | 28% | 20% | 81.2% | $214.0M (24%) | $200.6M | about 120% | 12% |
| Q4 2025 | $953.2M | 29% | 23% | 81.4% | $291.0M (31%) | $205.4M | about 120% | 停止披露 |
| Q1 2026 | $1,006.4M | 32% | mid-20s% | 80.2% | $289.1M (29%) | $196.8M | low-120%s | 停止披露 |
| Q2 2026 | $1,121.5M | 36% | high-20s% | 79.6% | $278.7M (25%) | $220.3M | low-120s | 停止披露 |
这张表里最该注意的三件事：(1) 收入增速从 25% 到 36%，非 AI 从 18% 到 high-20s，加速是真的；(2) 毛利率从 81.7% 一路降到 79.6%，管理层说是"投入创新和效率相互抵消"，但同期正是大客户续约降价、低价产品放量的时候；(3) SBC 从每季 $1.35 亿涨到 $2.2 亿，占收入一直在 20% 左右，FCF 率从 31–33% 降到 25%。收入越加速，利润率越没有扩张，这和 CRWD、NET 在 2026 年的走法不一样。
竞争格局
竞争格局，按威胁从低到高：
- Splunk (思科)、New Relic (PE)、Dynatrace：老牌厂商。Splunk 和 New Relic 被收购后客户不满，是 Datadog 替换的主要来源。Dynatrace 在复杂的大企业 (尤其金融) 更强，首单动辄 $20 万以上，而 Datadog 以 $1–2 万的小单落地再扩张，两家正面竞争不多。
- Elastic：日志和搜索起家。Elastic 前战略伙伴总监说，在全栈可观测性上"很少从 Datadog 或 Dynatrace 手里赢单"，赢的基本都是 Splunk。
- Grafana Labs (开源 + 云)：OTel (开源的数据采集标准) 普及后，越来越常进入候选名单；Grafana Cloud 一般比 Datadog 便宜约 20%，自建更便宜。最常见的不是整体替换，而是"分层"：重要应用留在 Datadog，开发测试环境、低价值日志放到 Grafana/Prometheus。
- Chronosphere (2025 年被 Palo Alto 收购)：专做大规模指标存储，卖点是"帮你降可观测性成本"，在数据入库前先过滤、聚合。它从 Datadog 手里拿走了最大客户的指标业务，ARR 据 TMTB 转述从收购时约 $8,500 万涨到超过 $5 亿，其中一家大客户贡献很大 (另一份 Acecamp 纪要说 2025 年 9 月末已有 $1.6 亿，两者口径不一)。
- ClickHouse、Databricks、Snowflake (数据平台)：长期最大的威胁。观测数据越多，存储和计算越是成本大头，客户就越想把数据放在自己的湖仓里、只把需要实时看的部分送进 Datadog。OpenAI 的日志据称已迁到基于 ClickHouse 的自研方案。Datadog 的回应是 Federated Logs：让客户在 Datadog 界面里直接查 ClickHouse/Databricks 里的日志。
- AI SRE 创业公司 (Resolve、Traversal 等) 和第三方 agent：如果客户用 Claude Code 或自建 agent 通过 MCP 读 Datadog 的数据来排障，Datadog 的界面和 Bits 的价值会被摊薄。Datadog 首席产品官自己说这是一场"有意思的争夺"。
Thesis：市场在定价什么 vs 我的假设
市场现在在定价什么
把卖方报告、TMTB 和 Funda 的买方情绪、管理层的说法放在一起，市场隐含的 thesis 是：
"Datadog 是 AI 时代的用量赢家。它按主机、数据量、span 收费，AI 应用和 agent 让被监控的东西更多、更复杂，计费单位随 AI 变多。这次加速是结构性的，不只是 AI 客户，传统企业也在加速。最大客户的减量已经被消化。增速 25–30% 可以持续多年。"
这个 thesis 的证据看起来很强：Whale Rock 和 Maplelane 在 5 月的 Sohn AI 论坛上说 DDOG 是"一切的明确赢家"；Truist 升级时的理由是"街上低估了消费增长的持久性"；Wolfe 的上行情景是 2025–28 年收入复合 32%，目标价 $340。
我的假设：量涨，谁吃到了
做 TEAM 的时候，我们的问题是"AI 让 seat 变多还是变少"。DDOG 不按 seat 收费，这个问题不成立。DDOG 对应的问题是：AI 让遥测数据暴涨，这些量有多少变成了 Datadog 的收入？
拆开看，Datadog 的收入 = 客户送进来的量 × 单价。AI 让第一项的"潜在量"暴涨 (一位埃森哲前专家说，传统应用每天约 100GB 遥测，AI 应用每天 4–5TB)。但有三股力量在压另外两项：
- 客户不把量全送进来：采样、过滤、分层存储，把高量低价值数据放到 S3、ClickHouse 或 Grafana。
- 大客户压单价：承诺越大单价越低，大 AI 客户有能力自建，谈判筹码更大。
- Datadog 自己降价：Flex Logs、Infinite Cardinality、BYOC、Federated Logs，这些产品都是为了让客户"多发数据不多花钱"，否则客户会走。
所以我的假设是：只有当"Datadog 收进来的量 × 单价"的增长快于通缩，而且增长来自广泛的客户 (不是几家 AI 实验室)，thesis 才成立。如果增长主要来自"更多应用上云"这种老逻辑，那 DDOG 是一家很好的 25% 增长公司，但不是市场在定价的"AI 用量复利机"。
把 thesis 拆成可以验证的 7 条预期
| 编号 | 预期 | 要找什么证据 |
|---|---|---|
| E01 | 非 AI 客户加速是结构性的 | 非 AI 增速数字、原因 (上云/多产品/企业自己的 AI 应用)、NRR/GRR、新客户贡献、云优化是否结束 |
| E02 | AI 原生客户分散、留得住 | AI 客户数和分层；最大客户合同和减量；其他 AI 大客户会不会走"压价 + 自建/换供应商"的路 |
| E03 | AI/agent 让计费单位变多 | MCP 调用、trace、日志、GPU 的具体倍数和机制；反例 (agent 少查、客户不增加采购) |
| E04 | 单价没有被通缩吃掉 | 续约降价、折扣、OTel 降低迁移成本、竞品价差、Datadog 自己的低价产品、毛利率 |
| E05 | 新 AI 产品能收钱 | Bits AI SRE、LLM/Agent Observability、GPU、安全的收入；或者反方：agent 让界面不值钱，Datadog 退化成"存数据的地方" |
| E06 | 预期和仓位 | 指引方法和保守程度、卖方/买方预期、股价反应 |
| E07 | 股东经济学 | SBC、稀释、回购、FCF、利润率、毛利率 |
反常识 Check List
按对 thesis 的冲击大小排序。每条先说人话，再放原文。
1. CEO 自己讲了"量涨价跌"的机制 → 冲击 E04
2025 年 11 月，分析师问最大 AI 客户提前续约的条款，CEO 不谈具体客户，但把普遍规律讲了出来：客户用得多 → 承诺更大 → 拿到更好的价格 → 续约当时收入下降，之后再靠用量涨回来。这句话说明单价下行不是外部推测，而是 Datadog 商业模式的一部分。关键在于"之后用量涨回来"是否成立；最大客户这次没有涨回来。
Yes. Yes. I mean, look, this is -- we're not getting to the detail of any specific customer that this is the motion is always the same. Like customers grow they commit to more, they get better prices. So you see again, talking about customers in general, you see growth of usage, drops in revenue as customers renew and get higher commit and a better price and then unit growth after that for those customers. That's the motion that we've had. We have about 30,000 customers so far.
Finally, we signed a 9-figure annualized expansion with a leading AI company. This company has been a long-time Datadog customer and has expanded their usage of multiple products, securing better economics for a higher commitment with an early renewal. Speaking of AI customers, we continue to help native customers big and small to grow and scale their businesses. And we continue to see this group broaden in number and size, with more than 500 AI- native companies in this group, about 100 of which are spending more than $100,000 annually with Datadog and more than 15 who are spending more than $1 million annually with us. While we know there's a lot of attention on this cohort, we primarily see it as an indication of what's to come as companies of every size and every single industry incorporate AI into their cloud applications. And that's it for another very strong quarter from our go-to-market teams, who ran our very hard at work as we have a really exciting pipeline for Q4. Before I turn it over to David for a financial review, I want to say a few words on our longer-term market.
2. 最大客户把不需要实时的数据挪出去了，竞品确认拿到了九位数 → 冲击 E02
2026 年 Q2 业绩会上，Morgan Stanley 的 Sanjit Singh 直接问：减量是因为新合同单价更低，还是流失/降级？管理层没回答。一个月后 CFO 在高盛会议上说，这个客户把"不需要实时的数据库或指标存储"从 Datadog 挪了出去。9 月 Palo Alto 业绩会说，Q4 净新增 ARR 里有一笔来自"一家大型 LLM 客户从现有供应商迁到 Chronosphere"的九位数收益。Datadog 当天跌 3%。
David, thank you for giving us the color on some of the guidance assumptions, particularly headed into Q3, with respect to the largest customer. I was wondering if you could share any additional details in terms of the new contract, was it of similar duration? And in terms of the lower usage, is that a function of the customer getting lower unit price because of making a new commitment? Or was there some churn or down sell that we need to think through not only for Q3, but for the balance of the year? Yeah. So maybe I'll take this one. I think we – so overall, we as usual, we don't want to comment too much on any specific customer because we also don't really control what's happening with any specific customer. We wanted to be transparent about this on the call, because we did see a reduction in usage and we took the liberty to fully de-risk the guidance for the rest of the year with respect to that customer. And again, the reason for that is we don't control what's happening to a specific customer, but we do have a great amount of control on what's happening to everything else in the business. And the business is booming. And we don't want that to overshadow basically the acceleration we see pretty much everywhere else in the business. So as we mentioned on the call, we renewed the customer, it's a longtime customer who uses many of our products. But there's not a lot more we can share. Yeah. I think... Okay. ...just to get specific on the guidance. We last quarter and previous quarters said that we essentially have a level of commit and we can de-risk our guidance by using that. And then as you know, in most of our large customers, we have variability relating to the commit, so take that into consideration. Yeah. The last thing I will say because I know so it's on people's minds, is if you back out our largest customer from our growth, you get pretty much the same growth rate, as the rest of the business has been accelerating very steadily. Actually we've seen, I think now five quarters of continuous acceleration from the rest of the business. And we feel very good about the – what we see in the market. Yeah.
On **DDOG -3%:** “*Keep in mind, as we noted last quarter, our Q4 net new ARR includes a nine-figure benefit from a large LLM customer migrating to Chronosphere from an incumbent vendor.*”
来源背景：TMT Breakout (TMTB) 日报，PANW 财报点评，引用 PANW 管理层原话 (CEO/CFO 未具名) 与 TMTB 自己的分析。关于 '大型 LLM 客户' 未点名 (市场普遍认为是 OpenAI)。二手转述，看不到底稿。
* **Identity.** CyberArk ACV growth accelerated to approximately 27% y/y from ~20% around the acquisition, shared leads exceeded 400, PANW’s installed base generated >200 new logos and $5M-plus TCV deals increased 50% y/y. Cost synergies remain three to six months ahead of plan. Modern PAM broadens the addressable privileged-user base, while Secure IAM and non-human identity controls create additional land-and-expand opportunities, although “true” independent agentic identity remains early. * **Observability.** Chronosphere ARR surpassed $500M and more than doubled following the acquisition, but the quarter included over $100M of NGS ARR benefit from one large LLM customer’s migration. The remaining migration benefit should be below $100M in F1Q before normalizing materially, creating a difficult quality-of-growth and concentration comparison after F1Q. Offsetting that concern, PANW cited a separate ~$20M ACV AI-inference customer, 50% XSIAM attachment among Chronosphere net-new logos, and a path to broader enterprise selling once feature parity is reached over the next six months. * **Margins.** F4Q non-GAAP gross margin was 74.8%, down ~100 bps y/y and below Street 76.1%, reflecting faster-growing cloud/SASE revenue, higher hosting expense, acquired-business mix and memory/storage inflation. Non-GAAP operating margin nevertheless reached 29.6% vs.
来源背景：TMT Breakout (TMTB) 日报，PANW 财报点评，引用 PANW 管理层原话 (CEO/CFO 未具名) 与 TMTB 自己的分析。关于 '大型 LLM 客户' 未点名 (市场普遍认为是 OpenAI)。二手转述，看不到底稿。
3. Datadog 主动让客户"多发数据不多花钱" → 冲击 E03/E04
Infinite Cardinality 的官方说法是让客户在 AI agent 产生更多数据时"不产生任何额外费用"。CFO 说他们会主动去找客户："你设错了，你往 Datadog 里送了太多用不上的数据"。Federated Logs 让日志不必进 Datadog。这些对客户是好事、对长期留存是好事，但每一项都在切断"AI 产生更多数据 → Datadog 多收钱"这条链。
And with Bring Your Own Cloud or BYOC, customers can now use the full Datadog experience on logs that are kept within their infrastructure. And we've also announced that we are bringing BYOC to metrics and traces as well. In the digital experience space, joining monitoring automatically gives a single shared view for every critical user flow. And for custom metrics data, we introduced Infinite Cardinality Metrics, which allow our users to answer arbitrarily complex questions as they generate larger amounts of data with AI agents without incurring any extra costs. Finally, we launched a number of innovations to secure the AI stack and defend against a new class of AI powered attacks. AI Guard agent discovery finds and maps every known and unknown custom agent to security teams can see what is protected and what is not. AI Guard for custom agents provides runtime protections to block attacks that can only be detected with real time observability data.
Yeah. We've analyzed that. We actually thought, okay, for – we want the – a map – the logs that need to be accessed there to be in Datadog, but we want the platform to be able to use as a central point. The more we have our users centralized on Datadog, we proven that we get higher average revenue per customer. So, to close Datadog off and force them to go to another platform is not as optimal as staying in Datadog all day long and using the data efficiently. We actually – getting back to how we've evolved, we actually go to clients and we say, you set this wrong. You're flowing too much data into Datadog that you don't need for this purpose. And it creates a much better long-term value of customer, and also, in many cases, is margin enhancing for us because to store all that data in Datadog that is never used also is a burden on our system. Every two to three quarters, we all have to ask you about some new competitor in the space... Sure.
I'll talk about Federated Logs. It's in a group of innovations where we used to have the – all the data have to flow into Datadog to be used, but we know that's not efficient. So, Federated Logs, Data Observability Pipelines, Bring Your Own Cloud. What this means in the case of Federated Logs is that we know that not all logs have to flow into Datadog to enhance the platform, that we can link into and logs can be stored, in this case, it was ClickHouse and Databricks. So, what we're doing is, we'll say (00:20:00) we understand certain – what we want to be able to have more data available in the platform without having to port all of your data into Datadog. That, we've proven, and we've gotten feedback from clients, enhances the value of the platform, eliminating one of the objections is, I have to bring all my data into Datadog even though I'm storing it elsewhere. And this is a better way to use Datadog, to have what flows in to be the most important, but to leave what isn't – doesn't need to flow in there.
4. CEO：agent 带来的流量还不是今天业务的驱动 → 冲击 E03/E05
市场 thesis 最核心的一环是"agent 让计费单位变多"。CEO 9 月在 Citi 会议上说，增长最快的是基础设施和应用层的建设，agent 的 trace 爆发"是告诉我们世界往哪走，而不是今天已经在驱动业务"。5 月他还说，加速"从数学上"主要来自现有产品。
And if we look historically at the S-curves that the business has gone through, maybe you started with metrics actually really kind of collaboration and then metrics, and then you know APM, and then logs, and these have each kind of grown to being kind of billion dollar-type opportunities. There's several things that you've been putting out that you, I think, have been very hopeful that could become some of those next S-curves, whether or not it's security, whether or not it's AI monitoring, the GPUs, monitoring that you guys have been putting out. When you look at the acceleration that's going on right now, how much can you kind of think those new things for this versus kind of your existing footprint? What's interesting is most of the acceleration like mathematically, it comes from the existing stuff. And when you think of like we have the – I think we said like 14 of the top 20 AI companies using us, including the very largest of those, like, they use the same traditional products as the rest of our customer base, so by and large. So that gives you a sense of the shape of the market. When you look at the observability category, so there's this Gartner number that we put every year. But we're number one in observability according to Gartner in – at least in the closest – the category they have that's closest to observability for that exact name of the category.
They were so impressed. They emailed us straight on that. So, it's pretty transformative for them. In terms of where it is on the results, you won't really see it yet on the results. We're actually in the process right now of repackaging these products. We used to have one SKU we sold for – specifically for investigations. And now, we're packaging it to a set of AI credits instead that can work across a lot of different surfaces, because as I said, we expanded the product to cover a lot of different surfaces. And so, you should – we probably will comment on that in the future. But, for now, there's a – the transformation's happening right now in terms of the way we're charging those. And by the way, if you zoom out, the world as a whole doesn't quite know how to charge for agents just yet.
5. 客户的观测预算锚定在云支出的约 10%，不是锚定数据量 → 冲击 E03
一位前 Datadog 工程经理 (现初创公司 CTO) 说，经验法则是观测支出约为云支出的 10%，现在 LLM 支出单列以后，他仍按 10:1 管理。一位独立顾问的说法更直接：年付 Datadog $2,000 万的客户，环境扩大 3 倍也不会花到 $6,000 万；他预计最后约 20% 的数据全量采集，其余采样。如果预算锚定云支出，Datadog 的增速长期会向客户云支出增速收敛 (目前 AWS 约 20% 出头)，而不是向数据量增速靠拢。
JC: Yeah, I think the common rule of thumb that I'm familiar with from the old days is that if your cloud provider spend is X, then your observability spend should be about 10% of that. I think we now obviously have this whole category of LLM spend and then observability spend within that. And if we break those out into their own separate line items, that 10 to 1 relationship is still similar to what I'm trying to maintain in our organization.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。
+ All vendor solutions combined cover only about 20% of production environments; the other 80% is homegrown and open-source tooling — what DDOG and DT are really competing for is that 80%, not each other + That 80% exists precisely because customers will not pay incumbent vendor pricing across their whole environment; his example: a client spending $20m a year on DDOG is not going to spend $60m when the environment triples in size + The answer is for vendors to use AI to distinguish which data matters and tier collection accordingly — he expects an 80/20 outcome where roughly 20% of data (transaction-level, agent-behavior-level) is collected in full, and the rest
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。
6. CEO 承认通用大模型 + MCP 也能做类似的排障 → 冲击 E05
被问到"agent 会不会取代 Datadog"，CEO 一方面说观测数据比大模型输入大 5–6 个数量级、要求近实时，通用模型做不好；另一方面承认，"眯着眼看"，通用模型通过 MCP 取数也能得出类似结论，Datadog 的差异要靠下一步把专用模型放进数据平面里。首席产品官 10 月也说，客户是通过轻量 MCP、Datadog 自己的 agent 还是第三方跨厂商 agent 来消费遥测，是一场"有意思的争夺"。
And the part of the job there is to make sure to get to the benign ones super quickly, so they can focus on other things. And so, we do that as well. So, that's what we do today. If you squint, you could say, hey, some of that you could also do from a (39:07) and you can then – or MCP tools to acquire data and maybe they can come to the similar reasoning and come to a similar conclusion. The next level of what we're building though is about having that intelligence directly inside the data plane, with all of our real time data and to drive the automation itself. So, models are specialized that can run on very large volumes of data at a very attractive cost and with very low latency. So, we can fully automate the operations there. There's some first outputs of that we can see.
On how much context to surface via MCP, Li says Datadog has had to curate formatting and query mechanics rather than ship a raw API wrapper, and notes an emerging “interesting battle” over whether customers consume telemetry via lighter MCP access, Datadog’s own curated AI agent, or third-party agents synthesizing data from multiple vendors. She argues dashboards aren’t disappearing — some customers still find clicking through known dashboards faster than natural-language queries — but consumption surfaces are proliferating (Slack, CLI, IDE), and AI agents increasingly generate dashboards just-in-time for a specific investigation rather than reading pre-built ones.
来源背景：TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。
7. CEO 预警：AI 支出的"合理化"可能比想象来得早 → 冲击 E06
We're not at all in the phase of, okay, so now what else can I build on that or how do I rationalize it or how do I optimize? That's for later. That later might come sooner. If we see the explosion of the AI company's revenue continue at the rate it is right now, that money is coming from somewhere and there's probably going to be a push for rationalization sooner rather than later. But the mode right now is still very much. Let me get it to work. Yeah.
Does that affect also your margins? Can you keep up that 30% growth? Yeah, I think the – so on our end, it's a big opportunity because everybody right now across the whole customer base, from the startups to the AI natives and with all of the big enterprises like the big banks and train companies and whatnot in the middle, everybody is busy adopting AI. And everybody is seeing their costs skyrocket and everybody is also finding that they are not in a position where they can really understand what's actually needed. Are they spending the right amount to get the right outcome? What are the outcomes, by the way? Like it's super hard. When you zoom in into specific parts of the business, you can see the outcomes, you can see value, you can see things. When you try to add everything up across the whole business, it's a lot more difficult, and I expect that to remain difficult for maybe a year or two, basically. But for us, that creates a huge opportunity because we can be there to help understand who's doing what with AI, what outcomes they're reaching, where they – how much they're spending to do that and how they can get better at it. One analogy I would give is right now, we just invented air travel and we're sending postcards by business, building – buying business class tickets for them.
8. 渠道和客户层面的反例
一位北美渠道商看到的客户消费增速是 Q1 mid-to-high teens、Q2 high teens 到 low 20s，明显低于公司报告的数字 (他的客户偏传统企业)。一家用 Datadog 的公司运维负责人说，他们的指标、看板、日志量都在减少，迁到了 Grafana，费用自然不涨；他还说 AI 降低了维护开源方案的门槛。这两条都是单一视角，但说明"全面加速"在客户层面并不均匀。
对。 嗯。First question, what was the growth growth rate of customers' consumption on DataDog in this year Q1 and Q2, and what are the reasons behind the changes in growth rate? Yeah, um, so the consumption in Q1 was um, in general mid to high teens, I would say percentage year over year. um, I think the Q is, Q2 is showing um, medium acceleration, uh, probably in high, um, teens or you can call it low 20s. um, you know, customers, uh, reasons and the drivers from customers scaling AI and cloud native workload. Primary drivers were, um, more Kubernete adoption. Um, there was a higher trace volume, there was broader APM deployment.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。
问：公司在 Datadog 上的投入是否会显著增加？ 暂时没有增加。公司以前的核心指标等内容其实在减少，很多都迁移到了开源的 Grafana系统。相对来说，Datadog 上的指标数、dashboard数据、日志量等都减少了很多。Datadog 的计费方式是根据节点数、指标数和数据量来计费，只要这些减少，费用自然不会增加。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。
9. Q2 收入创纪录，RPO 却环比没涨 → 冲击 E06
Q2 环比收入增量 $1.15 亿创纪录，但 RPO (已签未确认的合同额) 从 $34.8 亿变成 $34.7 亿，同比增速从 51% 降到 43%，cRPO 从 mid-40s 降到约 40%。CFO 的解释是"收入比 RPO 更能代表趋势"。一种解释是最大客户续约时承诺额降了。
We believe this metric highlights the mission critical nature of our platform for our customers. Now, moving on to our financial results. Billings were $1.18 billion, up 38% year-over-year. Remaining performance obligations, or RPO, was $3.47 billion, up 43% year-over-year. Current RPO grew about 40% year-over-year and RPO duration increased year-over-year. As we previously mentioned, we continue to believe revenue is a better indication of our business trends than billing and RPO. Now let's review some of the key income statement results. Unless otherwise noted, all metrics are non-GAAP.
Evidence：逐条验证
按 7 条预期组织。每节先给判断，再列关键原文，最后可以展开该节全部原文。绿色边框是支持 thesis 的原文，红色是反对，灰色是口径不齐或中性。
E01 非 AI 加速是结构性的吗
判断：大部分成立。加速是广泛的、连续的，而且有渠道数字印证；但驱动是"上云 + 多产品整合 + 销售扩张"这组老逻辑，企业自己的 AI 应用只是小头。一年前 CEO 说的是"环境不再逆风"，这里面有云优化周期结束的成分。
数字：非 AI 客户收入增速 19% (2025 Q1) → 18% → 20% → 23% → mid-20s → high-20s (2026 Q2)，剔除最大客户后的业务连续 5 个季度加速。存量客户环比用量增长是 2022 年以来最强；新客户对同比增长的贡献从 Q1 的 25% 升到 Q2 的 30%，企业新客户 bookings 翻倍。NRR 从约 120% 回到 low-120s，但 Q2 没有继续上升；GRR 一直在 mid-to-high 90s。
管理层怎么解释：三个原因。第一，上云：CFO 多次说传统企业只有 25–30% 的负载上了云并现代化，70% 以上还没动；AI 让企业更急着现代化。第二，多产品：用 4 个以上产品的客户从 51% 升到 56%，8 个以上从 13% 升到 20%；客户把 PagerDuty、xMatters 这类工具合并进 Datadog。第三，销售产能：过去两年招了很多销售、开了新区域和新的大客户团队。
需要打折的地方：2025 年 11 月 CEO 第一次讲加速时的原话是"我不知道我们看到了上云的大幅加速，但至少环境不再往反方向推"，这是周期性的说法 (2022–24 年的云成本优化结束了)。CEO 自己也说，现在的增速"远低于 2021 年"，所以不会出现客户连续几年消化巨额涨幅的问题。
外部印证：多数渠道数字对得上：Scicom (F500 渠道) Q1 +31%、Q2 +33%，全年目标 35%，下半年冲 37–40%；一家北美 SI Q1 +33%、Q2 约 +35%；BTIG 最大的渠道联系人 Q1 +32%、Q2 约 +35%；SHI 前员工看到的客户今年支出 +25–28%。偏低的是一位北美渠道商 (Q1 mid-to-high teens) 和 Funda 的全球渠道 B (今年 +20%，和去年持平，"上限是预算而不是需求")。Bernstein 7 月降级时认为占收入 85% 的非 AI 需求会从 Q3/Q4 走平，9 月又根据 AWS 数据说 Q4 非 AI 云负载可能再弱约 100bp。
And, this quarter included new land deals with two of the world's biggest AI research teams, helping them improve, and optimize their training workflows. I'll talk more about that in a bit. Even more impressive was the growth in our non-AI customers. Non-AI customer revenue growth accelerated again this quarter to mid-20% year-over-year, up from 23% last quarter, and 19% in a year-ago quarter. We think this is a sign of strong continued cloud migration, greater adoption of our products, and customers of all kinds accelerating their use of AI. Finally, churn has remained low with gross revenue retention stable in the mid-to-high 90s highlighting the mission critical nature of our platform for our customers. Regarding our Q1 financial performance, and key metrics, revenue was $1.01 billion, an increase of 32% year-over-year, and above the high end of our guidance range. We ended Q1 with about 33,200 customers up from about 30,500 a year ago.
Yes. I said there's 3 points to it. One part is the demand environment is -- is positive in general. I don't know that we see massive acceleration of cloud migration, but at least the environment is not pushing the other way. We know which happens from time to time. So that's point number one. Point number 2 is we've been growing sales capacity quite a bit, and we've created new go-to-market motions to go after the kind of customers who were not getting before like -- we've done quite a bit of investment over the past couple of years, and we see that starting to pay off. As I said also, we feel good about the Q4 in terms of pipeline on the sales side. So it's too early to tell yet. We still have to close those deals, but we feel good about the scaling of our go-to- market. And point number 3 is we have a number of products that we've been developing over the years. Some of them are early, some of them a little bit further along, they are really clicking. We see -- we have a lot of success with getting large surprises to adult Flex Logs, for example, we have a lot of success, some of new products such as analytics that we mentioned on the call, we're seeing some large land deals or a lot team. So all of that is contributing to the picture you're seeing today.
It sounds like the enterprises are getting on board with their AI initiatives. And so just in terms of the enterprise AI, app dev cycle, what does that look like for Datadog over the last couple of quarters? Well, we do see broad adoption, and we see it in two ways. One is we see it manifest itself is in just more transformation, more cloud adoption, more workloads, more modernization from customers. And that's what drives the majority of the known AI customer acceleration, so we mentioned also we've seen continuous acceleration from customers that existed before AI and that are not majority AI businesses. And that's been pretty remarkable. Like the acceleration that we gave the numbers on the call, but the acceleration since last year has been constant and very significant. And it keeps happening as far as we can tell.
I mean, help me understand the big deal activity. Yeah, it's definitely a confluence of both. So you have – in a customer base that has been around for a long time, so therefore they're not cloud-natives, they didn't just get birthed, so they have legacy technologies and they've – there's – they have a long way to go. There's somewhat 25%, 30% of workloads in the cloud right now and modernized. So they are continuing to find use cases and modernizing. Datadog's platform is getting bigger and better so we're consolidating market share onto our platform, and we're getting better about delivering the enterprise service model, whether that be channel partners, customer service, technical help, the whole ecosystem to be able to deliver. So all of that has been what we've been investing behind for some time that is bearing fruits. And that's resulted in some large lands, some – we still are largely a land and expand, some very substantial expansions.
These customers generated about 90% of our ARR and we generated free cash flow of $289 million with a free cash flow margin of 29%. Turning to product adoption. Our platform strategy continues to resonate in the market. For example, 56% of our customers now use four or more products up from 51% a year ago. Thirty-five-percent of our customers use six or more products up from 28% a year ago, and 20% of our customers use eight or more products, up from 13% a year ago. So we're landing more customers, and delivering value across more products, and our business continues to grow. Our total ARR now exceeds $4 billion, and our quarterly revenue exceeded $1 billion for the first time in Q1. This is a big achievement for all of us at Datadog, and is a product of years of investment in building, and innovating for our customers, but we're still just getting started.
Meanwhile, our AI customers continue to grow rapidly and diversify in the quarter. This 750 strong customer group includes a broad range of AI startups as it has in the past, but now also includes hyperscalers using Datadog for in-house AI labs. In Q2, this includes 31 customers spending more than $1 million annually, of which eight customers spent more than $10 million annually. We also achieved strong new logo dollar bookings with particular strength in enterprise, where new logo annualized bookings more than doubled from the year ago, and we are seeing new logos ramping faster and contributing more to revenue growth. The portion of our year-over-year revenue growth that relates to new customers was about 30% in Q2, up from 25% in Q1. Geographically, we're performing well in all regions with growth acceleration across the regions. We see particular strength in the Americas as much of the AI activity is occurring in the US as well as – in addition, we are executing strongly in LatAm. Regarding retention metrics, our trailing 12 month net revenue retention percentage was in the low-120, similar to last quarter, and churn remains low with gross revenue retention in the mid to high-90s.
The portion of our year-over-year revenue growth that relates to new customers was about 30% in Q2, up from 25% in Q1. Geographically, we're performing well in all regions with growth acceleration across the regions. We see particular strength in the Americas as much of the AI activity is occurring in the US as well as – in addition, we are executing strongly in LatAm. Regarding retention metrics, our trailing 12 month net revenue retention percentage was in the low-120, similar to last quarter, and churn remains low with gross revenue retention in the mid to high-90s. We believe this metric highlights the mission critical nature of our platform for our customers. Now, moving on to our financial results. Billings were $1.18 billion, up 38% year-over-year.
And it's a largely driven by existing customers because that's the majority, like, when you think of what it takes to move that number, that's basically the majority of our business. Like we're not just going to move that with a few newer customers like this there, it's largely driven by the existing customers. And it's driven by both increases in volume because they're moving more workloads to the cloud and adoption of our newer products as they consolidate onto us. We think it's sustainable for one thing, if you compare it to what we have seen in the heady days of 2021 or the growth rates are accelerating, but they're still far below what we were seeing at that time. And so we don't create the same issue of customers having to digest very large increases multiple years in a row. I think in this case, we're very well within the realms of sustainability. And as has been a theme in this call, remember that when customers adopt and they consolidate, they have an eye towards the financial side of the equation basically, how much money are they going to make or save by doing that at the end? And we are very good at helping customers understand that and making that case and helping them save money at the end of the day. So we feel good about that. And I want to just add one thing, and we talked about this last quarter, that some of this has to do with the investments that we're making in our platform and our product, but it also has to do with the investments that we're making in our go-to-market. We've successfully expanded quota capacity, the geography of it and essentially, that's, as we talked about last quarter, providing returns. So that's also being a growth driver in our non-AI or enterprise type business.
Sid Roy: Right. So for Datadog, we have a 35% annual growth target. In the second quarter, we did 33%. In the first quarter, we did 31%. So you could almost say like our...
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。
我们想先问一下,就是专家这边的客户的话,在datadog上面的consumption在今年的一季度和今年二季度的整个增涨,的整个同比增长情况是什么样的? 然后增涨背后主要的一个驱动因素是什么? Yeah, so what is the growth rate of your customer spending on datadog in Q1 and Q2 this year and what are the reasons behind their changes in the growth rate? Yes, so uh, I mean Q2 is not completely finished yet, of course, but in in Q1 we saw a growth rate of uh uh slightly north of 30%, so uh 33%. Uh, and for Q2, we expect uh a similar growth, somewhat accelerating, but not not like a massive difference. So, uh, let's say 35% for um for Q2. Uh, and the reason I would say is uh mainly that existing customer go deeper, right? I mean, I think I always think about uh data dog as an as an ecosystem of different modules and products, right? And and they are uh there are synergies in between them and they are very good also at sort of uh following a land and expand uh strategy. So it's mostly uh sort of uh net retention and growth with existing customers driving it. Not exclusively, of course, but I think that's that's a growth growth story that is interesting. Mhmm. Okay.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。
Bernstein raised its long-term estimates for Datadog, citing increasing confidence in the company’s AI platform opportunity, but downgraded the stock to Market Perform on valuation and elevated near-term expectations. The firm expects a strong 2Q, but believes growth could disappoint beginning in Q3/Q4 as ex-AI demand (~85% of revenue) flattens and AI lab growth starts to normalize. Bernstein expects Q4 revenue growth to slow roughly 500bps to ~29% y/y, well below investor expectations for 30%+ growth into next year and peak growth in the high-30% to 40%+ range.
来源背景：TMT Breakout (TMTB) 日报，转述 Bernstein 降级报告摘要。二手转述，看不到底稿。
展开本节全部 131 条原文 (支持 102 / 反对 10 / 其他 19，按时间)
- 支持 Q3 非 AI 原生客户收入增速由 18% 加速到 20%，且 10 月延续
I will note that over time, we think this metric will become less relevant as AI usage in production broadens beyond this group of customers. Our year-over-year revenue growth also accelerated amongst our non-AI native customers. In Q3, our revenue growth, excluding AI native customer group, was 20% year-over-year, accelerating from 18% year- over-year in Q2, and we have seen this trend of accelerating growth continue in October. Regarding retention metrics. Our trailing 12-month net revenue retention percentage was 120% similar to last quarter and our trailing 12-month gross revenue retention percentage remain in the mid- to high 90s.
- 支持 非 AI 原生存量客户环比用量增长为 12 个季度最强
To start, our Q3 revenue was $886 million, up 28% year-over-year and up 7% quarter-over-quarter. To dive into some of the drivers of our Q3 revenue growth. First, overall, we saw sequential usage growth from existing customers in Q3 that was higher than our expectations and the strongest in 12 quarters in our non-AI native customer base. We saw year-over-year growth acceleration broadly across our business, including in new logos and existing customers, both enterprise and SMB with customers across our spending bands, big and small, and customers in a wide variety of industries. Next, we saw strong and accelerating contribution from new customers.
- 支持 新客户贡献上升：新 logo 年化 bookings 翻倍以上，新客户占收入增量约 25% (Q2 20%)
We saw year-over-year growth acceleration broadly across our business, including in new logos and existing customers, both enterprise and SMB with customers across our spending bands, big and small, and customers in a wide variety of industries. Next, we saw strong and accelerating contribution from new customers. New logo annualized bookings more than doubled year-over-year and set a new record driven by an increase in average new logo land size, particularly in enterprise. We believe we are starting to see the benefits of our growth of sales capacity. And we are seeing new logos ramping faster, contributing more to revenue growth. The portion of our year-over-year revenue growth that related to new customers was about 25% in Q3, up from 20% in Q2. Next, our AI native customers continue to exhibit rapid growth, while more customers in this group are growing to be sizable customers. As Olivier discussed, we extended the contract of our largest AI native customer.
- 口径不齐 CEO 承认没看到上云大幅加速，只是环境"不再逆风"——周期性因素 (优化结束) 与结构性因素混在一起
Yes. I said there's 3 points to it. One part is the demand environment is -- is positive in general. I don't know that we see massive acceleration of cloud migration, but at least the environment is not pushing the other way. We know which happens from time to time. So that's point number one. Point number 2 is we've been growing sales capacity quite a bit, and we've created new go-to-market motions to go after the kind of customers who were not getting before like -- we've done quite a bit of investment over the past couple of years, and we see that starting to pay off. As I said also, we feel good about the Q4 in terms of pipeline on the sales side. So it's too early to tell yet. We still have to close those deals, but we feel good about the scaling of our go-to- market. And point number 3 is we have a number of products that we've been developing over the years. Some of them are early, some of them a little bit further along, they are really clicking. We see -- we have a lot of success with getting large surprises to adult Flex Logs, for example, we have a lot of success, some of new products such as analytics that we mentioned on the call, we're seeing some large land deals or a lot team. So all of that is contributing to the picture you're seeing today.
- 支持 加速归因于销售产能扩张和新 go-to-market 投入开始回报 (结构性)
Yes. I said there's 3 points to it. One part is the demand environment is -- is positive in general. I don't know that we see massive acceleration of cloud migration, but at least the environment is not pushing the other way. We know which happens from time to time. So that's point number one. Point number 2 is we've been growing sales capacity quite a bit, and we've created new go-to-market motions to go after the kind of customers who were not getting before like -- we've done quite a bit of investment over the past couple of years, and we see that starting to pay off. As I said also, we feel good about the Q4 in terms of pipeline on the sales side. So it's too early to tell yet. We still have to close those deals, but we feel good about the scaling of our go-to- market. And point number 3 is we have a number of products that we've been developing over the years. Some of them are early, some of them a little bit further along, they are really clicking. We see -- we have a lot of success with getting large surprises to adult Flex Logs, for example, we have a lot of success, some of new products such as analytics that we mentioned on the call, we're seeing some large land deals or a lot team. So all of that is contributing to the picture you're seeing today.
- 支持 加速是广泛的：跨企业/SMB、跨消费档位，剔除 AI 公司后 SMB 也在走强
Yes, I want to -- I'll add that it's across the customer base, enterprise SMB. And when we look at it, it's not just an AI SMB. If you remove these AI companies, you still see a strengthening SMB demand cycle going on. And unlike in previous periods, it also is across spending ranges. We're not seeing larger spenders or smaller spenders. We're just seeing a broad trend of improved demand across the spending trends. .
- 中性 NRR 约 120% 与上季持平，GRR 维持 mid-to-high 90s
In Q3, our revenue growth, excluding AI native customer group, was 20% year-over-year, accelerating from 18% year- over-year in Q2, and we have seen this trend of accelerating growth continue in October. Regarding retention metrics. Our trailing 12-month net revenue retention percentage was 120% similar to last quarter and our trailing 12-month gross revenue retention percentage remain in the mid- to high 90s. And now moving on to our financial results. Our billings were $893 million, up 30% year-over-year.
- 支持 该 GSI 在北欧的可观测性/Datadog 业务 2025 年约 +30%，而且在加速
A:它在加速。可观测性行业正在从被动监控向主动、自愈方法转变。增长大约为 30% 同比，这是健康的。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 支持 Q4 比 Q3 (约 25%) 加速，超出他们 28–29% 的预期
A:高于预期。之前的季度表现良好——大约 25%——而这种情况延续并有所加速。我们的预期可能大约是 28% 或者 29%，但不是 30%。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 口径不齐 渠道 4Q25 Datadog 业务 +20.8%，低于预期
A:We closed Q4-the quarter ended 25 December-a little over 20%. We basically closed at 20.8%, very close to 21%, but still below expectations. (The December numbers have just been finalised. It is 20.8%.)
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 反对 Q4 走弱的原因是季节性，而且年底预算冲量没有落到可观测性
A:Over the last two quarters of 2024 across my observability practice, that has been the trend. In Q4, I saw a general dip, attributable to two factors: seasonality because of year-end events, and the year-end budget flush. In our last discussion, we talked about large deals and Black Friday. The year-end budget flush did not benefit observability as much.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 支持 渠道 2025 全年约 +22.7–22.8%
A:The full-year Datadog growth for 2025, including Q4, will come to around 22.7 or 22.8%-basically a little under 23%. The full year.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 支持 24% 的目标中至少 21 个百分点来自非 AI 部分 (他自称不科学)
Cloud security SIEM is also doing well. If you asked me to break down a 24% growth target, I believe-this is not scientific-that at least 21 percentage points would still come from the non-AI part.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 支持 北美专家A 2025 年 DDOG 业务约 3500 万美元，+24%，符合预期
2025 DDOG业务规模约US$35m, **24%yoy，符合预期**
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 支持 传统行业客户重启 IT 现代化、加速上公有云
传统行业客户今年再重启IT modernization, 加速向公有云迁移
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 支持 1H26 的增量中 85% 来自非 AI workload
预计85%增量来自非AI workload
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 支持 欧洲专家C 4Q25 +30% 以上，比 3Q 的 25% 加速
* 该专家DDOG业务体量较小，客户总体observability需求在2025年有所加速 * 4Q25 DDOG业务增长30%以上，**超预期1-2ppts, 较3Q25的25%yoy加速；2025全年增长30%yoy**
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 口径不齐 渠道 Q4 约 +25%，落在预期下限
A:它落在我们预期的下限，约为 25%。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 支持 增长主力仍是基础设施监控，以及虚拟机向容器/微服务的迁移
A:是的。退一步讲，尽管我对日志管理做了一些评论，但仍在推动和带动可观测性增长的是基础设施监控和基础设施现代化。从传统虚拟机向容器和微服务的转变正在推动增长。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 支持 非 AI 原生客户收入增速由 20% 加速到 23%，1 月延续
Now to dive into some of the drivers of our Q4 revenue growth. First, overall, we saw robust sequential usage growth from existing customers in Q4. Revenue growth accelerated with our broad base of customers, excluding the AI natives to 23% year-over-year, up from 20% in Q3. We saw strong growth across our customer base with broad-based strength across customer size, spending bands and industries. And we have seen this trend of accelerated revenue growth continue in January.
- 支持 加速趋势在 1 月延续
First, overall, we saw robust sequential usage growth from existing customers in Q4. Revenue growth accelerated with our broad base of customers, excluding the AI natives to 23% year-over-year, up from 20% in Q3. We saw strong growth across our customer base with broad-based strength across customer size, spending bands and industries. And we have seen this trend of accelerated revenue growth continue in January. Meanwhile, we are seeing continued strong adoption amongst AI-native customers with growth that significantly outpaces the rest of the business. We see more AI-native customers using Datadog with about 650 customers in this group.
- 支持 Q4 bookings 创纪录 $1.63B (+37%)，18 笔 >$1000 万 TCV，2 笔 >$1 亿
This acceleration was driven in large part by the inflection of our broad-based business outside of the AI-native group of customers we discussed in the past. And we also continue to see very high growth within this AI-native customer group as they go into production and grow in users, tokens and new products. Our go-to-market teams executed to a record $1.63 billion in bookings, up 37% year-over-year. This included some of the largest deals we have ever made. We signed 18 deals over $10 million in TCV this quarter, of which 2 were over $100 million and 1 was an 8-figure land with a leading AI model company. Finally, churn has remained low, with gross revenue retention stable in the mid- to high 90s, highlighting the mission-critical nature of our platform for our customers. Regarding our Q4 financial performance and key metrics, revenue was $953 million, an increase of 29% year-over-year and above the high end of our guidance range.
- 支持 财富 500 强渗透 48%，中位 ARR <$50 万，空间大
55% of customers use 4 or more products, up from 50% a year ago, 33% of our customers use 6 or more products, up from 26% a year ago, 18% of our customers use 8 or more products, up from 12% a year ago, and as a sign of continued penetration of our platform, 9% of our customers use 10 or more products, up from 6% a year ago. During 2025, we continued to land and expand with larger customers. As of December 2025, 48% of the Fortune 500 are Datadog customers. We think many of the largest enterprises are still very early in their journey to the cloud. The median Datadog ARR for our Fortune 500 customers is still less than $0.5 million, which leaves a very large opportunity for us to grow with these customers. So we're landing more customers and delivering more value, and we also see that with the ARR milestones we're reaching with our products. We continue to see strong growth dynamics with our core 3 pillars of observability: infrastructure monitoring, APM and log management as customers are adopting the cloud, AI and modern technologies.
- 支持 核心 APM 加速到 mid-30%，三大支柱各自 ARR 规模
And this includes continued rapid growth with Flex Logs, which is nearing $100 million in ARR. And our third pillar, the end-to-end suite of APM and DEM products also crossed $1 billion in ARR. This includes an acceleration of our core APM product into the mid-30% year-over-year and currently our fastest-growing core pillar. We have now enabled our customers with the easiest onboarding and implementation in the market, while delivering unified deep end- to-end visibilities into their applications. Now remember that even with these 3 pillars, we are still just getting started as about half of our customers do not buy all 3 pillars from us, or at least not yet.
- 中性 客户自建/外包：CEO 称"有时会优化一下"，但总体越用越多
No. I mean, look, it's something that happens sometimes, but it's a small minority of the cases. Like the general notion is customers start with some homegrown or attempts to do things themselves, then they move to our product, then they scale with our products. Sometimes they optimize a little bit along the way, but the general notion is they do more and more with us. They rely on us for more of their -- solving more of their problems, and they outsource the problem, and increasingly, the outcomes, to us. So, I don't think that's changing. Look, we'll still see customers here and there that choose to in-source it and do it themselves, again, hugely for cultural reasons. I would say, economically or from a focused perspective, it doesn't make sense for the vast majority of companies. And we even see teams at hyperscalers that have all the tooling in the world, all the money in the world, all the know-how in the world and that still choose to use our product because it gives them a more direct path to solving their problems.
- 支持 4Q25 收入 +29%，比 3Q 加速 1pt
4Q25收入增长29%yoy，比3Q25加速1pt并超买方预期1pt
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 支持 非 AI 的 core cohort 从 20% 加速到 23%，1 月继续加速
+ **Core cohort（非AI客户）的收入继续加速增长至23%yoy (3Q25%: 20%) 且在一月份持续加速**
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 支持 客户在搭建观测基础设施 (监控、排障所有系统)，推高 AWS 消耗
DS: And this is partially due... not primarily, partially due to... This is some specific clients, but really the storage of data and the bringing up of some core infrastructure pieces for multiple clients around their observability needs, you know, monitoring all of their systems and being able to track and troubleshoot all of those systems. So a lot of times those systems when they're installed, especially the cheaper data, if you will, goes into AWS. So we're seeing a lot of that in our particular business. So that's like Kubernetes stuff, that type of work, we're seeing a lot there. We're still
来源背景：Dave Sidwell，Improving Holdings 芝加哥区总裁 (President for Chicago)，多云咨询/实施伙伴 (Industry Consultant，标题明示)；主持人 Brian Oh。视角局限：只看自己公司的客户，增速是他们自己 AWS 业务的数字，不是 AWS 整体。 - 中性 伙伴的 AWS 业务 3 月季度预测 8%，实际 10%
DS: I think we were forecasted about 8% this quarter. We realized 10%. So that'll be our story on GCP, too, in terms of beating the forecast. This one's a little different in the growth reasons, which I think is what you're going to ask next.
来源背景：Dave Sidwell，Improving Holdings 芝加哥区总裁 (President for Chicago)，多云咨询/实施伙伴 (Industry Consultant，标题明示)；主持人 Brian Oh。视角局限：只看自己公司的客户，增速是他们自己 AWS 业务的数字，不是 AWS 整体。 - 支持 UBS 客户检查：支出稳定增长约 8–15%，AI 预算重配下无份额流失
The firm remains constructive ahead of 1Q26, with customer checks (covering ~88% of ARR) indicating steady spend growth (~8–15%) and no share loss despite broader AI-related budget reallocation. Core demand remains supported by hyperscaler-driven infra trends, while AI is pressuring non-core spend but not Datadog itself. UBS views DDOG as a relative AI beneficiary with limited disruption risk, noting low appetite for DIY observability and no signs of OpenAI/Anthropic moving into the space, maintaining Buy with valuation still reasonable vs. growth.
来源背景：TMT Breakout (TMTB) 日报，转述 JPM (渠道检查及一位行业联系人) 与 UBS (客户检查) 的报告摘要，二手转述，看不到底稿。 - 反对 传统和中型企业因预算被 AI 项目挤占而对可观测性支出更谨慎
We spoke with three former / current Datadog (DDOG) sales and channel experts this quarter, and additionally collected first-hand usage feedback from a large North American enterprise customer. Overall, DDOG’s core business held up well in 1Q26, but with clear bifurcation: the AI-native cohort (including AI startups within the AWS ecosystem) grew above expectations as cloud consumption accelerated, while traditional and mid-market enterprises turned more cautious on observability spending as more of their budget was reallocated toward AI projects. Customer feedback on Bits AI was muted; enterprise customers concluded that the cost / benefit tradeoff did not yet pencil after testing, and only the AI-native cohort showed relatively positive uptake. OpenAI’s migration off DDOG is progressing somewhat more slowly than we previously assumed, with full migration now expected around 4Q26, while Anthropic’s usage continues to grow rapidly off a low base. **We expect 1Q26 revenue and 2Q26 guidance to extend the trend of beating consensus by ~5–6 ppts seen in recent quarters, and view the setup heading into print as healthier given the delayed OpenAI migration gap.**
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 企业 AI 投入在增，但可观测性支出还没跟上，被视为可以等
~2/3 of global enterprises have adopted some level of AI applications, but <5% have deployed corresponding observability / security layers
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 支持 前销售 B 估其 book 新业务 1Q26 +30%，并预期 3Q–4Q 加速到 35–40%
* Previously covered DDOG’s Commercial business (customers <25k employees) on the East Coast / Central US, managing 15 AEs and 6 SDRs * Expert estimates new business y/y for that book: 3Q25 +22–25%, 4Q25 +22–25%, **1Q26 +30% (above the 20–25% expectation)**; **expects 2Q26 +25–27%, with 3Q–4Q potentially accelerating further to +35–40%** * Existing-customer expansion rate: 1Q26 ~105%; 2Q26–3Q26 ~110%; 4Q26 potentially up to 115–125% * Acceleration is driven primarily by AI-native customers
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 支持 DDOG 自 2024 年底起月消费过 $500 就签约，再快速 upsell 到 $10–20k+
+ AI-startup cloud consumption is accelerating sharply — free credits that previously lasted 1–2 years are now burned in 3–6 months, triggering paid observability adoption earlier + Since late 2024, DDOG has increased focus on new logos, signing customers as soon as monthly consumption exceeds ~US$500 and then driving rapid upsell to ~US$10–20k+ * Open-source tools such as Grafana are cheaper, but the setup and maintenance burden is high; many customers still choose DDOG once they factor in total cost
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 客户开始用 AI 替代部分人头，AI 预算挤占其他软件预算，拖慢 DDOG 传统业务
+ Customers are starting to substitute AI for some headcount; software is no longer the top investment priority. AI budgets are crowding out other software budgets, slowing DDOG’s traditional business + Only a few legacy industries (e.g., banking) are accelerating cloud migration in response to AI; the SMB customers under coverage are largely already cloud-native + AI budgets are mostly directed toward LLM and internal AI tooling development, with most still in POC and not yet in
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 口径不齐 只有银行等少数传统行业因 AI 加速上云，其覆盖的 SMB 基本已云原生
+ Customers are starting to substitute AI for some headcount; software is no longer the top investment priority. AI budgets are crowding out other software budgets, slowing DDOG’s traditional business + Only a few legacy industries (e.g., banking) are accelerating cloud migration in response to AI; the SMB customers under coverage are largely already cloud-native + AI budgets are mostly directed toward LLM and internal AI tooling development, with most still in POC and not yet in
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 支持 非 AI 客户收入增速连续加速：19% (Q1'25) → 23% (Q4'25) → mid-20s (Q1'26)，管理层归因于上云、多产品采用、客户自己用 AI；mid-20s 未给精确数
I'll talk more about that in a bit. Even more impressive was the growth in our non-AI customers. Non-AI customer revenue growth accelerated again this quarter to mid-20% year-over-year, up from 23% last quarter, and 19% in a year-ago quarter. We think this is a sign of strong continued cloud migration, greater adoption of our products, and customers of all kinds accelerating their use of AI. Finally, churn has remained low with gross revenue retention stable in the mid-to-high 90s highlighting the mission critical nature of our platform for our customers. Regarding our Q1 financial performance, and key metrics, revenue was $1.01 billion, an increase of 32% year-over-year, and above the high end of our guidance range.
- 支持 CFO 重复同一口径 (剔除 AI natives)，并称增长在客户规模、消费档位、行业上都广泛
These included wins in Observability and included some of our newer products like security, data observability and Flex Logs. And our new logo average land size also set a record and more than doubled year-over-year as we continue to land larger deals. Revenue growth accelerated with our broad base of customers, excluding the AI natives to mid-20s percent year-over-year, up from 23% last quarter and 19% in the year-ago quarter. We saw robust growth across our customer base with broad-based strength across customer size, spending bands and industries. Meanwhile, our AI native customer growth continues to significantly outpace the rest of the business. This group continues to diversify and grow, including 22 customers spending more than $1 million annually and five spending more than $10 million annually.
- 支持 TTM NRR 回升到 low-120s (上季约 120)，GRR 维持 mid-to-high 90s
This group continues to diversify and grow, including 22 customers spending more than $1 million annually and five spending more than $10 million annually. This group includes the leading companies and foundational models, code-gen tools, and vertical specific AI solutions. Next, regarding our retention metrics, our trailing 12-month net revenue retention percentage was in the low 120% percent, up from about 120 last quarter, and our trailing 12-month gross retention percentage remains in the mid-to-high 90s. Now moving on to our financial results. Billings were $1.03 billion, up 37% year-over-year, and remaining performance obligations, or RPO, was $3.48 billion, up 51% year-over-year, with current RPO growing in the mid-40s percent year-over-year.
- 支持 Q1 存量客户环比用量增长是 2022Q1 以来最强；ARR 每个月都在加速，4 月延续
Our Q1 revenue was $1.01 billion, up 32% year-over-year. Our 6% quarter-over-quarter revenue growth is the highest for Q1 since 2022, and our $53 million quarter-over-quarter revenue added is the highest ever for Q1. That included the strongest quarter of sequential usage growth from existing customers since the first quarter of 2022. We also delivered an all-time record for sequential ARR added to the quarter. ARR growth accelerated in each month of Q1, and we see a continuation of these healthy growth trends in April. We also achieved strong new logo bookings. New logo annualized bookings, set a new all-time record by a significant margin and more than doubled versus a year-ago quarter.
- 支持 新客户年化 bookings 创纪录、同比翻倍以上；平均 land 规模也翻倍以上
ARR growth accelerated in each month of Q1, and we see a continuation of these healthy growth trends in April. We also achieved strong new logo bookings. New logo annualized bookings, set a new all-time record by a significant margin and more than doubled versus a year-ago quarter. These included wins in Observability and included some of our newer products like security, data observability and Flex Logs. And our new logo average land size also set a record and more than doubled year-over-year as we continue to land larger deals. Revenue growth accelerated with our broad base of customers, excluding the AI natives to mid-20s percent year-over-year, up from 23% last quarter and 19% in the year-ago quarter. We saw robust growth across our customer base with broad-based strength across customer size, spending bands and industries.
- 支持 多产品渗透提升：4+ 产品 56% (去年 51%)、6+ 产品 35% (28%)、8+ 产品 20% (13%)
Turning to product adoption. Our platform strategy continues to resonate in the market. For example, 56% of our customers now use four or more products up from 51% a year ago. Thirty-five-percent of our customers use six or more products up from 28% a year ago, and 20% of our customers use eight or more products, up from 13% a year ago. So we're landing more customers, and delivering value across more products, and our business continues to grow. Our total ARR now exceeds $4 billion, and our quarterly revenue exceeded $1 billion for the first time in Q1.
- 支持 非 AI 加速的原因：平台扩展/整合 + 销售产能扩张且没牺牲人效 (CFO 回答 Lenschow，没拆分上云 vs 份额)
Thank you. Yeah, it's a number of things including. One is the expansion of the platform, the consolidation, the successful ramping of sales capacity, while not jeopardizing productivity, which has resulted in increasing ARR and a good environment as well. And I think that's what we said last time. There are a number of factors.
- 支持 管理层称剔除 AI 投资顺风后仍有明显加速，并自称在规模化竞争对手中份额领先 (未量化 AI 顺风占多少)
And certainly what we're proving out here is the investments we've made in go-to-market and are continuing are paying off and were the right decision. Oli, anything there? Yeah, and look, at the end of the day there's clearly some market tailwinds with the adoption of AI. But also we are outperforming all of our competitors at scale and we're taking share and that relates to the structure of our platform, to where we expand with new products, the way these products are maturing and starting to win in their respective categories, and the way we've successfully grown the sales capacity. Certainly the AI investment trend has helped, what we're trying to do is separate that. And AI investment is probably helping the overall as well. But when you really take that out, you see a very pronounced acceleration here. And that has to do with the factors that I mentioned and Oli talked about. Perfect.
- 中性 宏观：未见消费/电商受地缘影响，SMB 很强 (定性)
we had a very strong quarter across the board. We have, a multi-industry, multi-geography type of quarter. SMB was very strong. And that the source of our guidance and our raises are at the core, that type of performance. We haven't seen any particular effect in the consumer businesses or e-commerce businesses yet. We basically have a continuation of trends in those businesses, travels and things like that that are very similar to the other industry. So, we haven't seen it yet.
- 支持 核心非 AI 客户增速升到 mid-20s，4Q25 为 23%，3Q25 为 19%
+ The “core” non-AI customer cohort **grew at a mid-20% year-over-year rate, up from 23% in 4Q25** and 19% in 3Q25. + This implies **AI-native customer growth accelerated to 17% QoQ**, after growth was impacted by OpenAI’s migration out in 4Q, demonstrating Anthropic and other AI native customers’ consumption picking up. * 4Q adj OPM beat expectation by 1ppt, demonstrating strong operating leverage despite GPM being lower, likely due to the discount offered to large customers * 2Q26 revenue guidance is set at 30% year-over-year growth, surpassing market expectations by 6-7 percentage points. **Based on a normal beat cadence of 4-5 percentage points, actual growth could reach 35% year-over-year, the highest** **since 2022.** * Full-year revenue guidance is now 26-27% year-over-year growth. Factoring in **typical beat cadence, FY26 revenue is projected to accelerate to low-to-mid 30% growth.**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 CEO 称收入加速在 AI 与非 AI 客户群都广泛；Q1 每月 ARR 增速都在加速，4 月延续
*DDOG: ARR growth accelerated in each month of Q1 and “we see a continuation of these healthy growth trends in April”; CEO says revenue growth acceleration was broad-based across both AI and non-AI customer cohorts, signaling stronger
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 1Q26 公布数字与街上预期 (Street 口径来源未注明)，转述 CEO 评论。二手转述，看不到底稿。 - 支持 NRR 从约 120% 低点回升到 low-120s；$100K+ 客户 4,550 家
* Revenue $1.006B, +32% y/y (last q +29%) vs Street $960M, with growth accelerating driven by continued enterprise adoption, AI-native customer strength, and core observability demand. Calculated billings were $1.03B, +37% y/y vs Street $932M, a notable reacceleration and well ahead of expectations. * Operating margin came in at 22.2% vs Street ~21.3%.FCF was $289M vs Street $274M. * 2Q revenue guide of $1.07-1.08B (+29-31% y/y) came in ahead of Street at $994M (+20%), while FY26 revenue guidance was raised to $4.30-4.34B (+25-27%) vs Street $4.12B. The implied guide raise was materially above the size of the 1Q beat, pointing to a stronger 2H setup. * FY26 operating margin guidance was also lifted to 21.9-22.6% vs Street ~21.4% * Customer metrics remained healthy, with $100K+ ARR customers growing to 4,550 from 4,310 last quarter, while net retention rate improved into the low-120% range after bottoming near ~120% last quarter. Total customers reached 33.2K. * Mgmt commentary and investor focus centered around AI-native customer contribution, momentum in newer products like LLM Observability, Cloud SIEM and CNAPP, as well as whether broader enterprise AI deployments are beginning to drive incremental consumption. Investors also focused on whether FY guide conservatism implies further upside potential if AI-related workloads continue ramping through 2H.
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 1Q26 公布数字与街上预期 (Street 口径来源未注明)，转述 CEO 评论。二手转述，看不到底稿。 - 反对 同一 panel：软件已不是 CIO 名单首位，token 和 AI 才是
Software will have winners and losers. Datadog is viewed as the clear winner from everything. Over the last 45 days, more separation has emerged. New ways of writing code are fundamentally different. Software is sticky for a period, but it is no longer at the top of the CIO list — tokens and AI are.
来源背景：TMT Breakout (TMTB) 日报，转述 Sohn 投资大会上 Whale Rock (Alex Sacerdote) 和 Maplelane (Leon Shaulov) 的 AI panel 观点。买方观点，TMTB 摘要。二手转述，看不到底稿。 - 支持 Q1 加速覆盖 AI 原生和非 AI 两部分，管理层称非 AI 加速“更有意思”
Maybe what surprised you, and what do you see kind of persisting going forward? Yeah. So, look, the – and as we said on the call, I think on the earnings call, we saw acceleration across every single part of our business. We saw acceleration with the brand new AI native companies, whether they're small or very big. We saw also acceleration with the rest of our business, which is even more interesting. So the Non-AI part, all the companies that were around before AI starting – taking over. I think the drivers are little bit different there.
- 口径不齐 非 AI 客户加速的主体是上云/现代化和销售扩张，企业自身 AI 上线仍是小驱动 (未量化)
I mean a year ago to now, what is going on in this core drivers that's pushing that a little bit higher. Yeah. I mean, look, so we do have some proof points that those customers are starting to adopt some AI in production. It's still small relative to the size of those customers and still a small driver of our growth there like we see traffic, for example, MCPs and traffic also to our (00:19:26) product. We see a very strong growth. I think we released some numbers in the earnings call on that, but that's still a small number.
- 支持 非 AI 业务增长来源：上云、销售产能、新区域、多产品达到 PMF
We see a very strong growth. I think we released some numbers in the earnings call on that, but that's still a small number. Just give us an idea that these customers are moving there, but the bulk of it for these customers is that they are still modernizing, moving to the cloud. We're getting to more of them because we have successfully built up sales capacity. We go to market in more regions and more segments. We get great return on investment on that and then we've been expanding the set of products we can sell in the categories we sell into, and we have enough of those products that are reaching product market fit...
- 支持 Gartner 口径市场份额仅 13.6%，管理层认为不靠 AI 也有 5x 空间
And part of that is those companies have to understand they need to modernize and modernize faster so they can be ready for AI. Part of it is just that we're still very early in cloud migration. There's a number we like coding which is that I think Gartner has this report every year on the market share in ITOM and a number of other fields, and we're the leader in observability which we are the number one there but we still have only 13.6% of the market according to them. So this tells you how early it is in the market for us, how much opportunity there is ahead of us, even factoring out all of the new developments with AI, all of the new explosion of demand we can see there. And so we think we're still early in a super cycle there.
- 支持 Pomel：AI 原生和非 AI 业务都在加速
Pomel: *“We saw acceleration across every single part of our business. We saw acceleration with the brand new AI native companies, whether they’re small or very big. We saw also acceleration with the rest of our business, which is even more interesting — the non-AI part, all the companies that were around before AI started taking over.”*
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 支持 非 AI 加速原因：企业为 AI 加快现代化，上云仍处早期 (未量化)
*“What’s even more interesting for us is that the bulk of our business, the rest of our business that existed before AI became a thing, that business is accelerating as well. Part of that is those companies have to understand they need to modernize faster so they can be ready for AI. Part of it is just that we’re still very early in cloud migration.”*
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 支持 加速不来自特定客户，覆盖全面
They're all in the filings. But we've seen acceleration. What's the most exciting to us is that, it is not a specific customer or specific set of customer, specific side of the business. We've really seen acceleration across the board. And the way we look at the customer base today is we look at both what we call the AI natives, which are largely companies, they were started in the last few years built on AI or companies that we're seeing a little bit before that are, essentially, the large builders of the most important building blocks of AI today. So we see acceleration in that category, which sort of makes sense.
- 支持 AI 公司的基础设施需求不只 GPU，agent 调用工具带动 CPU 等通用设施，“无聊的”核心观测还有 5-10x
The other half, the more boring half, is that all of that is driving the need for standard observability. Those large AI companies are consuming large amounts of infrastructure and applications in the same way the other companies are. And as I said earlier, like we have less than 14% of that market and that market is growing, and is going to grow faster I think now as you see this huge CapEx investments in infrastructure in general, and it's not just GPUs anymore, it's GPUs and CDUs and everything else as these agents need to use tools. And so, I think that part of the business, the boring existing part, that's not new form factors, new applications, new anything. That part has another 5x or 10x into it, and that's very exciting. Well, you heard it here first. The only business that will remain after the robots take over, everything else is observability and Datadog.
- 支持 DDOG 的核心客户应该是部署 AI agent 的传统企业 (不会自建)
**这意味着 OpenAI 流失对 DDOG 的影响主要在 2026-2027 年的"短期阵痛",长期来看 DDOG 的 AI Observability 故事仍然成立**——因为 DDOG 的核心客户从来就不应该是超大 AI 公司,**而是部署 AI Agent 的传统企业**(这些企业不会自建)。
来源背景：Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。 - 支持 传统企业只有 25-30% 工作负载上云并现代化，持续迁移是大单来源
Yeah, it's definitely a confluence of both. So you have – in a customer base that has been around for a long time, so therefore they're not cloud-natives, they didn't just get birthed, so they have legacy technologies and they've – there's – they have a long way to go. There's somewhat 25%, 30% of workloads in the cloud right now and modernized. So they are continuing to find use cases and modernizing. Datadog's platform is getting bigger and better so we're consolidating market share onto our platform, and we're getting better about delivering the enterprise service model, whether that be channel partners, customer service, technical help, the whole ecosystem to be able to deliver. So all of that has been what we've been investing behind for some time that is bearing fruits.
- 支持 70%+ 工作负载未上云，长期驱动；新技术 (AI) 历史上会加速迁移
So we always have believed and we're seeing that this is a very long time investment cycle because such a high percentage of applications and infrastructure are still in legacy technologies. So we've always said that if you want, look at this, this is a secular trend and it's cyclical – it can be cyclical but it's also secular, and it's got a lot. I mean, you have 70%-plus of workloads that aren't in the cloud, so we believe this has very long legs. In terms of is it going to be a straight line, meaning on every month, is it going to have the exact same thing? We don't know. It's likely, if proven, you're going to have periods of investment, you may have periods of optimization. It's cloud software, it's consumption cloud software. So we can't predict exactly what the line's going to be in every moment, but we have – confident that it's very durable and it's very long because we're attached to huge market drivers. Now we also – if history repeats itself, we found that new technologies of which AI is happening is a huge accelerant of this conversion.
- 支持 加速已持续 3-4 个季度，来自平台多产品、AI 原生需求和过去一年的 GTM 投入
And this has been building on itself. We've also been investing substantially in our go-to-market over the last year in order to deliver this to our customers. So this has been building on itself and essentially, in the first quarter, it continued that trend of acceleration in many areas, the non-AI-natives, the AI natives, different geographies, enterprise down to SMB, and then all – those were all contributing factors in producing the first quarter, but the seeds of that started three to four quarters ago and have been building on itself. Let's talk a little bit about the multi-products. You guys give a lot of metrics...
- 支持 AI 让新项目迁到云上，云规模扩大会增加 Datadog 的云上观测收入；但不能上云的传统客户 Datadog 覆盖不到
Datadog有一个限制，它只是云上的可观测解决方案，对云下没有提供产品。现在有一个趋势，随着AI浪潮，很多服务需要搭在云上，包括token消耗、租赁CPU做计算，很多原先客户有新的项目会迁到云上，云的规模会扩大。云规模扩大一定会让Data dog云上的观测营收增加。但对于传统无法上云的客户，Datadog影响不到。
来源背景：未具名；以 Elastic (ES) 内部视角回答 (清楚 ES 组织架构调整、API 文档策略、Jina 整合)，推测为 Elastic 现任或前任员工，带明显中国区视角 (国内定价、阿里/腾讯 ES 服务)。局限：竞争对手视角；对 Datadog 的判断是间接的。 - 支持 AI/agent 让安全和观测预算显著增长，同时也在找地方优化以抵消
Now that's very eye-opening, and that's what everyone is going after to find those mitigation for those, and people are pumping money. It all depends how mature is your organisation, are you regulated, how large, are you publicly traded, those all revolves around this. I've seen the spend allocation to cybersecurity and observability with this whole AI agentic actually is growing. The budget is growing.
来源背景：匿名专家 ([Expert])。Third Bridge 主持人介绍为 'former Head of CCoE [Cloud Centre of Enablement] and Platform Engineering at PNC Financial Services'；但专家自述 '我是' PNC Cloud CoE 和平台工程负责人、在该行两年 。负责容器、中间件、数据平台及平台安全 (本地和云)；上一家机构曾把 Splunk 整合迁移到 Google Chronicle 。视角局限：受严格监管的大型银行买方，偏 '最佳单品 + 多厂商冗余' 策略，不代表一般企业；角色偏平台/安全，不直接管观测账单细节；对 Datadog 的评价较笼统、无数字。 - 反对 Datadog 支出没有增加：核心指标也在减少、迁往 Grafana，Datadog 上的指标、dashboard、日志量都少了很多
暂时没有增加。公司以前的核心指标等内容其实在减少，很多都迁移到了开源的 Grafana系统。相对来说，Datadog 上的指标数、dashboard数据、日志量等都减少了很多。Datadog 的计费方式是根据节点数、指标数和数据量来计费，只要这些减少，费用自然不会增加。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 支持 Truist：非 AI 原生企业越来越多部署 agent，需要更多遥测和监控
Truist upgraded DDOG to Buy and raised its PT to $300, arguing investor concerns around AI optimization are being overwhelmed by the urgency of enterprise AI adoption. Recent customer and partner conversations suggest Datadog remains in the early innings of AI-driven growth, with non-AI-native enterprises increasingly deploying agents that require more telemetry, monitoring, and observability. The firm also highlighted improving confidence in the durability of DDOG’s relationships with frontier AI labs, reducing a key overhang on the stock. While valuation has expanded, Truist believes Street still underestimates the durability of consumption growth and the size of the opportunity as AI adoption broadens across the enterprise.
来源背景：TMT Breakout (TMTB) 日报，转述 Truist 升级报告摘要 (基于客户与伙伴交流)。二手转述，看不到底稿。 - 口径不齐 专家观察到的客户 Datadog 消费 Q1 同比在 mid-to-high teens，Q2 温和加速到 high teens/low 20s (明显低于公司报告的营收增速)
嗯。First question, what was the growth growth rate of customers' consumption on DataDog in this year Q1 and Q2, and what are the reasons behind the changes in growth rate? Yeah, um, so the consumption in Q1 was um, in general mid to high teens, I would say percentage year over year. um, I think the Q is, Q2 is showing um, medium acceleration, uh, probably in high, um, teens or you can call it low 20s. um, you know, customers, uh, reasons and the drivers from customers scaling AI and cloud native workload. Primary drivers were, um, more Kubernete adoption.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 中性 增长驱动是 K8s、trace 量、APM、安全监控和 AI/agent 可观测性，但最大的底层驱动仍是传统上云和应用现代化
um, I think the Q is, Q2 is showing um, medium acceleration, uh, probably in high, um, teens or you can call it low 20s. um, you know, customers, uh, reasons and the drivers from customers scaling AI and cloud native workload. Primary drivers were, um, more Kubernete adoption. Um, there was a higher trace volume, there was broader APM deployment. Um, I've seen expansion in security monitoring and also the observability observability requirements are growing for AI applications and agent-based architectures. Where growth, um, kind of slowed down, I think was typically due to FinOps initiatives, log volume optimization, um, cloud cost control, and where customers were shifting lower value telemetry to um, a cheaper storage, um, and analytics platform. But overall, the AI driven workloads, um, I think they are meaningful drivers to data dog consumption in 26. particularly, uh, where customers are deploying Gen AI applications into prod. Um, even though the traditional cloud transformation and application modernization, um, still remain the largest underlying drivers. 在Q1的话的消耗大概是 大概是就是他说的是mid to high teens,就是大概是15,就是十几,就是15到20之间的一个数字吧, 然后到Q2的时候的话是high teens to low 20,就是可能是接近20或者20几,就是比25低那样的一个一个水平。 然后它的一个增长的原因或者说驱动力的话,第一个呢就是对于K8S, Kubernetes的采用更多, 然后trace就是追踪的追踪的量也更多,然后更多的API的部署,然后安全监控方面的话也出现了比较大的增长, 然后AI的应用还有基于agent的架构的那种可视化需求也是在增长。
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 口径不齐 纪要称 Q1 消费同比 '二十个百分点低位'、Q2 '增速放缓' (与音频原话相反，应视为整理错误)
A:第一季度的消费同比增长在二十个百分点低位。随后第二季度增速放缓。增长的驱动因素包括更广泛的产品采用、更高的 Trace 流量以及大规模的 APM 部署。我也看到安全监控方面的扩展，随着 AI 应用和基于 agent 的架构增多，可观测性需求在增长。增速放缓通常是由于 FinOps 措施、日志量优化、云成本控制，以及客户将低价值的遥测数据转移到更便宜的存储层或数据平台上。
来源背景：标题写明 '北美软件渠道商离职专家' ，作者 Lantian；内容与 Merit Datadog 产品进展调研 I (2026-06-16 音频) 逐段对应，判断为同一专家同一场的英文转写再翻译版。局限同音频 I：只看到渠道客户，无样本量，回答偏框架化。 - 支持 其客户 Datadog 支出 Q1 同比 +33%，Q2 预计约 +35% 小幅加速
然后增涨背后主要的一个驱动因素是什么? Yeah, so what is the growth rate of your customer spending on datadog in Q1 and Q2 this year and what are the reasons behind their changes in the growth rate? Yes, so uh, I mean Q2 is not completely finished yet, of course, but in in Q1 we saw a growth rate of uh uh slightly north of 30%, so uh 33%. Uh, and for Q2, we expect uh a similar growth, somewhat accelerating, but not not like a massive difference. So, uh, let's say 35% for um for Q2. Uh, and the reason I would say is uh mainly that existing customer go deeper, right? I mean, I think I always think about uh data dog as an as an ecosystem of different modules and products, right? And and they are uh there are synergies in between them and they are very good also at sort of uh following a land and expand uh strategy. So it's mostly uh sort of uh net retention and growth with existing customers driving it. Not exclusively, of course, but I think that's that's a growth growth story that is interesting. Mhmm.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 支持 增长主要来自老客户加深使用 (净留存/land and expand)，新客户也有贡献
I mean, I think I always think about uh data dog as an as an ecosystem of different modules and products, right? And and they are uh there are synergies in between them and they are very good also at sort of uh following a land and expand uh strategy. So it's mostly uh sort of uh net retention and growth with existing customers driving it. Not exclusively, of course, but I think that's that's a growth growth story that is interesting. Mhmm. Okay.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 中性 其企业客户约 40% 在用专门的可观测性工具 (专家自称是猜测)
Uh, yeah, exactly. So um if we start with the you know percentage of customer that use observability tools. Um I would say maybe 40% or something like that that that's like a guess from my side. and and remember we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 口径不齐 其渠道的 Datadog 业务今年约 +25–28%，Q1 约 20%，Q2 明显加速
Yeah, so we're just wondering what was the year of year growth rate of your customer spending on datadog in Q1, Q2 this year and what are the reasons behind the changes in the growth rates? In terms of data dog, um, growth rate for this, for this year. Um, we're definitely seeing around 25 to 28% growth. Um, this year. Um, when we look at the reasons for growth, um, I think generally we've seen the increased amount of traffic. Um, within infrastructures. AI is definitely one of the key reasons. Um, and I think a lot of organizations are trying to be much more efficient in terms of their control in terms of visibility, in terms of management. Um, and data dog has actually resonated a lot with um, maybe beyond mid market, I would say in a lot of the uh or many cases in the enterprise in the enterprise market. Mhm, okay. 今年他放在一起讲了,然后说今年目前的增速是25%到28%。 呃,增速主要的原因就是AI嘛,AI的兴起导致流量的增加,因为现在企业都想要提效,所以他们也需要这些呃可是化呀,然后更多的控制啊,更好的管理啊等等,这个也符合Datadog公司的一个定位吧,所以在企业级市场和中部市场都是比较受欢迎的。 嗯,呃想问一下就是专家的话,二季度比一季度的话,它这个增速的话可能会高多少呢? 然后专家认为到年到年底的话,每个季度都会有一个明显的加速吗? Yeah, so could you break down the growth numbers by Q1 and Q2? Because we want to see if there's an acceleration. And also, um, to you know, by the end of this year, would you would you say that every quarters would see an acceleration? Yeah, I would say so, Q1 was around 20% for us. Um, and Q2 has definitely accelerated. Um, we've definitely have seen growth in new customers as well as existing customers. So the acceleration is coming from both sides, I would say.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 支持 加速来自新老客户；老客户在 AI、安全日志上扩用，客户在做平台整合降本
Yeah, I would say so, Q1 was around 20% for us. Um, and Q2 has definitely accelerated. Um, we've definitely have seen growth in new customers as well as existing customers. So the acceleration is coming from both sides, I would say. Um, you know, we're seeing an increase in, you know, AI, security logs, for example, with existing customers, absolutely. Um, but we're also seeing approaches of consolidation, you know, a lot of customers are consolidation, consolidating the platform versus having, you know, a lot of different platforms within environment because that is cost savings, cost savings for them, absolutely. So overall, we're also seeing a lot of cloud, a multi-cloud strategy. If you look at most organizations don't just have Microsoft now, Azure or or AWS, but they have Alibaba, they have other clouds in their environment also.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 支持 企业通常至少有 3 个云，多云带来统一可观测性需求
So overall, we're also seeing a lot of cloud, a multi-cloud strategy. If you look at most organizations don't just have Microsoft now, Azure or or AWS, but they have Alibaba, they have other clouds in their environment also. So, you know, they have at least three different cloud providers within their environment, if you look at enterprise for example. Um, so that's expanded. Um, so, but also the existing existing customers are expanding their use cases within within the product groups, I would say.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 反对 80–85% 企业已有可观测性实践，净新客户已饱和，增长要靠 agent 流量
Um, this is, you know, the entire growth is based upon AI. Uh, not if you look at the observability market, you know, some people will say saturated. I mean, I would say 80% of enterprises, or 85% of enterprises have observability uh practices within the environment. There's not really net-net new customers that don't have observability, uh, within their environment. So from that perspective, uh net new customers is really saturated because most of the market has some sort of tooling involved. What the challenge is next is that traffic that will be generated by the agents essentially.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 支持 客户今年在 Datadog 上的平均支出增长 25–28%
A:今年的增长率大约在 25% 到 28% 之间。就增长原因而言，我们看到基础设施内的流量增加。AI 无疑是关键驱动因素之一，许多组织都在努力在控制、可见性和管理上提高效率。数据显示增长集中在中端市场；我不会说超出中端市场。很多增长在许多情况下来自 DevOps 市场。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 口径不齐 增长集中在中端市场和 DevOps，没有超出中端市场
A:今年的增长率大约在 25% 到 28% 之间。就增长原因而言，我们看到基础设施内的流量增加。AI 无疑是关键驱动因素之一，许多组织都在努力在控制、可见性和管理上提高效率。数据显示增长集中在中端市场；我不会说超出中端市场。很多增长在许多情况下来自 DevOps 市场。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 Q1 约 20%，Q2 加速，新老客户都在增长
A:会的。第一季度大约是 20%，第二季度确实加速了。我们在新客户和现有客户中都看到了增长，所以加速来自两方面。例如，在现有客户中我们看到了 AI 和安全日志的增加。我们也看到客户倾向于在一个单一平台上进行整合，而不是在环境中使用许多不同的平台，因为这能为他们产生节省。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 反对 硬件涨价挤压了今年的 IT 预算
从客户的 IT 预算角度看，今年由于硬件成本上涨，情况非常具有挑战性，这影响了整个 IT 预算。客户正在评估整个基础设施中的 AI。就中心化数据而言，我们看到 AI 采用率在上升，但客户也在确保数据受保护并符合合规要求。因此我们有很强的势头。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 Q1、Q2 收入同比都在 30–35%，Q2 很轻微加速
Uh we just want to know if there's an acceleration in Q2 compared with Q1, and if so, what's the reason behind it? Sure. So, I would say the year-on-year growth in revenue has roughly been in the same ballpark uh of around 30 to 35% in both Q1 and Q2. There's been a very slight acceleration in Q2 versus Q1, and the driver has been similar to uh the the overall growth rates of uh Datadog, right? So uh one of them is that uh the adoption of AI case, AI use cases continue to accelerate. And uh some of the Datadog products which have been out in general availability for some time now have accelerated in terms of adoption.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 硬件涨价没有挤占软件预算，企业是整体加 IT 预算
With the hardware increasing its price so much, do you think this has put a pressure on the software budget? And if so, does this have a negative impact on Datadog? That's a that's a good question but I feel like what's happened is that hardware budget has not really eaten into software budget. What's happened is actually that uh people have just increased their IT budgets in general, right? So I think people are in an experimentation phase where they're like, oh what happens if we increase our IT budgets by 20%? Do we become much more productive? And then there's all that experimentation with uh with with AI tools in order to see if that uh improved efficiency can actually happen, right? So I would say, look, why would we say that we uh are increasing hardware cost, right? It will be uh in order to get GPUs and more uh cloud compute and stuff like that, right? But the problem is that that alone will not give you a production grade uh sort of like AI use case, right? You'll then going to have to include software like the AI models itself and uh also the fundamental parts of the uh you know developer life cycle that you had to invest in before still remains true. Uh you still have to uh you know do observability for uh the AI application that you have coded using this increased hardware cost if it is critical to you and if it is uh customer facing, right? So what I've seen is more that the overall budgets have increased rather than hardware budgets having eaten into the software budgets because you can't just expect to get more hardware and that to yield results if you're not uh you know building the right software ecosystem around it. So uh it it's been more a case of overall budget increase for some of these companies than just investing in hardware.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 反对 Bernstein：ex-AI 需求 (约 85% 收入) 从 3Q/4Q 起可能走平
Bernstein raised its long-term estimates for Datadog, citing increasing confidence in the company’s AI platform opportunity, but downgraded the stock to Market Perform on valuation and elevated near-term expectations. The firm expects a strong 2Q, but believes growth could disappoint beginning in Q3/Q4 as ex-AI demand (~85% of revenue) flattens and AI lab growth starts to normalize. Bernstein expects Q4 revenue growth to slow roughly 500bps to ~29% y/y, well below investor expectations for 30%+ growth into next year and peak growth in the high-30% to 40%+ range.
来源背景：TMT Breakout (TMTB) 日报，转述 Bernstein 降级报告摘要。二手转述，看不到底稿。 - 支持 传统企业连续四个季度加速，不靠大规模 GPU 建设；云原生加速由 coding agent 和自建 agent 驱动
Phenomenally, yeah. We also, obviously, the AI natives growing very, very fast. But then we see the cloud natives have accelerating growth, at least in their consumption of compute and their – the number of applications they're writing, and that's largely driven, in their case, by coding agents and by them building their own AI agents. But we also see the traditional enterprises accelerating. Like we've seen acceleration of growth for our business of the traditional enterprises every quarter for the past four quarters, I think... Yeah... ...and we've seen an inflection point.
- 支持 企业观测投入被应用现代化、上云等大项目带动，整体稳健
Sid Roy: It's not too dissimilar than what we spoke about last time. From my perspective, the enterprise market is robust. For the most part, organizations are continuing to funnel investment into observability, security, things like that; oftentimes these are dragged in by larger initiatives, like modernization, business service, application modernization, cloud migration, things of that nature, or the rewriting of older legacy applications onto newer platforms. In general, the observability market has been robust. It's across the board.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 其 Datadog 业务 Q1 +31%、Q2 +33%，全年目标 35%
Sid Roy: Right. So for Datadog, we have a 35% annual growth target. In the second quarter, we did 33%. In the first quarter, we did 31%. So you could almost say like our...
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 下半年瞄准 37–38%，甚至接近 40%，靠大单、替换和扩容
Sid Roy: First quarter, we did 31% annual growth. In the second quarter, we did 33%. We have an annual target of 35%. And so for the second half, with our Datadog, our numbers are actually tracking a little bit closer. So we also are very much back second half loaded with larger deals, more replacements, more expansions. But in the second half, we know we're probably targeting... I would say... closer to 37%, 38% growth, maybe even getting closer to the 40%, which is
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 客户把 xMatters、PagerDuty 等旧系统整合到 Datadog On-Call (交叉销售)
how we intend on hitting our 35% target. Datadog, it's a little bit more, I'd say. One aspect with Datadog versus Dynatrace is that there is, when we think about logging... again, a key theme here is logging, semi-structured, unstructured data... is that with Datadog, we also have the benefit, unlike Dynatrace, of having customers spend allocated for security logging. So what we consider like SIEM, for security incident event monitoring. So we're very confident on the Datadog number. In addition to the meat and potato type stuff, like infrastructure, cloud monitoring, APM, logging, we're also with Datadog definitely more consolidation opportunities around things where customers are looking to consolidate legacy implementations of things like X Matters or PagerDuty onto Datadog On-Call. And with it, it's not always as apparent with Dynatrace, but with Datadog, they have a very discrete SKU related to LLM observability, and we are starting to see uptick in customers explicitly acquiring capacity for LLM observability. In essence, it's the same thing from a Dynatrace standpoint, where the customers are requiring this capacity to basically enable LLM tracing and LLM logging within their telemetry data sets.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 BNP：DDOG 2Q 销售高于计划几个点，基础设施监控、APM 和 AI 用例强
BNP’s quarterly reseller survey found cybersecurity remained the top IT spending priority over the past three months, while Datadog sales tracked a few points above plan on product functionality, competitive wins and strength in infrastructure monitoring, APM and AI-related use cases. However, overall IT spending intentions weakened sharply from April, and resellers expect cybersecurity momentum to trail other categories over the next three months, suggesting demand may have been concentrated in 2Q. DDOG’s forward spending score also fell to 2.0 from 5.4, creating a more mixed setup despite solid recent execution.
来源背景：TMT Breakout (TMTB) 日报，转述 BNP Paribas 季度经销商调查与 UBS 客户检查 (7 位客户 + 3 位伙伴) 摘要。二手转述，看不到底稿。 - 支持 BTIG 最大联系人 2Q 增速约 35%，1Q 32%，2025 年约 20%
BTIG reiterates Buy on both names and raises DDOG’s price target to $289 from $255 after partner checks showed broad acceleration, with its largest contact seeing growth improve to roughly 35% in 2Q from 32% in 1Q and about 20% in 2025. BTIG says AI inference is the clearest demand driver, benefiting infrastructure and log monitoring, LLM observability and network performance monitoring, while consolidation away from SolarWinds is adding further upside. BTIG also says DT checks were consistently positive, with partners expecting high-teens to 20% growth as AI expands observability demand, supporting its Buy rating and $47 price target.
来源背景：TMT Breakout (TMTB) 日报，转述 BTIG (伙伴检查) 与 Morgan Stanley (企业检查) 报告摘要。二手转述，看不到底稿。 - 支持 Metaplane 用户几乎都是 Datadog 客户，说明交叉销售依附于现有客户
**TL** [00:48:51]: Certainly, a lot cheaper or it used to be before Datadog rolled it into their platform, but like Bigeye, really much cheaper to run when you're not looking at huge scale. I think generally speaking, Monte Carlo got rid of its entire mid-market organisation and its commercial business to focus solely on the enterprise. There's no outbounding to small companies trying to win 30,000-50,000 deals. Everybody is whale-hunting at the big three. Metaplane probably does quite well with its Datadog customers, but there's still a gap in the market for those that aren't using Datadog. It's quite rare somebody would have Metaplane without having Datadog, for example.
来源背景：Terry Lucy (TL)。Third Bridge 页眉 'Executive, Go-to-market at CodeRabbit Inc'，介绍为 'former Strategic Account Executive at Monte Carlo Data' (明示)；自述在 Monte Carlo 多年，是伦敦办公室区域第二号员工，负责企业/战略客户 。视角局限：数据观测 (非 APM) 销售视角，对 Datadog 的看法来自竞争一侧；主题是 Monte Carlo，涉及 Datadog 的内容只有几段。 - 支持 专家 A 的 book 1Q 最终约 26%，2Q 跟踪 33–35%，高于本人预期
* **This expert’s 1Q26 ultimately landed at about 26% y/y, above the preliminary 21% we had last quarter; 2Q26 is tracking to 33-35%, a clear acceleration and above his own expectations**
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 deal 滑单明显减少，2Q 大部分 organic bookings 当季落地
+ **Deal slippage has fallen off materially** — in prior quarters at least half of signings would push into the following quarter, whereas most 2Q organic bookings landed within the quarter, and July is following the same pattern * Larger-TCV deals remain concentrated in high-tech and parts of financial services, and have not yet spread across the broader customer base * On the source of growth: new logos will still contribute less than one-third in 2H and into next year, with **the bulk of the incremental dollars coming from product expansion within the installed base**
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 增量主要来自存量客户多产品扩张，新客户贡献不到 1/3
+ **Deal slippage has fallen off materially** — in prior quarters at least half of signings would push into the following quarter, whereas most 2Q organic bookings landed within the quarter, and July is following the same pattern * Larger-TCV deals remain concentrated in high-tech and parts of financial services, and have not yet spread across the broader customer base * On the source of growth: new logos will still contribute less than one-third in 2H and into next year, with **the bulk of the incremental dollars coming from product expansion within the installed base**
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 客户用的 DDOG 产品从 2–3 个增加到至少 5 个
+ Customers previously running 2-3 DDOG products are moving to at least 5; the land products are still Infra and APM, then Logs, followed by Synthetics and Cloud Security + The buying center is no longer just the CIO/CISO — SRE, QA, DevOps and platform engineering are now decision-makers, pulling through Synthetic Monitoring, CI Visibility and Error Tracking * 2H growth mix: traditional products (applications/infrastructure/logs) account for at least two-thirds; security contributes at least 15% and Dev-related products about 10%, together roughly one-quarter; AI-related products total 10-15%, which he concedes is on the optimistic side * The multiplier effect of AI on application count: before mid-2023, a typical enterprise had roughly 100 developers supporting 400 applications; today headcount is up only 5-10% (and skewed to senior engineers) while application count is close to 650, up more than 50%
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 口径不齐 传统企业渠道 B 今年 +20% (与去年持平)，上限受预算而非需求约束
* **The upside is capped by budget rather than demand** — absent IT budget constraints, he believes this year could have been 28-30%
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 DDOG 净新增客户数同比约 +30%，Dynatrace 为 20–25%
+ **The main driver is agents** — on a 0-100 scale, customer agent consumption is still only around 10%, yet it has already produced a step-change in cloud spend; customers cannot control that consumption themselves and their tooling is fragmented, which is why demand for standardized observability is so strong + Banking and high-tech are the two fastest-moving industries on AI investment, but most customers still lack the governance framework to support AI in production, so the real ramp is a next-year story * **DDOG net-new customer count is up about 30% y/y versus 20-25% for Dynatrace**; DDOG is stronger in mid-market, and its enterprise new-logo opportunity is widening
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 F500 渠道 C 的 DDOG 业务 1Q +31%、2Q +33%，全年目标 35%
* **His DDOG business grew +31% in 1Q26 and +33% in 2Q26, against a full-year target of 35% and a 2H target of 37-38%, with close to 40% not ruled out**
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 Wells Fargo 调查 (约 $95M 年支出)：多数大客户预计 2026 增加支出，继续向 DDOG 转移份额
Wells Fargo says its 2Q survey, representing roughly $95M of annual Datadog spend, showed most large customers expect to increase spending in 2026 and continue shifting share toward DDOG from other observability vendors. Wells Fargo notes the survey was modestly softer than 1Q, partly because the latest respondent mix included smaller customers with narrower platform usage, but still views demand as constructive into 2H26. Wells Fargo says the results reinforce Datadog’s durable share-gain story and continued enterprise expansion.
来源背景：TMT Breakout (TMTB) 日报，转述 FundaAI 预览 (同 8/3 Funda 原文) 和 Wells Fargo 2Q 客户调查 (约 $95M 年 DDOG 支出) 摘要。二手转述，看不到底稿。 - 支持 非 AI 客户增速再加速到 high-20s (Q1 mid-20s，去年同期 18%)
Our revenue growth in Q2 has accelerated across our customer base. On one hand, our AI native customer cohort continued to grow and diversify, both in the number of customers we serve and the scale of those customers. But on the other hand, and as a great illustration of the breadth of trends across our business, revenue growth for our non-AI customers also accelerated again this quarter to the high 20% year-over-year, up from the mid-20s last quarter and 18% in the year ago quarter. Overall, we continue to see healthy trends in customer demand. Our broad base of customers, from the most nimble startups to the largest and most established enterprises are all adopting AI.
- 支持 管理层归因：各类客户都在用 AI，带动上云和现代技术使用
But on the other hand, and as a great illustration of the breadth of trends across our business, revenue growth for our non-AI customers also accelerated again this quarter to the high 20% year-over-year, up from the mid-20s last quarter and 18% in the year ago quarter. Overall, we continue to see healthy trends in customer demand. Our broad base of customers, from the most nimble startups to the largest and most established enterprises are all adopting AI. We think this is accelerating the usage of cloud and modern technologies, as well as their usage of the Datadog platform to observe, secure and act on their cloud and AI workloads. Regarding on Q2 financial performance and key metrics, revenue was $1.12 billion, an increase of 36% year-over-year and above the high end of our guidance range. We ended Q2 with about 33,400 customers, up from about 31,400 a year ago.
- 支持 非 AI 加速主要来自存量客户，驱动是上云带来的量 + 整合采用新产品
I mean, from what we can tell, it's very broad-based. And it's a largely driven by existing customers because that's the majority, like, when you think of what it takes to move that number, that's basically the majority of our business. Like we're not just going to move that with a few newer customers like this there, it's largely driven by the existing customers. And it's driven by both increases in volume because they're moving more workloads to the cloud and adoption of our newer products as they consolidate onto us. We think it's sustainable for one thing, if you compare it to what we have seen in the heady days of 2021 or the growth rates are accelerating, but they're still far below what we were seeing at that time. And so we don't create the same issue of customers having to digest very large increases multiple years in a row.
- 支持 非 AI 加速的大部分来自更多转型/上云/工作负载/现代化 (转录 'known AI' 应为 non-AI)；疫情前就有的非 AI 业务客户也在持续加速
And so just in terms of the enterprise AI, app dev cycle, what does that look like for Datadog over the last couple of quarters? Well, we do see broad adoption, and we see it in two ways. One is we see it manifest itself is in just more transformation, more cloud adoption, more workloads, more modernization from customers. And that's what drives the majority of the known AI customer acceleration, so we mentioned also we've seen continuous acceleration from customers that existed before AI and that are not majority AI businesses. And that's been pretty remarkable. Like the acceleration that we gave the numbers on the call, but the acceleration since last year has been constant and very significant.
- 支持 剔除最大客户后的其余业务已连续 5 个季度加速
Yeah. The last thing I will say because I know so it's on people's minds, is if you back out our largest customer from our growth, you get pretty much the same growth rate, as the rest of the business has been accelerating very steadily. Actually we've seen, I think now five quarters of continuous acceleration from the rest of the business. And we feel very good about the – what we see in the market. Yeah.
- 支持 新客户对同比收入增长的贡献从 Q1 约 25% 升到 Q2 约 30%；企业新客户 bookings 翻倍，新客户 ramp 更快
This 750 strong customer group includes a broad range of AI startups as it has in the past, but now also includes hyperscalers using Datadog for in-house AI labs. In Q2, this includes 31 customers spending more than $1 million annually, of which eight customers spent more than $10 million annually. We also achieved strong new logo dollar bookings with particular strength in enterprise, where new logo annualized bookings more than doubled from the year ago, and we are seeing new logos ramping faster and contributing more to revenue growth. The portion of our year-over-year revenue growth that relates to new customers was about 30% in Q2, up from 25% in Q1. Geographically, we're performing well in all regions with growth acceleration across the regions. We see particular strength in the Americas as much of the AI activity is occurring in the US as well as – in addition, we are executing strongly in LatAm.
- 中性 NRR 持平于 low-120s，GRR mid-to-high 90s (NRR 没有随收入加速进一步上升)
Geographically, we're performing well in all regions with growth acceleration across the regions. We see particular strength in the Americas as much of the AI activity is occurring in the US as well as – in addition, we are executing strongly in LatAm. Regarding retention metrics, our trailing 12 month net revenue retention percentage was in the low-120, similar to last quarter, and churn remains low with gross revenue retention in the mid to high-90s. We believe this metric highlights the mission critical nature of our platform for our customers. Now, moving on to our financial results.
- 支持 可持续性：增速仍远低于 2021 年，不会让客户连续多年消化巨额涨幅
Like we're not just going to move that with a few newer customers like this there, it's largely driven by the existing customers. And it's driven by both increases in volume because they're moving more workloads to the cloud and adoption of our newer products as they consolidate onto us. We think it's sustainable for one thing, if you compare it to what we have seen in the heady days of 2021 or the growth rates are accelerating, but they're still far below what we were seeing at that time. And so we don't create the same issue of customers having to digest very large increases multiple years in a row. I think in this case, we're very well within the realms of sustainability. And as has been a theme in this call, remember that when customers adopt and they consolidate, they have an eye towards the financial side of the equation basically, how much money are they going to make or save by doing that at the end? And we are very good at helping customers understand that and making that case and helping them save money at the end of the day. So we feel good about that. And I want to just add one thing, and we talked about this last quarter, that some of this has to do with the investments that we're making in our platform and our product, but it also has to do with the investments that we're making in our go-to-market. We've successfully expanded quota capacity, the geography of it and essentially, that's, as we talked about last quarter, providing returns.
- 支持 销售配额产能和地域扩张也在带来回报
So we feel good about that. And I want to just add one thing, and we talked about this last quarter, that some of this has to do with the investments that we're making in our platform and our product, but it also has to do with the investments that we're making in our go-to-market. We've successfully expanded quota capacity, the geography of it and essentially, that's, as we talked about last quarter, providing returns. So that's also being a growth driver in our non-AI or enterprise type business. That's right. And you see it also contributing a bit, in a bit (00:48:35).
- 口径不齐 客户总数环比仅 +200 (33,200→33,400)，分析师称 '印象中最弱'；CFO 解释为免费/付费边界的低端波动，对收入影响很小
Our gross customer additions continue to be strong and on trend line. And that's the vast majority of our revenues. We have at the very low end, the border between free and contract and that has variability, very low effect on revenues. So if you – that accounts as we talked about in many quarters, that accounts for the variability of the customer count. And it really has to do with something that has very little effect on revenues.
- 支持 核心/非 AI 用量增速从 mid-20s 升到 high-20s，来自交叉销售、上云和 AI 应用开发
* 26Q2 revenue grew 35.6% YoY, in line with buy-side expectations of 35-36%, accelerating 3ppts QoQ. However, the US$46m beat was slightly below the US$50m last quarter. * 26Q3 revenue guidance was a touch light at 29.5% at the high end, where expectations were looking for a 3-handle. Applying a typical 5ppts beat would imply a marginal sequential growth deceleration in 26Q3. * **The miss appears mostly driven by usage optimization at OpenAI, DDOG’s largest customer. For 26Q2, we estimated a 20% QoQ decline in OpenAI’s spending, on which market consensus appeared too optimistic. Management confirmed it has renewed a 9-figure contract with OpenAI, but the latter’s usage will decline sequentially in 26Q3. We believe Chronosphere is gaining share within the OpenAI platform, as it prices 30-40% lower than DDOG.** * Core/non-AI usage growth continued to accelerate to high-20s from mid-20s, driven by strong cross-sell momentum and increasing usage from cloud migration and AI app development. * Full-year revenue guidance was raised by more than the 26Q2 beat flowed through, to 30-31% YoY, in line with expectations. * The margin trend remains strong — GPM dipped below 80%, but OPM beat expectations on operating leverage.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 支持 非 AI 增速从一年前约 18% 升到 high-20s，原因是 AI 推动上云、现代化、新应用
Barclays says CEO/CFO meetings reinforced that DDOG’s acceleration is broader than AI-native customers, with non-AI growth improving from roughly 18% a year ago to the high-20% range as AI drives cloud migration, infrastructure modernization and new application development. Barclays notes DDOG intentionally de-risked 3Q guidance by assuming its largest AI customer contributes only at contractual minimums, while broader usage and pipeline trends remain strong. Barclays also says GPU monitoring, agent observability, Bits AI, MCP and AI security are gaining early traction, but remain in the early innings of monetization and represent a multi-year growth opportunity.
来源背景：TMT Breakout (TMTB) 日报，转述 Barclays 与 DDOG CEO/CFO 财报后 fireside 会面的报告摘要。二手转述，看不到底稿。 - 支持 非 AI 业务连续 4-5 季加速至 upper-20s，季度环比增收 $115M
And is that driven by AI adoption? Yeah. So, this has been four or five quarters, as you mentioned, of acceleration up into the upper-20s, and that's enterprises and stuff. That's the adoption of the Datadog platform, that's winning market share. For instance, in the last quarter I think we said sequentially we grew $115 million of revenues. So, if you just want to on a quarterly basis, think of that as over $400 million of business. And you look at that versus the competitors, you'll see that the market share gains are very substantial.
- 支持 NRR 在 low-120s，长期稳定
Our cohorts are very, very long, meaning our cohorts signed five years ago are still expanding. The reason they can with us is they consolidate on Datadog and part of it is that we've expanded the product line so much. So, that's kind of what happens on an ongoing basis that produces the net retention that we talked about in the 120s, low-120s. And that's been pretty persistent over a long period of time. That's something very important to look out, like you said.
- 支持 老客户群 (5 年前签的) 仍在扩张，靠合同到期时整合
We sell on a credit basis, we sell $2 million of capacity and they can use the platform. So, what we find when we get that is we find that adoption. Our cohorts are very, very long, meaning our cohorts signed five years ago are still expanding. The reason they can with us is they consolidate on Datadog and part of it is that we've expanded the product line so much. So, that's kind of what happens on an ongoing basis that produces the net retention that we talked about in the 120s, low-120s.
- 支持 专家称 Datadog/Dynatrace 增速超过 20%
Um, and there needs to be a lot more, um, I think investment, um, so that customers can see that value. Um, but I think from an observability perspective, you know, there's there's quite a lot of competition. Um, but Datadog and Dynatrace are actually doing really, really well. You know, they're definitely growing over 20%. It's very difficult to compete with them, very, very difficult. Um, but it it, you know, it it really depends on what value Snowflake brings.
来源背景：未具名专家，从'我们这边的客户''我们围绕 Snowflake 的可观测性平台开展了大量市场推广培训'推测为 Snowflake 内部 GTM 人员或紧密合作伙伴 (未明示)。视角局限：站在 Snowflake 一方；只看自己客户，未必了解 Datadog 内部数据。 - 支持 10 万美元以上客户预计用 4–6 个产品，100 万美元客户用 8–12 个以上，扩展空间大
**EE** [00:06:02]: First of all, I think that alone is a very strong evidence for the power of the platform. From my experience and based on my memory, larger customers or customers that tend to spend more also adopt more products. It's not like they're just spending all this USD 100,000 on infrastructure monitoring, just one product. I think that's a good evidence for the strength of the platform and the adoption. Again, it's not like the USD 1m customers who probably already use even eight, 10, 12 products at minimum. I think with this customer, there's still a lot of upside. I assume they use maybe between four and six products. There's definitely more upside because we are seeing customers expand over time. When I was there, we were seeing that expansion continues. We can talk also about the go-to-market engine of Datadog and how it works, but knowing how it works, I think there's still a lot of upside there to growing that customer base for sure.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 口径不齐 短期客户可能把预算转向 GPU/API，Datadog 过去 5–6 年已经历过几次类似波动
**EE** [00:37:29]: I think potentially. I think we might see a wave. I mean there's two types of waves. The one you're describing. The second one is maybe just like, "Okay, observability costs are just going to become super high, and we'll have to manage that." I think in general, companies are allocating more money in budgets today for AI-specific initiatives. We see it also in security budgets, security can tap into other AI-related budgets. Potentially, yes, but I think long term, companies will still realise that observability is needed to run these systems properly, especially as they're not predictable. I don't know if there's no human in the loop or I think there's still going to be a big need to understand what these systems are doing. Maybe short term, yes, companies will be a bit more cautious. We've seen these types of waves, I think, with Datadog a couple of times in the last 5-6 years.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 支持 预计 NRR 维持在 125–130% 左右 (高于分析师引用的 '低 120%')
**EE** [00:52:46]: Again, I think there are a lot of things to look at here. I know post-COVID, there was an increase in NRR. For the longest time, Datadog was also at 130%-plus. I think it was 18 consecutive quarters or something like that, if I remember correctly. It was very healthy for a long time. I assume if it went down, it was because of, again, more macroeconomic things or bigger trends because for the most part, customers are very, very happy with Datadog. There's always going to be some that are more cost conscious or have this issue or that issue, but for the most part, the platform is very sticky and loveable by customers. I definitely see it staying around those numbers, probably 125-130%, but I think, again, the NRR is just one signal. I think that should be always looked at together with the multi-product adoption. If companies are growing from two products to four products to six and eight, if those numbers are growing, that's another important signal to correlate that with and just customer growth in general. If cross-product adoption continues to increase, then that's… It's not just about the usage level expansion or the NRR, it's also the platform expansion, which is critical, again, to retain those customers and get them to pay more.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 支持 新客户贡献占比还可能继续上升，潜在组织约 20 万家
**EE** [00:49:53]: I think there's definitely more potential to land. Potentially that percentage could also increase. I think, again, even with 35,000, I don't know how many customers, there are probably 200,000 organisations out there. There's a lot of competitors that I know Datadog is going after explicitly and see a lot of opportunity there as well. I definitely see still a lot of opportunity in terms of new logos and also with a lot of obviously, new companies being built these days much faster.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 支持 Q2 2026 收入 11.2 亿美元，同比 +36%，>10 万美元客户 4,720 家，FCF 2.79 亿美元
Agent 运行以后，每一步调用、延迟、错误、重试和工具结果都需要被记录。Q2 2026 Datadog 收入 11.2 亿美元，同比增长 36%；年化收入超过 10 万美元的客户达到 4,720 家，自由现金流 2.79 亿美元。公司同时推出 Bits Code、Bits Chat、Agent Builder、自治 Bits AI 与 AI Guard。[5](https://investors.datadoghq.com/news-releases/news-release-details/datadog-announces-second-quarter-2026-financial-results/)公开电话会转录显示，31
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 - 支持 非 AI 客户增速一年内从约 18% 加速到 high-20s
What's especially exciting to us is that this acceleration happens throughout the customer base. So, it's not just a matter of AI customers growing very fast, I mean, of course, they are, but the rest of the business, the non-AI customers, all the companies that existed beforehand and that are not primarily the business of AI, have been accelerating over the past year or two. I think we said on the call we accelerated since one – from one year ago from, I think, 18% year-over-year growth for that part of the business to the high-20s. So, massive acceleration there. And what this really shows is that the AI transformation is happening.
- 支持 渗透率低：约半数财富 500 强是客户，平均年合同约 $400K
So, I mean, first, I would say, we're – we have so much more penetration to be had in the market than we currently have. So, I mentioned earlier, even though we're the leader, we have 13%, 14% market share. If you look at the largest customers, we are in about half of the Fortune 500. And I think the average contract – annualized contract is $400,000 or something like that. It's a little bit higher than that. So, I think we can do so much more, like we can grow so much more with all of those customers. And so, there's a – again, it doesn't take AI, it doesn't take imagination to see the huge amount of growth in the 5x or even 10x we can add on top of that.
- 支持 非 AI 部分约占 85% 业务
I said earlier, I started the keynote with the fact that we've seen massive acceleration of the non-AI customers, and all of that is happening on the back of those infrastructure build-outs and those capabilities that are being shipped by the AI natives. And so, we see it's not an isolated benefit to one part of the – or an isolated runaway investment from one part of the ecosystem. We see it throughout the ecosystem with companies in the other 85% of our business that are very, very careful about the cost equation and making sure that they build sustainable businesses. Along those lines, this notion and idea of price deflation, I think you've been generally very candid about the fact that you're going to get more volumes from your customers in providing the single source of truth and eyes on glass on the health of their environments. Well, the more volumes you're going to see, there's going to be more price deflationary impacts.
- 支持 land-and-expand：1 个产品到 2 个产品的扩展动作显著，统一平台有黏性 (未给数字)
JC: That approach generally, I don't know off the top of my head the percentage-wise or numbers here, but there is a significant adoption motion, if you will, of 1 product to 2 products, so on and so forth over time. So I think that demonstrates that self-integration, if you will, is powerful and makes sense.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 支持 上云比例只有 upper 20s-30%，AI 提高现代化紧迫性，是非 AI 加速的原因之一
I think that's very true. So, despite the fact that cloud has been around, there are – one, there are so much legacy infrastructure. The research companies say that somewhere upper 20%, 30% of applications are in the cloud. And there are tons of enterprises that, you'd be shocked, are so immature, meaning, they really haven't even started their first material projects. But the fact that anytime there's been change of technology, of which AI is a major one, the impetus or the urgency of both modernizing applications and putting them in the cloud is enhanced. And we're seeing that. That's some of the reasons why we're seeing the non-AI business accelerate. At the same time, it's not going to happen by itself. Our bottoms up selling that's very effective has to increasingly be complemented by top down and between that.
- 支持 key accounts：一个销售只盯 1-2 家 IT 预算数千万的大企业
Our commission plans were done that way. But I think what we've learned is, some of these enterprises, this is a very long sales cycle, and you can't have an enterprise salesperson covering 10 accounts. They're covering one or two accounts where the IT budget and conversion in the cloud can be in the tens of millions. And we stick this person on that account, and they're going after all the projects there. And this is the evolution.
- 反对 Bernstein：4Q 非 AI AWS 增速可能再弱约 100bp，传统负载走弱
Bernstein says AWS web metrics remain soft through early September, though Q3 should still show ~50bps of ex-AI growth acceleration given the strong start to the quarter. Bernstein sees more risk in Q4, where non-AI AWS growth could face an additional ~100bps headwind versus current estimates as traditional workloads weaken. Bernstein says accelerating OpenAI/Anthropic consumption should cushion AWS and DDOG, while lower-AI-exposure usage names such as TWLO and NET look more vulnerable.
来源背景：TMT Breakout (TMTB) 日报，转述 Bernstein (AWS 网站流量数据) 与 Citi (DT) 报告摘要。二手转述，看不到底稿。 - 支持 选 Datadog 的多是想在云上站稳脚跟的传统企业 (企业上云驱动)
MC: I would say more of the legacy companies are Datadog customers or are drawn to Datadog. And I think right now the shift has been towards more legacy companies that are... and I wouldn't say... that are trying to maintain their cloud presence. I think the shift... they still support tons of cloud-native businesses, but also the shift towards kind of repurposing your existing gear, et cetera, is a big something big. And a lot of folks are keeping their data centers, and I think it was always a kind of a misnomer that people would get rid of their data centers entirely. So I think that is something that's important. I think even with Dynatrace though, their ability to be deployed alongside some of their current infrastructure, I think that's also kind of a big driver for them. But I think with Datadog and where they're, how far up-market they are, I think they're probably more inclined to... the customers to pick Datadog are going to be those more legacy companies that are trying to get a foothold in the cloud. And then the ones that are more apt for an open source tool are ones that have felt that they've been spurned by some of these other larger companies, whether it's Splunk or whether it's Dynatrace or Datadog, and are now looking to have more of an open source kind
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 支持 IR：非 AI 原生客户也在加速，'为 AI 做准备' 的投资进入生产
Truist says meetings with DDOG IR reinforced that guidance remains conservative around its largest AI-native customer, with 4Q assumptions based on contractual minimum usage even though 3Q is tracking above that floor. Truist says IR described growth as broad-based across products, cohorts and verticals, with non-AI-native customers also accelerating as “getting-ready-for-AI” investments move into production. Truist also sees security as an underappreciated growth leg after crossing $100M ARR, with SIEM leading adoption and security still <2% of spend among >$1M ARR customers, leaving substantial expansion runway.
来源背景：TMT Breakout (TMTB) 日报，转述 Truist 与 DDOG IR 会面后的报告摘要。二手转述，看不到底稿。 - 支持 业内共识：生产中重要的应用负载只有约 20% 被观测厂商监控，80% 未渗透
Bernd Harzog: Well, look, I think that, well, let me back up and give you the top level perspective here. I think the consensus number on the part of everyone is that something on the order of 20% of the application workloads running in production that are running things that you care about are, in fact, monitored by an observability vendor. In other words, from the perspective of a vendor, like a Datadog or a Dynatrace, the vendors have penetrated 20% of their market. And 80% remains unpenetrated.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 支持 AI 提高企业利润，预算不再是过去 10 年 '要花新钱先省旧钱' 的零和环境
Bernd Harzog: And spending it on making sure that you are effectively deploying AI without doing harm seems like a reasonable thing to spend some of the extra money on. So it feels to me like we're not dealing with the flat budget problem that we had been dealing with for the last 10 or so years, where if you wanted a big company to spend more on something, you have to tell them, show them how to spend less on something else.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 支持 Jefferies：核心 ex-AI 原生增速 2Q 加速到 high-20s，AI 原生客户 750+
Jefferies says DDOG core growth ex-AI natives accelerated to high-20s in 2Q26 and 750+ AI-native customers position it well for agentic observability. But its largest customer is expected to reduce usage from 3Q, creating a 2027 concentration/optimization overhang, while ~19x CY27 revenue leaves less attractive risk/reward. Hold, PT $300 from $260
来源背景：TMT Breakout (TMTB) 日报，转述 Jefferies 报告摘要 (DDOG 与 DT)。二手转述，看不到底稿。
E02 AI 客户分散、留得住吗
判断：一半成立。最大客户之外，AI 原生客户确实在变多、变大、变分散；但最大客户走的正好是我们担心的路，而 Datadog 拿下超大厂 AI 实验室的理由是"赶时间"，不是结构性壁垒。
分散的一面：AI 原生客户从约 500 家 (2025 Q3) 增到 750 家 (2026 Q2)；年付超 $100 万的从 15 家增到 31 家，超 $1,000 万的从 5 家增到 8 家 (Q2 起这个群体里加入了用 Datadog 做内部 AI 实验室的超大厂，口径变宽了)。前十大 AI 公司全部是客户。2026 年 Q4 签了"最大的基础模型公司之一"(市场推测是 Anthropic) 八位数年化新单，是 Datadog 史上最大的新客户；Q1 又签了两家全球最大科技公司的 AI 研究部门，用于训练和 GPU 监控。
集中的一面：2025 年 Q3 AI 原生占收入 12%，之后不再披露。最大客户 (公司不点名) 的轨迹：2025 年 Q3 提前续约，"用更高的承诺换更好的经济条件"；同期把日志迁到基于 ClickHouse 的自研方案 (MeritForum 调研)；2026 年 Q2 签九位数续约，用 17 个产品，但 Q3 起减量；指标存储迁到 Chronosphere。Funda 估计它占 Datadog 2026 年收入 7–8%，2027 年降到低个位数。按这个估计，2027 年约有 $1.5–2.5 亿的收入缺口要靠其他业务补，相当于拖累增速 4–5 个点。指引已经按合同最低用量算了 (IR 说 Q3 实际跟踪高于这个底线)。
为什么会自建：独立顾问 Bernd Harzog 的解释最清楚：能自建的公司有两个特征，规模很大，但应用很少 (一个业务、一个应用、一种架构)。这正是 AI 实验室的样子。一般企业有几百上千个应用，自建划不来。CEO 被问超大厂为什么不自建，回答是"AI 竞赛的紧迫性让人想清楚什么是核心"。这是一个依赖竞赛节奏的理由。Funda 的渠道专家还提到，Anthropic 底层存储也用 ClickHouse。
And our new logo average land size also set a record and more than doubled year-over-year as we continue to land larger deals. Revenue growth accelerated with our broad base of customers, excluding the AI natives to mid-20s percent year-over-year, up from 23% last quarter and 19% in the year-ago quarter. We saw robust growth across our customer base with broad-based strength across customer size, spending bands and industries. Meanwhile, our AI native customer growth continues to significantly outpace the rest of the business. This group continues to diversify and grow, including 22 customers spending more than $1 million annually and five spending more than $10 million annually. This group includes the leading companies and foundational models, code-gen tools, and vertical specific AI solutions. Next, regarding our retention metrics, our trailing 12-month net revenue retention percentage was in the low 120% percent, up from about 120 last quarter, and our trailing 12-month gross retention percentage remains in the mid-to-high 90s. Now moving on to our financial results.
Yes. There are 3 questions in there. The first is overall on guidance, except what we're going to speak about next, we took the same approach as we looked at the organic growth rates and the attach rates and then the logo accumulation rates and discounted that. So for the overall business, which is quite diversified, we talked about diversification by industry, by geography, by SMB, mid-market and enterprise, we took the same approach. We noted that with the guidance being 18% to 20% and the non-AI or heavily diversified business being 20% plus, that would imply that the growth rate of that core business assumed in the guidance is higher than the growth rate of the large customer. It doesn't mean the large customer is growing any which way. It's just that in our consumption model, we essentially don't control that. And so, we made a very conservative assumption there. And the last point, I think you mentioned is the highly diversified. We said 650 names in the AI cohort is quite diversified, essentially would be very similar to our overall business, which we have a range of customers, but not the concentration level. And what we're seeing there is significant growth. But like our overall distributed customer base, a growth and then potentially some working on how the product is being used, but nothing out of the ordinary relative to the overall customer base in the very diversified AI set of customers outside the largest customer. Hopefully that's helpful.
This customer chose to standardize on Datadog across its business, displacing four commercial and internal tools. Datadog also proved value beyond core observability with product analytics, CI visibility, data observability and cloud cost management. This deal includes our largest win to-date for Bring Your Own Cloud, displacing their legacy commercial logging tool at a petabyte scale. And finally, we signed a nine figure renewal with a leading AI company. This long time, very large customer uses 17 Datadog products to enable unified visibility on production workloads at a very large scale, albeit with a usage reduction starting in Q3, which we considered in our guidance and which David will speak to. Before I turn it over to David for a financial review, let me offer a few words on our longer-term outlook. There is no change to our overall view that digital transformation and cloud migration are long-term secular growth drivers for our business. But we now have an additional growth driver with AI as we help our customers deliver value with this transformative new technology.
We've always used those inputs and looked at the commitment and the usage of doing that. Yeah. (00:31:39) I say is. In this case, we did choose to fully de-risk our largest customer. And the reason for that is we don't want that to be an overhang on what is otherwise a business that is accelerating and performing extremely well. So we have the same overall conservatism as we always do when we look at our numbers, but in this case we also weighted (00:32:04) this one a little bit differently. Okay, perfect. That's very clear.
* 在上个季度的调研中我们提到，尽管DDOG已经在3Q将Log management迁移到自研的基于Clickhouse的OLog上，随着mid-training的兴起，DDOG对于LLM researcher重要性大幅提升，使得OpenAI与DDOG完成续约，并且2026年用量仍可能实现增长 * **然而我们最新的调研显示，Chronosphere在metrics上的功能持续提升，因此最近在OpenAI的内部用量也在快速上升，从而导致OpenAI对DDOG的依赖有所下降。**我们预计metrics在4Q25占OpenAI DDOG ARR的40%，该部分可能在2026年中完全迁移到Chronosphere中。余下的trace部分仍将继续在DDOG上使用。**总体而言，我们预计OpenAI DDOG ARR到2026年底将下滑20%以上。** * 另一方面，DDOG其他AI native客户的用量在快速提升，将弥补部分OpenAI用量的损失。其中Anthropic在小范围部署database monitoring及infra monitoring * **综合考虑，我们初步认为DDOG 2026收入增速有望达到24-25%，因此首次给出的指引可能落在18-19%yoy，符合买方市场预期**，低于卖方一致预期的21%。在目前市场对软件的悲观看法下，不够出彩的指引仍难以改善市场情绪
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。
A:My understanding is that OpenAI is going to use Chronosphere privately for infrastructure-that is basically metrics-because, for logging, they have their own log solution. Infrastructure will be with Chronosphere. That transition, and again this is all industry chatter, should happen sometime around Q2-this is my take based on what I am hearing. Maybe even Q1, but that looks a little aggressive.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。
Bernd Harzog: That would be 1 dimension. And the other dimension would be, and this is true for the ones that do it, they can be very large, but they have relatively few applications. So think of a customer that is in 1 business with 1 application. Right. Okay, that 1 application generates 100% of the revenue. Okay. So we have 1 application here. Maybe we use 3 or 4 different languages in that 1 application. It runs in 1 place. Let's say it runs in the cloud, right? It has 1 principal architecture. Let's say it's microservices. Okay.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。
Yeah. I was just going to say it, but I guess, the point is, you've emphasized that those largest customers have been able to go in-house on some other things. Is there something unique about AI that prevents them from doing that here? Well, I think the urgency of their development efforts focuses the minds. That's what I would put it. I would say, it forces you to figure out what's core and what's not core and who you want to get to, what you need to do to maximize your chances of success. And again, it's the same thinking all of our customers have all the time. I think the equation for hyperscalers has often been fairly different because they have basically unlimited access to staffing. And, they could sort of set their own time, their own time horizons for the developments they wanted to make. I think the situation is a little bit different with the AI race, maybe. Thank you. Our next question is coming from line of Gregg Moskowitz of Mizuho. Please go ahead.
What's new is we now have a product for GPU monitoring. It's a very new product. And we see the hyperscalers that are coming to us for training workloads in particular, being very interested in that. So, again, it's too early in the product lifecycle and the customer lifecycle for these specific customers, to call definitive victory there. But we see that as a very encouraging sign of where the market might go in the future, because we think this might be a bellwether of what, the next 10, 100, 500 companies that are going to start training workloads or going to want to do. We have some signs that go beyond the customers we signed this quarter that point that way too. Thank you so much.
So and what do you think this number will look like by the end of this year? And for the other customers of Datadog, like the other AI native companies or Antropic or enterprise companies, do you think they will do similar things so that they will reduce their, you know, revenue uh on Datadog? So, uh good question. So, first of all, for OpenAI, the revenue contribution is in the low, uh low to high single digits, right? So it's uh for a company of our scale, it's somewhat concentrated revenue, but it's not as if uh the contribution is massive, right? So even if we assume that OpenAI is going to completely replace Datadog, uh that is not going to have that meaningful a hit on the company's overall revenue numbers, especially given that the AI cohort is the one that's actually witnessing much more uh uh accelerated growth versus the rest of the cohort, right, in terms of traditional companies. So even if let's say OpenAI uh is hypothetically doing it, uh the the the growth from a lot of the other uh other uh AI native companies is actually going to uh counteract that. And I think OpenAI especially given their loss of market share in some of the... key uh markets that they're operating in, such as the uh business to consumer chat agent and also the coding agents etc.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。
展开本节全部 118 条原文 (支持 55 / 反对 35 / 其他 28，按时间)
- 中性 AI 原生客户占收入 12%，Q2 为 11%，一年前约 6%
In addition, we now have more larger AI customers, including 15 of them spending $1 million or more annually with Datadog, and about 100 spending more than $100,000 annually. Year-over-year revenue growth from our AI native customers, excluding the largest customer, again, accelerated in Q3. In Q3, this group represented 12% of our revenue, up from 11% last quarter and about 6% in the year ago quarter. I will note that over time, we think this metric will become less relevant as AI usage in production broadens beyond this group of customers. Our year-over-year revenue growth also accelerated amongst our non-AI native customers.
- 支持 AI 原生客户超过 500 家，约 100 家年付 >$10 万，15 家以上 >$100 万
This company has been a long-time Datadog customer and has expanded their usage of multiple products, securing better economics for a higher commitment with an early renewal. Speaking of AI customers, we continue to help native customers big and small to grow and scale their businesses. And we continue to see this group broaden in number and size, with more than 500 AI- native companies in this group, about 100 of which are spending more than $100,000 annually with Datadog and more than 15 who are spending more than $1 million annually with us. While we know there's a lot of attention on this cohort, we primarily see it as an indication of what's to come as companies of every size and every single industry incorporate AI into their cloud applications. And that's it for another very strong quarter from our go-to-market teams, who ran our very hard at work as we have a really exciting pipeline for Q4.
- 支持 剔除最大客户后，AI 原生客户增速再度加速
As Olivier discussed, we extended the contract of our largest AI native customer. In addition, we now have more larger AI customers, including 15 of them spending $1 million or more annually with Datadog, and about 100 spending more than $100,000 annually. Year-over-year revenue growth from our AI native customers, excluding the largest customer, again, accelerated in Q3. In Q3, this group represented 12% of our revenue, up from 11% last quarter and about 6% in the year ago quarter. I will note that over time, we think this metric will become less relevant as AI usage in production broadens beyond this group of customers.
- 口径不齐 最大 AI 客户 9 位数年化扩张 + 提前续约，但条款是"更高承诺换更好经济性"
This customer developing Datadog products faster than previously expected and segment support higher usage. With this expansion, they will adopt Datadog Incident Management and OnCall solution company-wide for a total of 5,000 users who support operational continuity across the business. Finally, we signed a 9-figure annualized expansion with a leading AI company. This company has been a long-time Datadog customer and has expanded their usage of multiple products, securing better economics for a higher commitment with an early renewal. Speaking of AI customers, we continue to help native customers big and small to grow and scale their businesses. And we continue to see this group broaden in number and size, with more than 500 AI- native companies in this group, about 100 of which are spending more than $100,000 annually with Datadog and more than 15 who are spending more than $1 million annually with us.
- 中性 CFO 拒绝给最大客户未来几个季度的贡献变化
No. I mean we don't provide that kind of information on individual customers. We're trying to give a picture of the overall business. Generally, I think as Oli mentioned, on our larger customers, we have a motion of the expansion of volume and then we talk when we work on the term and the volume-based pricing, but we don't give guidance like that on individual customers.
- 中性 最大客户合同期限：只说"年度+承诺"的典型路径，不确认是 1 年还是多年；称 RPO 增长并非来自该合同
Yes. On the first point, I think we give a lot of examples and our motion, which our customers would be following, including that one would be -- we fix out annual plus commits. We're not commenting on individual contracts here, but it would follow a typical path to other types of contracts. So that's what we would do.
- 中性 Q2'25 AI 原生客户约占收入 11%
* 披露了AI原生客户贡献了约 11% 的收入（Q2’25）。 * Datadog不以 RPO 作为主要披露指标；更多披露 大客户 ARR 数量（如 ARR>$100k 客户数）来体现扩张。
来源背景：Acecamp 作者整理的卖方报告和买方讨论 (不是专家访谈)；涉及 DDOG 的是转述的公开数字和高盛的观点。 - 支持 除 OpenAI 外，AI 原生客户没有减量
A:Yes, that is still the case. It is not that any of the AI-natives have necessarily stopped spending. OpenAI is a different beast altogether, but the AI-native cohort remains one of the biggest revenue-generating sources. Despite developments concerning OpenAI, we have not seen any pullback from the rest of the AI-native cohort.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 反对 OpenAI 的基础设施 metrics 迁往 Chronosphere，日志用自有方案，预计约 2026Q2 迁完
A:My understanding is that OpenAI is going to use Chronosphere privately for infrastructure-that is basically metrics-because, for logging, they have their own log solution. Infrastructure will be with Chronosphere. That transition, and again this is all industry chatter, should happen sometime around Q2-this is my take based on what I am hearing. Maybe even Q1, but that looks a little aggressive.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 支持 他预计 Anthropic 用量会增加，但这只是他的预期
That partnership is not there with OpenAI; they ended it. I do anticipate an increase in usage by Anthropic, but there is no information other than that-that is just my anticipation.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 反对 3Q 给 AI 原生公司做部署的一次性贡献在 4Q 大幅减少
3Q帮助部分AI native公司做了部署的一次性贡献在4Q大幅度减少
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 反对 OpenAI 已在 3Q 把日志管理迁到自研的、基于 ClickHouse 的方案
* 在上个季度的调研中我们提到，尽管DDOG已经在3Q将Log management迁移到自研的基于Clickhouse的OLog上，随着mid-training的兴起，DDOG对于LLM researcher重要性大幅提升，使得OpenAI与DDOG完成续约，并且2026年用量仍可能实现增长 * **然而我们最新的调研显示，Chronosphere在metrics上的功能持续提升，因此最近在OpenAI的内部用量也在快速上升，从而导致OpenAI对DDOG的依赖有所下降。**我们预计metrics在4Q25占OpenAI DDOG ARR的40%，该部分可能在2026年中完全迁移到Chronosphere中。余下的trace部分仍将继续在DDOG上使用。**总体而言，我们预计OpenAI DDOG ARR到2026年底将下滑20%以上。** * 另一方面，DDOG其他AI native客户的用量在快速提升，将弥补部分OpenAI用量的损失。其中Anthropic在小范围部署database monitoring及infra monitoring * **综合考虑，我们初步认为DDOG 2026收入增速有望达到24-25%，因此首次给出的指引可能落在18-19%yoy，符合买方市场预期**，低于卖方一致预期的21%。在目前市场对软件的悲观看法下，不够出彩的指引仍难以改善市场情绪
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 反对 作者估计 metrics 约占 OpenAI DDOG ARR 的 40%，2026 年中迁到 Chronosphere；OpenAI 的 DDOG ARR 到 2026 年底降 20% 以上
* 在上个季度的调研中我们提到，尽管DDOG已经在3Q将Log management迁移到自研的基于Clickhouse的OLog上，随着mid-training的兴起，DDOG对于LLM researcher重要性大幅提升，使得OpenAI与DDOG完成续约，并且2026年用量仍可能实现增长 * **然而我们最新的调研显示，Chronosphere在metrics上的功能持续提升，因此最近在OpenAI的内部用量也在快速上升，从而导致OpenAI对DDOG的依赖有所下降。**我们预计metrics在4Q25占OpenAI DDOG ARR的40%，该部分可能在2026年中完全迁移到Chronosphere中。余下的trace部分仍将继续在DDOG上使用。**总体而言，我们预计OpenAI DDOG ARR到2026年底将下滑20%以上。** * 另一方面，DDOG其他AI native客户的用量在快速提升，将弥补部分OpenAI用量的损失。其中Anthropic在小范围部署database monitoring及infra monitoring * **综合考虑，我们初步认为DDOG 2026收入增速有望达到24-25%，因此首次给出的指引可能落在18-19%yoy，符合买方市场预期**，低于卖方一致预期的21%。在目前市场对软件的悲观看法下，不够出彩的指引仍难以改善市场情绪
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 支持 其他 AI 原生客户用量快速提升，可部分弥补；Anthropic 小范围部署了数据库和基础设施监控
* 在上个季度的调研中我们提到，尽管DDOG已经在3Q将Log management迁移到自研的基于Clickhouse的OLog上，随着mid-training的兴起，DDOG对于LLM researcher重要性大幅提升，使得OpenAI与DDOG完成续约，并且2026年用量仍可能实现增长 * **然而我们最新的调研显示，Chronosphere在metrics上的功能持续提升，因此最近在OpenAI的内部用量也在快速上升，从而导致OpenAI对DDOG的依赖有所下降。**我们预计metrics在4Q25占OpenAI DDOG ARR的40%，该部分可能在2026年中完全迁移到Chronosphere中。余下的trace部分仍将继续在DDOG上使用。**总体而言，我们预计OpenAI DDOG ARR到2026年底将下滑20%以上。** * 另一方面，DDOG其他AI native客户的用量在快速提升，将弥补部分OpenAI用量的损失。其中Anthropic在小范围部署database monitoring及infra monitoring * **综合考虑，我们初步认为DDOG 2026收入增速有望达到24-25%，因此首次给出的指引可能落在18-19%yoy，符合买方市场预期**，低于卖方一致预期的21%。在目前市场对软件的悲观看法下，不够出彩的指引仍难以改善市场情绪
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 口径不齐 管理层不再披露 AI 原生客户占收入比例
We didn't -- have not put it in there.
- 支持 AI 原生客户约 650 家，19 家年付 >$100 万；前 20 大 AI 原生公司中 14 家是客户
And we have seen this trend of accelerated revenue growth continue in January. Meanwhile, we are seeing continued strong adoption amongst AI-native customers with growth that significantly outpaces the rest of the business. We see more AI-native customers using Datadog with about 650 customers in this group. And we are seeing these customers grow with us, including 19 customers spending $1 million or more annually with Datadog. Among our AI customers are the largest companies in this space, as today 14 of the top 20 AI-native companies are Datadog customers. Next, we also saw continued strength from new customer contribution.
- 支持 新签"最大 AI 基础模型公司之一"8 位数年化 land，是历史最大新 logo，替换 5+ 开源/商业/自研工具
So let's move on to sales and marketing. I want to highlight some of the great deals we closed this quarter. First, we landed an 8-figure annualized deal and our biggest new logo deal to date with one of the largest AI foundational model companies. This customer has a fragmented observability stack and cumbersome monitoring workflows leading to poor productivity. This is a consolidation of more than 5 open source, commercial, hyperscaler and in-house observability tools into the unified Datadog platform that has returned meaningful time to developers and has enabled a more cohesive approach to observability.
- 支持 CFO 称剔除最大客户的 AI 群体分散，集中度与整体业务类似
Yes. There are 3 questions in there. The first is overall on guidance, except what we're going to speak about next, we took the same approach as we looked at the organic growth rates and the attach rates and then the logo accumulation rates and discounted that. So for the overall business, which is quite diversified, we talked about diversification by industry, by geography, by SMB, mid-market and enterprise, we took the same approach. We noted that with the guidance being 18% to 20% and the non-AI or heavily diversified business being 20% plus, that would imply that the growth rate of that core business assumed in the guidance is higher than the growth rate of the large customer. It doesn't mean the large customer is growing any which way. It's just that in our consumption model, we essentially don't control that. And so, we made a very conservative assumption there. And the last point, I think you mentioned is the highly diversified. We said 650 names in the AI cohort is quite diversified, essentially would be very similar to our overall business, which we have a range of customers, but not the concentration level. And what we're seeing there is significant growth. But like our overall distributed customer base, a growth and then potentially some working on how the product is being used, but nothing out of the ordinary relative to the overall customer base in the very diversified AI set of customers outside the largest customer. Hopefully that's helpful.
- 反对 AI 群体中也有客户"在调整产品用法"(优化)，CFO 称不异常
Yes. There are 3 questions in there. The first is overall on guidance, except what we're going to speak about next, we took the same approach as we looked at the organic growth rates and the attach rates and then the logo accumulation rates and discounted that. So for the overall business, which is quite diversified, we talked about diversification by industry, by geography, by SMB, mid-market and enterprise, we took the same approach. We noted that with the guidance being 18% to 20% and the non-AI or heavily diversified business being 20% plus, that would imply that the growth rate of that core business assumed in the guidance is higher than the growth rate of the large customer. It doesn't mean the large customer is growing any which way. It's just that in our consumption model, we essentially don't control that. And so, we made a very conservative assumption there. And the last point, I think you mentioned is the highly diversified. We said 650 names in the AI cohort is quite diversified, essentially would be very similar to our overall business, which we have a range of customers, but not the concentration level. And what we're seeing there is significant growth. But like our overall distributed customer base, a growth and then potentially some working on how the product is being used, but nothing out of the ordinary relative to the overall customer base in the very diversified AI set of customers outside the largest customer. Hopefully that's helpful.
- 支持 CEO：业务不是靠少数客户建立的；更多超大规模 AI 实验室在接洽
Yes. I mean, look, we've never been -- we're not built as a business to be concentrated on a couple of customers. That's not how we've become successful. That's probably not how we'll be successful in the long term. So yes, I mean we -- at the end of the day, it should be irrational for customers -- for all customers in the AI cohort not to use our product. So we see -- we have some great successes with the customers currently in that cohort. We see more. By the way, we have more that are more inbound there and more customers that are talking to us from the largest even hyperscaler level AI labs. And we expect to drive more business there in the future. I think there's no question about that.
- 反对 作者推算 AI 原生 cohort 在 OpenAI 优化影响下环比只增长个位数
这意味着AI native cohort在OpenAI优化的情况下环比仅增长个位数
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 口径不齐 签下一家模型公司的八位数单子 (推测为 Anthropic)，短期填不上 OpenAI 的缺口，但有助于改善集中度
* APM产品增速达到mid-30%，体现了AI native客户们对LLM Monitoring和Real User Monitoring等产品的强劲需求，以及传统客户们AI应用逐渐上量的趋势 * 我们此前指出，OpenAI尽管完成了续约，但由于Chronosphere的使用体验提升，其对DDOG的用量在2026年仍将下降，部分抵消了core cohort的强劲增长 * 与此同时，其他AI native客户的使用进展良好；公司提到与某大模型公司签下了八位数订单（预计为Anthropic), 尽管该用量短期内无法填补OpenAI流失的空缺，但有助于客户集中度的改善
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 支持 UBS：没看到 OpenAI/Anthropic 进入可观测性领域，客户自建意愿低
The firm remains constructive ahead of 1Q26, with customer checks (covering ~88% of ARR) indicating steady spend growth (~8–15%) and no share loss despite broader AI-related budget reallocation. Core demand remains supported by hyperscaler-driven infra trends, while AI is pressuring non-core spend but not Datadog itself. UBS views DDOG as a relative AI beneficiary with limited disruption risk, noting low appetite for DIY observability and no signs of OpenAI/Anthropic moving into the space, maintaining Buy with valuation still reasonable vs. growth.
来源背景：TMT Breakout (TMTB) 日报，转述 JPM (渠道检查及一位行业联系人) 与 UBS (客户检查) 的报告摘要，二手转述，看不到底稿。 - 反对 AI 时代 cardinality 更重要，Chronosphere 在架构上优于 DDOG/DT；也有客户在试 OTel
+ Main reason for the miss: the observability industry is moving from observability 2.0 to 3.0, with customers reassessing how to deploy LLM-related telemetry in the AI / agentic era + Pre-2024, customers focused on traditional Infra / APM. From 2025, AI budgets expanded sharply, bringing entirely new telemetry / log / trace requirements; customers are now rethinking their existing observability architecture and carefully evaluating what can and cannot be consolidated + AI further raises the importance of cardinality (the ability to handle multiple metric / trace combinations simultaneously), where Chronosphere has an architectural edge over DDOG / DT; that said, it is not the sole determinant — some customers are also experimenting with OTel * Customers are increasing AI investment, but observability spend has yet to catch up
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 支持 AWS 生态 AI 初创云消耗加速，免费额度 3–6 个月就烧完，更早转为付费可观测性
+ AI-startup cloud consumption is accelerating sharply — free credits that previously lasted 1–2 years are now burned in 3–6 months, triggering paid observability adoption earlier + Since late 2024, DDOG has increased focus on new logos, signing customers as soon as monthly consumption exceeds ~US$500 and then driving rapid upsell to ~US$10–20k+ * Open-source tools such as Grafana are cheaper, but the setup and maintenance burden is high; many customers still choose DDOG once they factor in total cost
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 Funda 判断 OpenAI 到 2026 年底核心功能完全迁出，仅 RUM、Synthetics 等新产品留存
+ North America Channel Expert A: last quarter we expected OpenAI to fully migrate off DDOG by 2Q26; the latest read is that full migration will take at least until 4Q26 + **Our current view: OpenAI will fully migrate core functionality off DDOG by end-2026; that said, OpenAI continues to use a number of newer DDOG products (RUM, Synthetics Monitoring, etc.)**, where usage is still growing and could
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 Funda 估 OpenAI 占 DDOG 2026 收入 7–8%，2027 年降到低个位数
+ **We estimate OpenAI will still contribute ~7–8% of DDOG revenue in 2026 and low single-digits in 2027** + The core driver remains Chronosphere’s superior architecture at the metrics layer; however, RUM and parts of APM still require OpenAI to either build internally or stay on DDOG * **Anthropic usage is accelerating, but absolute scale remains far below OpenAI**
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 支持 Anthropic 在 DDOG 的 ARR 季增 15–20%，Funda 预计年底超 $100m
+ Former North America Sales Expert B: Anthropic’s current DDOG ARR is growing +15–20% QoQ + **We expect Anthropic’s DDOG ARR to exceed US$100m by end-2026** + Product-wise, Anthropic uses Infra / APM and Developer platform / API features, similar to OpenAI’s earlier usage
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 Anthropic 底层存储用 ClickHouse，长期依赖路径可能和 OpenAI 一样，难点在 metrics
Anthropic uses ClickHouse for underlying storage, suggesting Anthropic’s longer-term reliance trajectory may resemble OpenAI’s — with the key challenge again concentrated on the metrics side
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 支持 AI native 客户增速显著快于其余业务；>$1M 的 AI 客户 22 家、>$10M 的 5 家
Revenue growth accelerated with our broad base of customers, excluding the AI natives to mid-20s percent year-over-year, up from 23% last quarter and 19% in the year-ago quarter. We saw robust growth across our customer base with broad-based strength across customer size, spending bands and industries. Meanwhile, our AI native customer growth continues to significantly outpace the rest of the business. This group continues to diversify and grow, including 22 customers spending more than $1 million annually and five spending more than $10 million annually. This group includes the leading companies and foundational models, code-gen tools, and vertical specific AI solutions. Next, regarding our retention metrics, our trailing 12-month net revenue retention percentage was in the low 120% percent, up from about 120 last quarter, and our trailing 12-month gross retention percentage remains in the mid-to-high 90s.
- 支持 Q1 新签两家全球最大科技公司的 AI 研究部门 (一个七位数、一个八位数年化)，用于训练工作流和 GPU 监控
So our teams were hard at work again, and we're looking forward to sharing many new products, and future announcements at our DASH User Conference on June 9th, and June 10th in New York City. Now, let's move on to sales and marketing and highlight some of the deals we closed this quarter. First, we landed two large deals, a seven figure, and an eight figure annualized deals with the AI research divisions at two of the world's largest technology companies. These organizations are building and training the most advanced AI models in the world. It is critical for them to reduce engineering friction, and increase training velocity, but fragmented internal and open source tooling made it harder to identify and solve issues, and reduced engineering and research productivity.
- 支持 ARR 增量分散；剔除 Q4 签下、对 Q1 收入贡献最大的那个客户后 Q1 仍是 ARR 净增纪录
And we said we had a record. So, essentially, at the bedrock of this is sort of the run forward of ARR that we've already signed. The ARR add was very broad based and was not very concentrated. So, whereas we pointed out some very significant adds, I would say that the first quarter in that ARR add was really diversified and from lots of different places. So, I think Oli will come in here. But the confidence that we have is you're right. We essentially take what we already have, we discount the growth trends that we've seen and that produces what you exactly said, which is, whatever your assumptions are on Bits, a very impressive, sequential, really, due to what happened in Q1 and the rate of business accumulation by Datadog. Oli, do you want to add? Yeah. I mean, if you want to hold on what David just said. The adds were broad based. I mean, look, and when you look at why do we have a great Q1, we also landed great customers in Q4. We had talked about it a quarter ago. But even if you take out the customer we landed in Q4, that added the most revenue in Q1, we still had a record quarter in terms of ARRs. So, this is really broad based, and we landed a few more customers in Q1 that don't contribute any revenue yet, but we expect to be big contributors in the future. So, when you put all that together, we feel very confident about Q2, hence the numbers you've seen.
- 支持 超大规模云厂商有自建文化、资源充足，却也来用 Datadog 替换原有工具
So it's a win for everyone. The thing that's also interesting in particular this quarter is that we also landed some large parts of hyperscalers. And hyperscalers typically have a culture of building everything themselves, and, they certainly have the balance sheet and the human capital to support, some of that build up, like if there was ever a set of companies for whom it makes sense to do it themselves, that would be those companies. And yet, we see that they have the same issues, when it comes to going as fast as they can, being as efficient as they can with their resources, like they come to us, to replace some of the things they were using before. Two things, two metrics to look at that, to make the points, Oli, you're making. If you look at our platform adoption and you see both the growth of the different categories and the extension of the categories out to lots of products, that shows you that the consolidation on the Datadog platform has continued and there is a very strong trend.
- 反对 CEO 承认对 hyperscaler 训练客户 '宣布胜利' 还太早
It's a very new product. And we see the hyperscalers that are coming to us for training workloads in particular, being very interested in that. So, again, it's too early in the product lifecycle and the customer lifecycle for these specific customers, to call definitive victory there. But we see that as a very encouraging sign of where the market might go in the future, because we think this might be a bellwether of what, the next 10, 100, 500 companies that are going to start training workloads or going to want to do. We have some signs that go beyond the customers we signed this quarter that point that way too.
- 口径不齐 被问 hyperscaler 为何不自建时，CEO 的理由是 'AI 竞赛的紧迫性'，即留存依赖竞赛紧迫度而非结构性壁垒
I was just going to say it, but I guess, the point is, you've emphasized that those largest customers have been able to go in-house on some other things. Is there something unique about AI that prevents them from doing that here? Well, I think the urgency of their development efforts focuses the minds. That's what I would put it. I would say, it forces you to figure out what's core and what's not core and who you want to get to, what you need to do to maximize your chances of success. And again, it's the same thinking all of our customers have all the time. I think the equation for hyperscalers has often been fairly different because they have basically unlimited access to staffing. And, they could sort of set their own time, their own time horizons for the developments they wanted to make. I think the situation is a little bit different with the AI race, maybe. Thank you. Our next question is coming from line of Gregg Moskowitz of Mizuho.
- 支持 Funda 推算 AI 原生客户收入 QoQ 加速到 17%
+ The “core” non-AI customer cohort **grew at a mid-20% year-over-year rate, up from 23% in 4Q25** and 19% in 3Q25. + This implies **AI-native customer growth accelerated to 17% QoQ**, after growth was impacted by OpenAI’s migration out in 4Q, demonstrating Anthropic and other AI native customers’ consumption picking up. * 4Q adj OPM beat expectation by 1ppt, demonstrating strong operating leverage despite GPM being lower, likely due to the discount offered to large customers * 2Q26 revenue guidance is set at 30% year-over-year growth, surpassing market expectations by 6-7 percentage points. **Based on a normal beat cadence of 4-5 percentage points, actual growth could reach 35% year-over-year, the highest** **since 2022.** * Full-year revenue guidance is now 26-27% year-over-year growth. Factoring in **typical beat cadence, FY26 revenue is projected to accelerate to low-to-mid 30% growth.**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 口径不齐 管理层称 AI 原生增长广泛、不集中于某一客户，但 Funda 认为主要受 Anthropic 推动
* Management observes broad strength in new AI-native customer growth, without concentration in any one customer. **However, we believe the momentum is likely influenced by Anthropic’s ARR acceleration since January.**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 ARR >$1M 的 AI 原生客户 22 家 (4Q25 为 19)，其中 >$10M 5 家
+ Both Anthropic Claude Code and OpenAI Codex rely on DDOG for telemetry traces, with coding agents likely boosting significant APM/trace layer usage. * DDOG now has 22 AI-native customers with ARR over $1m (up from 19 in 4Q25), including 5 with ARR over $10m. * Most interestingly, DDOG secured two large ARR deals with AI research divisions at leading technology companies, showcasing its technology advantage and expected meaningful revenue contribution in 2H26**.**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 与领先科技公司的 AI 研究部门签下两笔大单，用于 AI 训练，超过自建方案
**This has demonstrated DDOG’s superior tech stack over competitors, as hyperscalers have adopted DDOG over in-house solutions for AI training**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 管理层称训练正成为市场，hyperscaler 尽管有自建工具也选择 DDOG
* **DDOG +30%** after their stellar beat and raise with revs reaccelearting to 32% from 29% in prior q. FY26 guide now calls for 26% ahead of street which was closer to 20%. The midpoint of FY26 was raised more than combined Q1 beat and upside to Q2 guide, implying a strong 2H and mitigating worries of growth cliff related to OpenAI. Overall, call was very bullish with mgmt saying that before they were more focused on inference. Now they are increasingly seeing training become a market as they are landing hyperscalers who are using DDOG despite having a whole suite of homegrown capabilities. As *Josh S* put it in TMTB Slack: *feels like this call was pretty much as bullish as u could hope for. only thing they could've done more was maybe just spend 5 minutes saying "anthropic" 50x*
来源背景：TMT Breakout (TMTB) 日报，作者市场评论 + TMTB Slack 成员 Josh S 的评论，转述业绩会要点。二手转述，看不到底稿。 - 支持 前 8-10 家 AI coding 公司几乎全部用 Datadog
Did you see that dynamic? Yes, definitely. And, look, (00:08:01) most of the – actually, pretty much all of the top 8 or 10 coding companies or (00:08:08) companies, whether you're talking about the models people use in production in large companies or the products that are built more for the consumers, all of these companies use Datadog behind the scenes. Yeah. And so that tells you about the very specific need there is in there.
- 中性 超大厂用 Datadog 是“背书”，但管理层自己说超大厂太少、善变，不是商业模式
So what are we doing? Which are we doing there? So, again, our business model in the end is not to serve the hyperscalers. There's not enough of them. They are too fickle or so for cultural reasons, as I mentioned earlier, but for us it's a great proof point. Even the companies that have virtually unlimited access to top talent and a strong cultural bias for building use our product. It means that it doesn't make sense for anybody else to build their own.
- 支持 hyperscaler 都用 DDOG，说明自建对任何人都不划算
*“If even the companies that have virtually unlimited access to top talent and a strong cultural bias for building use our product, it means it really doesn’t make sense for anybody else to build their own.”*
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 支持 前 8–10 家 coding/vibe coding 公司都在幕后使用 DDOG
Pretty much all of the top eight or ten coding companies or vibe coding companies, whether you’re talking about the models people use in production or the products built more for consumers — all of those companies use Datadog behind the scenes.
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 支持 超大厂文化上什么都自建，也开始买 Datadog (AI 开发/训练)
I'll give you a few more examples that we've had more recently. So we mentioned on our last earnings call that we also had a number of hyperscalers setting out optimal products in their case for helping with the development and training of their AI models. And these are companies that culturally, basically don't use any commercial software, like, they write everything themselves. They use open source. They're completely homegrown by nature.
- 中性 最大 AI 实验室可无限用推理、面对“垂直需求墙”，其做法不一定能推广到其他客户
So, most of our business is with large enterprises. But we also serve a lot of startups in AI, whether they are the super large-scale model makers or the companies one level below that are largely AI native and building on AI. So, when you look at the largest, it's interesting to see who they work, but it might not be completely transferable to the rest of the market because those companies have unlimited access to inference, for one thing. And also, for a number of them, like almost vertical wall of demand in front of them. They're not in a situation that most other companies are in. And so, they might make different choices and the way they operate might not carry across. When you look at the tier below that, all of the companies that are, I mean, still very successful growing fast in AI, but are still buying most of their inference from somebody else and have to make some hard prioritization call in terms of would they walk on or not, and have more reactive – more ratable growth for the rest of the market.
- 中性 模型厂运维超大 GPU 集群，关注结构化指标和单一格式日志
主要原因是模型厂商在可观测性方面的需求比较独特。可观测性通常分为三个部分：前端追踪、RUM，以及底层基础设施的指标（metric）和日志。模型厂商运维的是超大规模GPU集群，他们更关注集群产生的数据指标，都是结构化数据和单一格式的日志，这个 不同于传统操作系统日志记录每个人的操作。
来源背景：未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。 - 反对 模型厂选 ClickHouse 是因为它是列存、适合结构化和指标型数据
ClickHouse本身是列存平台，数据类型偏向结构化和指标型，这也是模型厂商选择这些产品的主要原因，Elastic并不是传统意义上的日志客户。如果未来有合规、检测等需求，或企业内部用户行为检测、比如OpenAI自己的内部装了EDR系统后需要做安全日志分析，或者其他外部的系统需要做日志分析的，那才可能会用Elastic的产品。但AI native企业如模型厂商，属于新型企业，不像传统企业有历史遗留的这种像CRM或采购过来的系统，他自己没办法做运维，所以需要保留大量日志，然后他们可能有上百套系统，可能受监管要求必须存储数据，这个时候就需要搜索引擎做异常检测分析，这时就会用到Elastic。传统企业内部有很多系统，产生不同日志，都是非结构化数据，这些是Elastic或Splunk的目标客户。
来源背景：未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。 - 口径不齐 Datadog 在模型厂主要用于用户体验监控，Anthropic 签的是 RUM 部分
如果是互联网企业，内部系统自建，或者他是做硬件然后仅提供单一的SaaS服务，像Anthropic和阿里的服务是不一样的，Anthropic只提供模型服务，而不像互联网企业有多种服务和复杂的日志检测需求，这时会考虑Elastic、Splunk 等产品。如果业务单一、主要是指标监控，需求是成本低、速度快，ClickHouse更适合做底层基础设施。Datadog主要用于用户体验监控，Anthropic和Datadog签的是可观测性里的RUM部分。另外一个比较好的是Datadog本身是 纯SaaS服务，他们前端监控也做得不错。
来源背景：未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。 - 支持 Datadog 的 RUM 覆盖全链路，对 C 端模型服务很重要
例如OpenAI提供的面向C端用户的服务，他会手机每个用户使用的性能体验数据，用户体验对他们非常重要。例如，客户访问ChatGPT时，响应时间是否足够快，推理时间如何分配，比如30秒推理中多少时间用于GPU推理、网络传输、网页渲染等。另外， 不同数据中心（如AWS、GCP）对客户响应速度也有差异，还涉及到cache。在命中cache之前可能还会有CDN对用户体验的影响。因此需要RUM产品去做全链路追踪，帮助优化用户体验。用户在不同模型间转换时，除了关注token价格和答案质量， 也关注网站交互体验，如登录速度。以前GPT和Anthropic都出现过网站无法访问的情况，因为用户量巨大，RUM对他们来说非常重要。Datadog在这方面做得很好，因为它覆盖全链路。
来源背景：未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。 - 支持 Anthropic 合同是八位数；当季还有两个 PCV 过亿的客户
从公开信息看到是八位数合同，提到本季度有两个PCV过亿的客户。
来源背景：未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。 - 反对 有自有云基础设施的公司 (GCP、AWS、阿里、腾讯) 不会用 Datadog
GCP肯定是用谷歌自己的可观测方案，就像阿里、腾讯这些，只要有自己的云基础设施，包括AWS，也不会用Datadog。
来源背景：未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。 - 支持 OpenAI 自建是因为超大规模下第三方监控不经济，属 outlier
**关于 OpenAI 流失的关键判断**:笔者在上篇专题 3 已经解释——OpenAI 自建的本质是"超大规模导致第三方监控费用不经济",**这是 outlier,不代表所有 AI 客户都会效仿**。
来源背景：Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。 - 反对 Anthropic、xAI 等头部 AI 公司会在 2027–2029 年陆续效仿自建
但**部分头部 AI 公司(Anthropic、xAI)会效仿**——2027-2029 年陆续发生。笔者的处理方式:
来源背景：Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。 - 口径不齐 作者把超大 AI 公司约 20–30 亿美元的 AI 观测需求视为 DDOG 拿不到
* 把 AI Observability 市场池 $10B 中,超大 AI 公司(OpenAI + Anthropic + xAI + Meta AI + Google AI)的需求 ~$2-3B 视为"DDOG 拿不到" * **DDOG 实际可竞争市场 = $7-8B**(扣除超大客户自建)
来源背景：Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。 - 中性 一家超大厂和一家大科技公司把 Datadog 用于训练；CFO 称是“背书”但不一定广泛使用
And this includes some of the foundation model companies, companies in database, companies that are vertical. So already, Datadog has been used pretty pervasively in the monitoring side, so production work environments, inference production, and I think we said over 650 and we gave a lot of statistics on spending over $10 million in 10-plus products. And what we added to that is that two larger companies, a hyperscaler and another very large tech company, that have model creation within their businesses, foundational models, had used Datadog for training as well. Most of the time, our end market has been for production workloads. But as a couple of things have happened, as there is a boundary between training, research or training and when it has to go into production, that we've now been able to have some customers buy from us, and in this case, they were some large customers, hyperscaler, who traditionally does more things on their own but use Datadog.
- 口径不齐 CFO 强调不过度承诺：收入大部分仍来自生产推理，训练是否成核心市场不知道
I mean, you mentioned earlier in our conversation that Datadog does well with inference and now you're saying training, too. And so... We're saying maybe – we're saying we have, but we're not – we don't believe – we haven't said they're – we generally don't overpromise. So when we have like a certain amount of training that is spread out, we'll tell you. Right now, it's more of sort of centralized. It's good, but we have said most of our revenues we expect to be from production inference. But we'll see what happens. We'll bring everyone along.
- 支持 未见大规模客户从 Datadog 迁往 Palo Alto (Chronosphere)，Datadog 地位稳固
回 答 者：我们尚未观察到大规模从 DataDog 迁移过来的情况，DataDog 的市场地位依然非常稳固。Palo Alto 目前的策略主要是通过产品捆绑等方式，在现有客户群中进行业务扩展。此外，Chronosphere 主要专注于与安全相关的可观测性问题，而 DataDog 的业务范围更广，涵盖了 IT 基础设施、用户行为体验、数字体验和真实用户监控等领域。因此，Palo Alto 无疑会在安全可观测性领域占据一席之地，但我认为 Chronosphere 不会与 DataDog 或 Dynatrace 构成直接竞争，其定位将更加聚焦于安全。
来源背景：未具名专家，Capgemini 前任人员 (由同场 MeritForum 06-05 标题'Capgemini 离职专家'推得)；以'we/our customers'口吻谈 Palo Alto 增长预期，推测为 Palo Alto 渠道/集成伙伴侧负责人。视角局限：伙伴视角只看到自己的客户群；对 Datadog 仅顺带评论。 - 反对 Chronosphere 原生做 GPU 级方案，因此得到 Anthropic、OpenAI 认可
它有两个优势：第一是降本增效能力，包括对数据点做采样，减少数据量。很多遥测数据需要采样，比如秒级、毫秒级采集的数据要上卷到分钟级、五分钟级，去看更粗的趋势，历史数据会丢弃，这里需要采样算法帮助有效上卷，减轻存储压力。Chronosphere 在数据进入存储前就能做过滤、聚合、塑形、采样，这些能力很强。第二，它原生做GPU解决方案，非常了解不同GPU型号及相关外设，包括铜缆、连接、交换机等，整套指标体系都做了，知道哪些指标重要，代表什么意义，并制定了对应规则，做了GPU级的解决方 案，所以能得到Anthropic、OpenAI等公司的认可。ClickHouse只是一个OLAP数据库，强调数据分析能力。
来源背景：未具名；以 Elastic (ES) 内部视角回答 (清楚 ES 组织架构调整、API 文档策略、Jina 整合)，推测为 Elastic 现任或前任员工，带明显中国区视角 (国内定价、阿里/腾讯 ES 服务)。局限：竞争对手视角；对 Datadog 的判断是间接的。 - 支持 模型大厂是 C 端场景，特别在意 RUM，ES 在这块不如 Datadog
这个不好说。云上的增速不太好判断，现在很多OP的业务，企业都在做agent，这会涉及到新的observability，比如LLM的observability。OP这块也会有增量。ES在指标方面刚出的PromQL支持，而且还没有看到有大型参考 客户可以用来讲故事。ES的IUM这块肯定不如Datadog强。Anthropic、OpenAI这些模型大厂本身是2C场景，特别在意IUM。
来源背景：未具名；以 Elastic (ES) 内部视角回答 (清楚 ES 组织架构调整、API 文档策略、Jina 整合)，推测为 Elastic 现任或前任员工，带明显中国区视角 (国内定价、阿里/腾讯 ES 服务)。局限：竞争对手视角；对 Datadog 的判断是间接的。 - 支持 Truist：对与前沿 AI 实验室关系的持久性信心提升，消除主要压制
Truist upgraded DDOG to Buy and raised its PT to $300, arguing investor concerns around AI optimization are being overwhelmed by the urgency of enterprise AI adoption. Recent customer and partner conversations suggest Datadog remains in the early innings of AI-driven growth, with non-AI-native enterprises increasingly deploying agents that require more telemetry, monitoring, and observability. The firm also highlighted improving confidence in the durability of DDOG’s relationships with frontier AI labs, reducing a key overhang on the stock. While valuation has expanded, Truist believes Street still underestimates the durability of consumption growth and the size of the opportunity as AI adoption broadens across the enterprise.
来源背景：TMT Breakout (TMTB) 日报，转述 Truist 升级报告摘要 (基于客户与伙伴交流)。二手转述，看不到底稿。 - 支持 AI 原生客户 Datadog 消费同比 +35% 到 50%以上，传统客户 +10–20% (即 '35–50%' 数字出处)
So if we talk about the consumption on Data Log, if we compare the AI native customers and non-AI native customers, what are their consumption growth in Q2 comparatively? Yeah, um, So, look, there is a, I would say noticeable difference between AI native and non-AI native customers. From what I have observed, the AI native customers see data dog consumption probably growth in somewhere from 35 to more than 50% year-over-year. Whereas, uh, if you're looking at traditional customers, you know, they are growing in 10 to 20% range. I think the gap is driven by um, very high, uh, telemetry generation from the AI workload, which includes stuff like GPU, infra monitoring, model inference tracking, vector DBs, agent orchestration layers and the need, you know, for end-to-end observability across the application. application stacks which are becoming more and more complex.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 最大的 AI 模型公司可能继续减少对第三方可观测性工具的依赖 (自建)
And I think in many cases, they already operate internal tool that is more far more specialized than what a commercial observability platform gives you. The other question is whether that materially threatens Data Dog. I think for a hyper scale AI model providers themselves, uh, I would say potentially yes. The largest AI companies, you know, may continue to reduce dependence on third-party observability tool as they build purpose-built internal platform for them. Um, although I don't think that trend extends to a a broader market, you know, because most customers they lack the engineering resources to build and maintain observability platform at that scale. So, if you look at Datadog, their real customer base is not OpenAI or Anthropic alone, right? It's thousand of companies running complex cloud and AI workloads. Now, those customers, they generally prefer buying an integrated observability platform. rather than one building for themselves. So even if OpenAI and Anthropic become fully self-sufficient internally, I don't think that automatically translate into a viable alternative for enterprise customers.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 支持 自建不会扩散到广泛企业市场，多数客户缺乏工程资源
I think for a hyper scale AI model providers themselves, uh, I would say potentially yes. The largest AI companies, you know, may continue to reduce dependence on third-party observability tool as they build purpose-built internal platform for them. Um, although I don't think that trend extends to a a broader market, you know, because most customers they lack the engineering resources to build and maintain observability platform at that scale. So, if you look at Datadog, their real customer base is not OpenAI or Anthropic alone, right? It's thousand of companies running complex cloud and AI workloads. Now, those customers, they generally prefer buying an integrated observability platform. rather than one building for themselves. So even if OpenAI and Anthropic become fully self-sufficient internally, I don't think that automatically translate into a viable alternative for enterprise customers. And the biggest uh, bigger competitive risk for Datadog in my view is not AI model companies building their own tools.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 模型公司为自己 API 加 token/延迟/成本监控，会侵蚀最简单的 AI 可观测性
That because that they are basically selling API access. And and this is like very much connected to uh the service they provide and it's it's a sort of a cheap and uh and it's a like a retention hook. So so so some of that could erode like the simplest form of AI observability. However, uh the the real enterprise problem that Data Dog is addressing, that is not like, you know, observability or even AI observability within one platform. It's like, you know, cross-platform, cross vendor, etc, right? So so a real incident could involve, you know, the underlying model, but also the orchestration framework. You know, you may have uh an underlying database, you could have downstream services, you have the infrastructure underneath. And and no single model provider uh will see that whole picture. And and I think also it's too messy to to to monitor everything that I described for a model company.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 口径不齐 Citi 承认投资者担心 tokenization、出口管制、实验室客户集中度，但认为风险被高估
Following DASH ‘26 and customer checks, Citi raised its PT to $270, arguing Datadog’s competitive position continues to strengthen as enterprises scale AI and agentic workloads. Customer feedback was broadly positive on AI observability demand, with Bits AI and Cloud Cost Management receiving strong interest, while displacement opportunities against Splunk and New Relic remain robust and competitive pressure from open-source alternatives appears limited. Citi acknowledges investor concerns around tokenization, export controls, and lab-customer concentration, but believes these risks are overstated relative to Datadog’s expanding AI opportunity, growing enterprise adoption, and ability to capture incremental observability spend.
来源背景：TMT Breakout (TMTB) 日报，转述 Citi (DASH 后与客户检查) 报告摘要。二手转述，看不到底稿。 - 反对 若 Anthropic 与云厂商集成做可观测性，将是重大威胁
对,然后小 呃 please go ahead. Yeah. I think at the at the hyperscalar level, I think if there is, if there is any integration, um I think with Anthropics then this would be a ma a major threat, yes. Um if there is any development from that side, um because that would be a a huge opportunity to to give the enterprise, I think. OK,对,他就是还是重申了他这个观点,如果anthropic决定在这方面发力的话,就未来不仅提供告诉你漏洞在哪,还能提供解决方案的话,那对Datadog来说会是一个很大的威胁。
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 支持 Anthropic 目前只能发现问题、不能采取行动，暂时替代不了 Datadog
A:目前还不会。Anthropic 能够发现问题并提供大量信息，但它目前不能采取行动。在不久的将来，也许它能够近似一套可观测性解决方案，但在目前阶段它还没有能力做到那样。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 AI 实验室自建完整的观测方案还很遥远
A:我认为这还相当遥远。要构建像 Datadog 那样的完整解决方案需要时间。短期内不会；也许将来会。到目前为止，对任何实验室来说要做到这一点都是具有挑战性的。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 所有顶级基础模型公司 (含前两名) 都用 Datadog，另有大量 AI 应用公司
So the observability layer is actually going to be the fundamental spend driver for most of the customers and that's the product that uh uh they start with, right? And in terms of the AI uh, customers that Data Dog has, they they have actually a very wide roster when it comes to AI customers. All the top, uh fundamental model making companies are using uh DataDog for observability including the top two uh sort of like model providers that we all know about, right? And in addition, a lot of AI applications companies are also uh sort of like uh using Data Dog. Now coming to the application that they're observing the most, right? These are usually critical revenue driving applications, right? So for example, if we talk about an e-commerce company, the main uh the main applications that they would be uh monitoring with with this observability suite will be the uh revenue driving application like mostly the uh website which is customer facing, right? You don't want the payment gateway to start failing, you don't want the uh card functionality to start failing. So anything that will our major revenue impact is what we see uh companies uh usually monitor first of all and then maybe then they'll expand to other internal applications also which are uh uh critical to them.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 口径不齐 AI 原生公司占总收入 high single digits，但对增长贡献显著且上升
The percentage contribution of uh applications that are driven by AI agents is actually uh increasing. But uh, a lot of a lot of the growth also comes from core observability security uh positions right now. So, uh, in terms of the overall revenue contribution, I would say that AI native companies would right now probably be uh in the high single digits in terms of overall revenue contribution, not uh contribution to revenue growth. But uh, uh for the for the growth itself, the contribution is significant and the contribution is actually uh increasing for uh uh AI native companies specifically, right? Given they're going through a massive uh user growth as well, if you look at the uh major AI native companies right now. And in addition, there's also the fact that a lot of companies are increasingly using their traditional products like application performance monitoring in order to monitor applications that they might have created with uh with uh uh AI or created using AI models or which have AI models in the backend, right? So, given a lot of new applications are also being written as a result of AI and using the AI models, that is also driving uh further acceleration of the revenue, which the company is not tracking separately. Like they're tracking the revenue that they're getting from uh the AI company cohort, the AI native cohort companies, but uh you know, they they're not specifically uh monitoring the revenue that they're getting from traditional products like application performance monitoring, but which are driven by, let's say applications that are actually uh with uh you know, the the uh fundamental AI models in the uh in the backend. Hmm, okay. 他说是这样的，就是，应该是一个非线性的，就是肯定不是1比1，然后是1比就是超过了这个1的这样的一个比值。
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 OpenAI 占收入低到高个位数，即使完全替换也影响有限，可被其他 AI 原生客户增长抵消
And for the other customers of Datadog, like the other AI native companies or Antropic or enterprise companies, do you think they will do similar things so that they will reduce their, you know, revenue uh on Datadog? So, uh good question. So, first of all, for OpenAI, the revenue contribution is in the low, uh low to high single digits, right? So it's uh for a company of our scale, it's somewhat concentrated revenue, but it's not as if uh the contribution is massive, right? So even if we assume that OpenAI is going to completely replace Datadog, uh that is not going to have that meaningful a hit on the company's overall revenue numbers, especially given that the AI cohort is the one that's actually witnessing much more uh uh accelerated growth versus the rest of the cohort, right, in terms of traditional companies. So even if let's say OpenAI uh is hypothetically doing it, uh the the the growth from a lot of the other uh other uh AI native companies is actually going to uh counteract that. And I think OpenAI especially given their loss of market share in some of the...
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 其他 AI 原生/模型公司过去 5–6 个月签了多年期承诺用量合同
So it's understandable that they might uh go ahead with moves like this. But I also want to uh surface that uh you know, some of the other companies, some of the other AI native companies have signed contracts very recently. like the other major model providers, some of the ones we're discussing, the contracts have been signed in the last five, six months. So, uh and and these are not single year contracts, they're relatively uh longer term commitments, right? So, there's some amount of guaranteed committed usage from there and I believe that the risk of complete replacement from uh uh, you know, some AI tools that people are using to write observability software is actually relatively lesser. And I'll tell you why there are like three, four reasons. It's a big, big lift.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 Anthropic 去年成为客户，几个月前已续约扩品，占收入低到中个位数，是 Datadog 史上最大单之一，随其用户增长成比例增加
So do you think these products will be a threat to DataDog. So yeah, I can take that question. So, they became a customer sometime in the last year, uh for the first time and they have actually already signed an extension. Uh, I think a few months back. So, uh, they're already seeing increased demand within the company for more data usage. They've expanded in terms of scope, they've expanded in terms of uh the uh products that they're using from Data Dog. So this definitely looks like a longer term uh commitment and the the revenue is as I said like similar to the other model provider that we talked about, right? In the low uh to medium single digit uh uh of of the overall uh revenue, relatively smaller actually, but a big big uh one of the biggest deals that data has uh signed historically. So, uh definitely see this further increasing. First of all, because there's going to be uh as as you see like anthropic themselves are on a very exponential revenue growth trajectory, right? So as their users increase naturally, they're going to have to scale up their infrastructure, so they'll have more infrastructure to monitor, more hosts where their application is deployed on. So, uh their costs associated with data are actually going to increase uh proportionally with their users, right? So, definitely there's going to be an increase. putting a specific number to it is very uh difficult and not something that I think anybody can very uh predict in terms of what this customer will end up spending over the next uh year or so.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 中性 专家推测 OpenAI 减量的原因：在 C 端聊天和编码 agent 市场输给 Anthropic、IPO 前成本压力
So even if we assume that OpenAI is going to completely replace Datadog, uh that is not going to have that meaningful a hit on the company's overall revenue numbers, especially given that the AI cohort is the one that's actually witnessing much more uh uh accelerated growth versus the rest of the cohort, right, in terms of traditional companies. So even if let's say OpenAI uh is hypothetically doing it, uh the the the growth from a lot of the other uh other uh AI native companies is actually going to uh counteract that. And I think OpenAI especially given their loss of market share in some of the... key uh markets that they're operating in, such as the uh business to consumer chat agent and also the coding agents etc. The loss of market share that they've had to Anthropic. under cost pressures in the short term due to uh due to uh the the IPO filings etc, right? So it's understandable that they might uh go ahead with moves like this. But I also want to uh surface that uh you know, some of the other companies, some of the other AI native companies have signed contracts very recently.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 MS：一家大 LLM 客户、两家研究实验室和更广泛企业负载推动消费
Morgan Stanley reiterates Overweight and raises its price target to $300 from $225, saying stronger enterprise checks and robust public-cloud demand support mid-30% 2Q revenue growth versus Street near 30.5%. Morgan Stanley says AI-native demand is becoming more measurable, with one major LLM customer, two research labs and broader enterprise workloads driving consumption across observability, GPU monitoring, security and data tools. Morgan Stanley expects a roughly 32% 3Q guide and at least 30% FY26 growth, while Morgan Stanley raises FY26/FY27/FY28 revenue estimates by about 1%/4%/5%; Morgan Stanley acknowledges the stock is no longer cheap at roughly 16x CY27 sales but believes continued acceleration supports the premium.
来源背景：TMT Breakout (TMTB) 日报，转述 BTIG (伙伴检查) 与 Morgan Stanley (企业检查) 报告摘要。二手转述，看不到底稿。 - 支持 BofA：两家 hyperscaler AI 研究大单和 5 家 $10M+ AI 客户分散客户基础，降低续约波动担忧
BofA keeps DDOG as a top 2Q26 pick, saying its AWS revenue analysis suggests current estimates are conservative and leaves room for another acceleration from the 32% y/y growth reported in 1Q. BofA says additional frontier-lab wins, including two hyperscaler AI-research deals and five $10M+ AI customers, diversify the customer base and make renewal volatility less concerning. BofA also argues the recent OpenAI incident reinforces the need for agent observability as enterprises require greater visibility into autonomous AI behavior, expanding DDOG’s long-term opportunity.
来源背景：TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (AWS 收入分析)。二手转述，看不到底稿。 - 反对 Funda 判断 OpenAI 2Q 用量环比降约 10–20%，后续还会继续优化
We spoke with six experts this quarter: two channel/GSI delivery partners covering global enterprise accounts, one North American F500 channel partner, one observability industry consultant, and two North American enterprise customers on the user side. Overall, channel partners report healthy business trends. Growth is still led by the traditional three (Infra/APM/Logs), with security and Dev-related products taking over as the second growth curve, while AI products themselves are still early in their contribution. On the frontier LLM customers, we understand OpenAI’s 2Q usage declined roughly 10-20% q/q with further optimization ahead, while Anthropic’s 2Q usage should be quite strong — net net, limited impact on the quarter.
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 Funda 估 2Q26 AI 原生收入约 $170m，占总收入约 15%
+ Other LLM customers also stepped up model training activity in 2Q, providing a further growth tailwind * **Net net: we estimate 2Q26 AI native revenue of roughly $170m, or about 15% of total revenue**
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 最大客户 (一家领先 AI 公司，使用 17 个产品) 签了九位数续约，但用量从 Q3 开始下降
Datadog also proved value beyond core observability with product analytics, CI visibility, data observability and cloud cost management. This deal includes our largest win to-date for Bring Your Own Cloud, displacing their legacy commercial logging tool at a petabyte scale. And finally, we signed a nine figure renewal with a leading AI company. This long time, very large customer uses 17 Datadog products to enable unified visibility on production workloads at a very large scale, albeit with a usage reduction starting in Q3, which we considered in our guidance and which David will speak to. Before I turn it over to David for a financial review, let me offer a few words on our longer-term outlook. There is no change to our overall view that digital transformation and cloud migration are long-term secular growth drivers for our business.
- 反对 CFO：已看到最大客户用量下降，已计入 Q3 和全年指引
Our guidance philosophy overall remains unchanged. As a reminder, we base our guidance on trends observed in recent months and imply conservatism on these growth trends. Regarding our largest customer, we have seen a usage reduction, which is incorporated in our Q3 and full year 2026 guidance. As Olivier noted, this customer has recently renewed with us. For the third quarter, we expect our revenue to be in the range of $1.135 billion to $1.145 billion, which represents a 28% to 29% year-over-year growth. Non-GAAP operating income is expected to be in the range of $260 million to $270 million, which implies an operating margin of 23% to 24%.
- 口径不齐 CEO 称对该客户 '完全去风险'；对 Sanjit 问的 '是单价下降还是流失/减量' 没有回答
So maybe I'll take this one. I think we – so overall, we as usual, we don't want to comment too much on any specific customer because we also don't really control what's happening with any specific customer. We wanted to be transparent about this on the call, because we did see a reduction in usage and we took the liberty to fully de-risk the guidance for the rest of the year with respect to that customer. And again, the reason for that is we don't control what's happening to a specific customer, but we do have a great amount of control on what's happening to everything else in the business. And the business is booming. And we don't want that to overshadow basically the acceleration we see pretty much everywhere else in the business. So as we mentioned on the call, we renewed the customer, it's a longtime customer who uses many of our products. But there's not a lot more we can share. Yeah. I think...
- 中性 大客户合同结构：有承诺 (commit) 的基础量，但用量相对承诺有波动；指引用承诺来去风险
Okay. ...just to get specific on the guidance. We last quarter and previous quarters said that we essentially have a level of commit and we can de-risk our guidance by using that. And then as you know, in most of our large customers, we have variability relating to the commit, so take that into consideration. Yeah. The last thing I will say because I know so it's on people's minds, is if you back out our largest customer from our growth, you get pretty much the same growth rate, as the rest of the business has been accelerating very steadily.
- 支持 剔除最大客户后增速 '差不多一样'，即最大客户已不再拉动增长
And then as you know, in most of our large customers, we have variability relating to the commit, so take that into consideration. Yeah. The last thing I will say because I know so it's on people's minds, is if you back out our largest customer from our growth, you get pretty much the same growth rate, as the rest of the business has been accelerating very steadily. Actually we've seen, I think now five quarters of continuous acceleration from the rest of the business. And we feel very good about the – what we see in the market.
- 口径不齐 续约条款 (期限、折扣) 拒绝评论；只说企业客户一般是年度以上合同、按量阶梯定价，续约 '有很多延续性'
And then, David, maybe just for you. On the largest customer renewal, is there anything you can tell us around maybe just any changes around the duration or anything that makes this new contract maybe a little stickier in terms of the discounted rate card, the amount of products that they're able to kind of use for better value, anything that increases the conviction level around stickiness. I won't comment on this other than to say that, most of our enterprise customers, as we've talked about for a long time, have annual plus and then the pricing is generally volume-based pricing. So, I would say overall, our customers transact with us in that way. And then we have that level of commitment. And then as we've talked about over a lot of years, then there's usage and then we transact. So it's similar to what we have with most of our larger enterprise customers. Oli, anything you want to add there? No, I think there's a lot of continuity in that renewal. I think that's what you can. It's one way to put it.
- 支持 AI 客户组 750 家，>$1M 31 家 (Q1 22 家)、>$10M 8 家 (Q1 5 家)；但口径扩大到包含 hyperscaler 内部 AI 实验室
We saw robust growth across our customer base with broad-based strength across customer size, spending bands and industries. Meanwhile, our AI customers continue to grow rapidly and diversify in the quarter. This 750 strong customer group includes a broad range of AI startups as it has in the past, but now also includes hyperscalers using Datadog for in-house AI labs. In Q2, this includes 31 customers spending more than $1 million annually, of which eight customers spent more than $10 million annually. We also achieved strong new logo dollar bookings with particular strength in enterprise, where new logo annualized bookings more than doubled from the year ago, and we are seeing new logos ramping faster and contributing more to revenue growth. The portion of our year-over-year revenue growth that relates to new customers was about 30% in Q2, up from 25% in Q1.
- 口径不齐 分析师 (Stifel) 推算剔除最大客户的下滑后指引可能高 300–400bp，CEO 拒绝推测
Thanks very much. Oli, given your commentary around how strong the core is and that your largest customer was not additive to growth here in 2Q. Should we assume that if we ex-out the sequential down tick in that customer, that the core guide would have been probably 300 basis points or 400 basis points higher? Well, I can't speculate, but what I will say is, look, the business overall is growing at the same rate, if you exclude that customer, as I said. And the business has been accelerating overall. That's why we feel good.
- 支持 剔除最大客户的 AI 客户仍在增长且快于非 AI
So it's a good time to be in business. Yeah. I think we commented in the remarks that the non-AI has accelerated and the AI excluding the largest customer continues. So I think we gave those trends in describing the business. Good.
- 反对 Funda 估 OpenAI 2Q 支出环比降 20%，市场共识太乐观
* 26Q2 revenue grew 35.6% YoY, in line with buy-side expectations of 35-36%, accelerating 3ppts QoQ. However, the US$46m beat was slightly below the US$50m last quarter. * 26Q3 revenue guidance was a touch light at 29.5% at the high end, where expectations were looking for a 3-handle. Applying a typical 5ppts beat would imply a marginal sequential growth deceleration in 26Q3. * **The miss appears mostly driven by usage optimization at OpenAI, DDOG’s largest customer. For 26Q2, we estimated a 20% QoQ decline in OpenAI’s spending, on which market consensus appeared too optimistic. Management confirmed it has renewed a 9-figure contract with OpenAI, but the latter’s usage will decline sequentially in 26Q3. We believe Chronosphere is gaining share within the OpenAI platform, as it prices 30-40% lower than DDOG.** * Core/non-AI usage growth continued to accelerate to high-20s from mid-20s, driven by strong cross-sell momentum and increasing usage from cloud migration and AI app development. * Full-year revenue guidance was raised by more than the 26Q2 beat flowed through, to 30-31% YoY, in line with expectations. * The margin trend remains strong — GPM dipped below 80%, but OPM beat expectations on operating leverage.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 口径不齐 管理层确认与 OpenAI 续签 9 位数合同，但 3Q 用量继续环比下降
* 26Q2 revenue grew 35.6% YoY, in line with buy-side expectations of 35-36%, accelerating 3ppts QoQ. However, the US$46m beat was slightly below the US$50m last quarter. * 26Q3 revenue guidance was a touch light at 29.5% at the high end, where expectations were looking for a 3-handle. Applying a typical 5ppts beat would imply a marginal sequential growth deceleration in 26Q3. * **The miss appears mostly driven by usage optimization at OpenAI, DDOG’s largest customer. For 26Q2, we estimated a 20% QoQ decline in OpenAI’s spending, on which market consensus appeared too optimistic. Management confirmed it has renewed a 9-figure contract with OpenAI, but the latter’s usage will decline sequentially in 26Q3. We believe Chronosphere is gaining share within the OpenAI platform, as it prices 30-40% lower than DDOG.** * Core/non-AI usage growth continued to accelerate to high-20s from mid-20s, driven by strong cross-sell momentum and increasing usage from cloud migration and AI app development. * Full-year revenue guidance was raised by more than the 26Q2 beat flowed through, to 30-31% YoY, in line with expectations. * The margin trend remains strong — GPM dipped below 80%, but OPM beat expectations on operating leverage.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 支持 AI 原生 ARR>$1M 客户 31 家 (1Q 22)，>$10M 8 家 (1Q 5)，均创新高；两家 NeoLab 签 7 位数
* DDOG now has 31 AI-native customers with ARR over US$1m (up from 22 in 26Q1), including 8 with ARR over US$10m (up from 5 in 26Q1). Both net adds were record highs. DDOG also signed two NeoLabs to 7-figure deals during the quarter. * If we were to assume a similar 5-ppt beat on overall revenue, continued acceleration in non-AI revenue growth, and a 20% decline in OpenAI revenue, **we would see non-OpenAI AI-native revenue accelerate further in 26Q3 and exceed OpenAI in revenue contribution for the first time. By then, OpenAI’s optimization would become a less material driver of revenue growth.** However, we need to be aware that Anthropic’s DDOG usage growth likely peaked in 26Q2, as it is indirectly linked to its own ARR growth, which may have slowed sequentially in 26Q3. * Management pledged to accelerate its internal AI lab investments in world models and LLM post-training for observability, helped by the Adaptive ML acquisition that closed in late June.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 支持 Funda 推算：3Q 非 OpenAI AI 原生收入将首次超过 OpenAI
* DDOG now has 31 AI-native customers with ARR over US$1m (up from 22 in 26Q1), including 8 with ARR over US$10m (up from 5 in 26Q1). Both net adds were record highs. DDOG also signed two NeoLabs to 7-figure deals during the quarter. * If we were to assume a similar 5-ppt beat on overall revenue, continued acceleration in non-AI revenue growth, and a 20% decline in OpenAI revenue, **we would see non-OpenAI AI-native revenue accelerate further in 26Q3 and exceed OpenAI in revenue contribution for the first time. By then, OpenAI’s optimization would become a less material driver of revenue growth.** However, we need to be aware that Anthropic’s DDOG usage growth likely peaked in 26Q2, as it is indirectly linked to its own ARR growth, which may have slowed sequentially in 26Q3. * Management pledged to accelerate its internal AI lab investments in world models and LLM post-training for observability, helped by the Adaptive ML acquisition that closed in late June.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 反对 Anthropic 的 DDOG 用量增速可能在 2Q 见顶，因与其自身 ARR 增速挂钩
* DDOG now has 31 AI-native customers with ARR over US$1m (up from 22 in 26Q1), including 8 with ARR over US$10m (up from 5 in 26Q1). Both net adds were record highs. DDOG also signed two NeoLabs to 7-figure deals during the quarter. * If we were to assume a similar 5-ppt beat on overall revenue, continued acceleration in non-AI revenue growth, and a 20% decline in OpenAI revenue, **we would see non-OpenAI AI-native revenue accelerate further in 26Q3 and exceed OpenAI in revenue contribution for the first time. By then, OpenAI’s optimization would become a less material driver of revenue growth.** However, we need to be aware that Anthropic’s DDOG usage growth likely peaked in 26Q2, as it is indirectly linked to its own ARR growth, which may have slowed sequentially in 26Q3. * Management pledged to accelerate its internal AI lab investments in world models and LLM post-training for observability, helped by the Adaptive ML acquisition that closed in late June.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 支持 AI 原生 750+ 家，30+ 家达百万级，前 10 大 AI 公司全是客户
The AI-natives themselves, where we've also won a lot of market share, have complemented that. And in that, I think we said it's pretty diverse. We had over 750 names, over 30 of them having a million, 10 of the top 10. So, that means that that group of cohorts is adopting Datadog on top of the overall environment and the overall business accelerating. So, we've had this stat for a number of years how to judge breadth and product adoption.
- 支持 AI 原生占 ARR 比例远低于疫情时的云原生
So, we've always been very successful in what we used to call cloud-natives. Now they're AI-natives. This group is a much smaller group as a percentage of ARR than when COVID happened. But it is growing very fast and it's the who's who. And it basically says that if you're essentially investing in modern technology, those companies do not have legacy systems.
- 反对 CFO 承认 AI 原生可能有波动
So, it's a really good forward-looking sign. In terms of a question that's come up. Is there going to be volatility there? Yeah, there might be. I mean, there's volatility in all these markets. But what is most important is you're compounding where technology is going over 5, 10, 20 years, which we've been doing successfully.
- 反对 大客户是“组合”：关键负载给 Datadog，其余可能自建，“总会有来回”
That being said, talk about like maybe usage trends among largest customers, like how did you get to that scale, and then just even how that conversation goes when you get to eight figures or nine figures of spend? Yeah. I mean, With larger customers, it's always been not one or the other. They do a portfolio of things and they tend to put their more modern workloads and mission-critical workloads being served by Datadog. So, there's always going to be the back and forth. The net weight has been towards not doing it yourself, but buying Datadog, that's what's produced from zero Datadog over $4.5 billion, that's produced Datadog for not being in the industry to being the largest player. So, the weight of this has been that way.
- 反对 研究员提问的前提：OpenAI 因 Datadog 成本高而自建替代
嗯。 And for do you think the top AI labs have the motivation or have the capability to um to um develop their own AI infra like Snowflake and the Databricks. because we know that like AI infra like DataDog, because the cost is very high, so OpenAI has been developing its own AI infra and trying to replace DataDog. So do you think Databricks also face a similar threat or risk of being replaced by those AI uh top AI labs themselves? And another question is you mentioned that the Genie's foundational models actually have this um multiple model routing mechanism. And this is a a bit similar like the cursor. And you know that last year cursor was very popular, but this year its market share dropped by about a half. and that's because of the because the large language model's coding capability has make a lot of progress. And do you think for uh Databricks' coding agent, uh it it may also face a similar threat when uh the large models coding capability further improve? I'll I'll first answer the question around the capability to develop one infra and platform, especially I think when you talk about players like um OpenAI, Claude, uh who are um moving very faster into these uh whole platform uh uh, you know, comparison, especially when they are competing with players like Microsoft uh in terms of foundry and uh, you know, uh agent 365 and other products. So, right now, I think uh there are certain areas where um the uh OpenAI, Anthropic, Google, uh they of course have the capital and engineering capability to build something very similar to Databricks. But right now their focus uh is more on some other areas. Like they are not focusing a lot on data pipelines and ETL.
来源背景：未具名专家，从'我们已经进行了大量优化''我们还推出了自己的 AI Gateway'推测为 Databricks 现任员工 (产品/解决方案方向，未明示)。视角局限：谈 Databricks；Datadog 只出现在研究员提问里，专家没有回应 Datadog 部分。 - 反对 非头部 AI 原生公司业务波动大，可能更在意成本
**EE** [00:07:51]: I think AI-native companies obviously are newer, they're more modern, their stack is more modern. They move faster, they're more agile. I think because they are more modern, they might explore different ways of doing different things. It depends. I mean it's like some native AI companies are booming and already worth billions of dollars and maybe spend is not the most critical thing for them, and they're not thinking about it right now. Maybe smaller commercial companies that are still AI-native, obviously, they're not the OpenAIs or the Lovables of the world, I think they could be more cost conscious at times based on how their business is going because even AI-native companies or SaaS companies, we're seeing that concern where Claude or Anthropic or OpenAI could just go and build something their selves, or things are easy to replicate. I think because their business could fluctuate then they might be a little bit more cost conscious. I don't know.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 反对 承认对最大 AI 客户 '确实有依赖'
**EE** [00:39:47]: I think we both know the company we're talking about. There's definitely a dependency. I can say that it's healthy to depend on one large customer or another, but I don't know.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 反对 PANW：4Q 净新增 ARR 含大型 LLM 客户从现有供应商迁到 Chronosphere 的 9 位数收益，DDOG -3%
On **DDOG -3%:** “*Keep in mind, as we noted last quarter, our Q4 net new ARR includes a nine-figure benefit from a large LLM customer migrating to Chronosphere from an incumbent vendor.*”
来源背景：TMT Breakout (TMTB) 日报，PANW 财报点评，引用 PANW 管理层原话 (CEO/CFO 未具名) 与 TMTB 自己的分析。关于 '大型 LLM 客户' 未点名 (市场普遍认为是 OpenAI)。二手转述，看不到底稿。 - 反对 Chronosphere ARR 超 $500M，4Q 有超 $100M 来自一家 LLM 客户迁移，F1Q 还有不到 $100M
Cost synergies remain three to six months ahead of plan. Modern PAM broadens the addressable privileged-user base, while Secure IAM and non-human identity controls create additional land-and-expand opportunities, although “true” independent agentic identity remains early. * **Observability.** Chronosphere ARR surpassed $500M and more than doubled following the acquisition, but the quarter included over $100M of NGS ARR benefit from one large LLM customer’s migration. The remaining migration benefit should be below $100M in F1Q before normalizing materially, creating a difficult quality-of-growth and concentration comparison after F1Q. Offsetting that concern, PANW cited a separate ~$20M ACV AI-inference customer, 50% XSIAM attachment among Chronosphere net-new logos, and a path to broader enterprise selling once feature parity is reached over the next six months. * **Margins.** F4Q non-GAAP gross margin was 74.8%, down ~100 bps y/y and below Street 76.1%, reflecting faster-growing cloud/SASE revenue, higher hosting expense, acquired-business mix and memory/storage inflation.
来源背景：TMT Breakout (TMTB) 日报，PANW 财报点评，引用 PANW 管理层原话 (CEO/CFO 未具名) 与 TMTB 自己的分析。关于 '大型 LLM 客户' 未点名 (市场普遍认为是 OpenAI)。二手转述，看不到底稿。 - 支持 31 家 AI 原生客户年化收入超过 100 万美元，其中 8 家超过 1,000 万美元
家 AI 原生客户年化收入超过 100 万美元，其中 8 家超过 1,000
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 - 中性 前沿实验室有用户单独操控成千上万个 agent；但它们推理免费，与按 token 算账的企业不同
Everybody starts with infrastructure and applications and logs, everybody has to get things working into production, everybody has to make sure that humans actually understand what's happening at the end of the day. Some of the day-to-day usage patterns are different. So, especially when you go into the frontier labs, for example, like you see customers that are – or users that are manipulating thousands or tens of thousands of agents individually, some of that is transferable to the rest of the market, and we expect to see it happen to other customers. Some of that is not. And obviously, when you have free inference and all the incentive in the world to push your models to the max, you're not in the same world as the enterprises that get the deal at the end of the month or every single token of inference they use.
- 支持 疫情时云原生峰值约 40% 业务，现在 AI 原生敞口小得多，其余 85% 业务谨慎控成本
We've seen the – a lot of those cloud native companies, at the same time, digital native companies expand like crazy and then contract quite a bit. We were very, very exposed to it at the time. And I think it was about 40% of our business at peak. And then, we suffered from the compression there. So, we are very, very careful about how we assess these accounts, how we handle their – how we watch for unhealthy behavior, for example, from customers and things like that. And we're very, very good now at getting in front of all of that in a way that we've learned over the past couple of years. The thing I would say, though, is that the exposure we have to these AI natives is much smaller. So, our business has grown quite a bit. It's very diversified.
- 支持 少数自建客户关掉后有回来的；GRR 在 high 90s；多家超大厂为 AI 建设找来
So, that's number two. Number three, like when you – when people reach large scale, there's often folks who want to build it themselves. And in some cases, they will, like we've had some customers that say, we're going to build it ourselves, and a number of them have turned off and some of them have come back. We've seen that, we've called some of those out on our calls. It's a very small fraction of our customer base. If you look at our aggregate numbers, our gross retention is extremely high, it's in the high 90s, including the whole business, from SMB all the way to enterprise, which is incredibly high. There's basically no upside there, like you can't imagine of anything that's substantially better in terms of retention. And what we've seen on a – fairly recently, I think we called that out a couple of earnings calls away ago.
- 中性 AI 观测需求与 AI 支出周期绑定，不会消失，但新增 workload 的速度可能放缓
JC: I mean, by definition, it's tied to the AI spending cycle, but I don't expect it to go away. There are already a huge amount of AI and LLM-powered workflows that have been put into production and they need to be monitored. As long as they are in production and serving production traffic, I think, depending on what happens in terms of the cost of tokens over time and using these solutions, we might expect the rate of new workloads to perhaps not continue at the pace that it has over the past quarter, 2 quarters. But I don't see technology executives saying at some point in the future, we need to cut the observability over this huge swath of systems.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 反对 最大客户：曾超承诺增长，已续约延长；但用量迁移，非实时的指标存储/数据库挪出 Datadog，带来波动；指引只计承诺额
But it isn't the case all the time, and it also isn't 100% or 0%. Many customers are doing a combination. And what we said all along, which is that from our largest customer, is that we have commitment contracts, and that customer was growing rapidly and spending above the commitment. But we know that in certain customers, there'll be sort of a tradeoff between. And what we said in our last earnings call was that, that customer renewed with us, staying with Datadog, extending the contract. And the usage there has moved around relative to what they're doing. Sometimes it makes sense to hook up a database that doesn't have to be or a metric store, doesn't have to be real-time with Datadog and put the other work that has to be real time. So, we had some volatility. And what we decided to do, given the investor focus on that, is to de-risk our guidance by putting in our guidance only the commitment can't go below the commitment. But we've always said that's a fringe case. For the most part, our business is not about outsourcing the hyperscalers' observability.
- 支持 GRR 在 upper 90s，大企业在高端
...a more complex stack or more DIY stack because the cost savings make it worth it. Yeah. I think when you look – step back and you look at our gross retention, means are you staying on Datadog or not, it tends to be, we said, in the upper 90s, with larger enterprise being at the high end of that. So, we don't retain every customer, but the vast majority of customers have been making the decision to use the Datadog platform. And in fact over time, as I mentioned, the trend has been that it is economic, priority-oriented to instead of building an observability platform yourself to put your scarce R&D dollars into your product and your business.
- 中性 很多客户是自建和 Datadog 的组合，不是全有或全无
And in fact over time, as I mentioned, the trend has been that it is economic, priority-oriented to instead of building an observability platform yourself to put your scarce R&D dollars into your product and your business. So, that's been the weight of it. But it isn't the case all the time, and it also isn't 100% or 0%. Many customers are doing a combination. And what we said all along, which is that from our largest customer, is that we have commitment contracts, and that customer was growing rapidly and spending above the commitment. But we know that in certain customers, there'll be sort of a tradeoff between.
- 支持 AI 原生从开源/自研转向 Datadog
And what they're increasingly realizing is, it's not efficient to spend that, building your own observability. Use Datadog. So, we're seeing a movement from open source or initial efforts or cobbling it together towards Datadog. We're seeing a use of – because they're newer customers, a use of many of our products, the platform. So, we're seeing the average use of products to be at the higher end.
- 中性 训练/post-training 来自部分超大厂和 AI 原生，CFO 说没有证据说明会成为“训练公司”
But then we thought back and we said, you know what's happening? The speed of this is such that the preproduction or late training into production is starting to merge, which created this revenue stream. We don't believe we now have the evidence to say that we're going to be the training company, but we're also seeing in some of these AI native companies the use of Datadog in training, which has been an additional source of revenues. The question we get, and you address this every earnings call with your comments on large customer... Yeah.
- 支持 Bernstein：OpenAI/Anthropic 消耗加速可缓冲 AWS 和 DDOG
Bernstein says AWS web metrics remain soft through early September, though Q3 should still show ~50bps of ex-AI growth acceleration given the strong start to the quarter. Bernstein sees more risk in Q4, where non-AI AWS growth could face an additional ~100bps headwind versus current estimates as traditional workloads weaken. Bernstein says accelerating OpenAI/Anthropic consumption should cushion AWS and DDOG, while lower-AI-exposure usage names such as TWLO and NET look more vulnerable.
来源背景：TMT Breakout (TMTB) 日报，转述 Bernstein (AWS 网站流量数据) 与 Citi (DT) 报告摘要。二手转述，看不到底稿。 - 反对 客户若不用 Datadog，替代方案几乎总是 OTel + 开源数据库自建，但需要大量人力
Bernd Harzog: Yeah, so what I would say is this, right? The first thing you have to understand is if you're not going to use Datadog or Dynatrace or something else in that category to watch an application that's doing something you care about, what are you going to do instead? Right? Because if it's something you care about, then the choice is not to do Datadog or do nothing. The choice is to do Datadog or Dynatrace or New Relic or something else, right? And the something else is almost always let's just leverage open source. Let's find a way to cobble together our own observability solution out of OpenTelemetry and some open source database. And the problem with that approach is that that requires you to burn people time to implement that and then manage that and support that and evolve that over time. You don't tend to have the people. Most companies don't have enough people to implement that approach at scale.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 反对 能自建的是 '规模很大但应用很少' 的公司 (单一业务单一应用)，这正是 AI 实验室画像
Bernd Harzog: That would be 1 dimension. And the other dimension would be, and this is true for the ones that do it, they can be very large, but they have relatively few applications. So think of a customer that is in 1 business with 1 application. Right. Okay, that 1 application generates 100% of the revenue. Okay. So we have 1 application here. Maybe we use 3 or 4 different languages in that 1 application. It runs in 1 place. Let's say it runs in the cloud, right? It has 1 principal architecture. Let's say it's microservices. Okay.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 反对 Chronosphere 成功吃走 '大规模部署的微服务应用'，常与 Datadog/Dynatrace 共存，其余应用留在原厂
Bernd Harzog: just applications running in the cloud, which is the easy thing to focus on. Right. Okay. So for them to compete with the entire company, Datadog or Dynatrace, would require a large amount of money and a large number of years, okay, for them to compete for a set of applications. You know, so, for example, Chronosphere very successfully competed for a set of applications, specifically ES: Mhm. Bernd Harzog: Microservices applications deployed at scale and in many cases had joint customers with Datadog and/or Dynatrace. So Chronosphere captured the set of applications that met their profile and everything else stayed with Datadog and Dynatrace, okay? So, you know, that kind of thing is much more likely than anybody replacing somebody who's got a true enterprise footprint.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 支持 Wolfe：DDOG 直接参与 Meta Muse 的可能性上升
Wolfe says ecosystem checks show agents materially increasing observability intensity as every tool call, retry and handoff needs to be traced, with traces costing roughly 2x metrics and already driving higher consumption. AI coding tools are also accelerating software-development velocity by ~2–3x, creating more code/services/pipelines for DDOG to monitor, while Wolfe’s initial conversations suggest an increasing likelihood of DDOG becoming directly involved with Meta’s Muse. Wolfe’s upside case now implies ~32% FY25–28 revenue CAGR vs. ~25.5% consensus and sees DDOG as one of the best-positioned ways to play inference/agentic demand.
来源背景：TMT Breakout (TMTB) 日报，转述 Wolfe Research 生态检查报告摘要。二手转述，看不到底稿。 - 反对 Jefferies：最大客户预计 3Q 起减量，形成 2027 年集中度/优化压制
Jefferies says DDOG core growth ex-AI natives accelerated to high-20s in 2Q26 and 750+ AI-native customers position it well for agentic observability. But its largest customer is expected to reduce usage from 3Q, creating a 2027 concentration/optimization overhang, while ~19x CY27 revenue leaves less attractive risk/reward. Hold, PT $300 from $260
来源背景：TMT Breakout (TMTB) 日报，转述 Jefferies 报告摘要 (DDOG 与 DT)。二手转述，看不到底稿。
E03 AI 让计费单位变多了吗
判断：方向成立，量的增长有扎实的例子；但从量到收入中间有两道闸门 (客户控制送进来多少、预算锚定云支出)，而且管理层自己说 agent 带来的量还没成为收入驱动。
成立的部分：AI 让企业写出更多应用：Funda 的北美渠道专家说，一家典型企业 2023 年中以前约 100 个开发者支持 400 个应用，现在开发者只多了 5–10%，应用数接近 650 (+50%)。应用多了，要监控的主机、容器、服务都多。一家北美 SaaS 基础设施公司的 DevOps 工程师说，用 Claude/Gemini 自动化以后，他每天用 Datadog 的强度和采集的遥测量都涨了 4–5 倍。CEO 的机制解释也讲得通：agent 大部分时间在调用工具，工具就是已有的应用，跑在 CPU 上，所以基础设施产品的消费在加速；技术栈 70–80% 和非 AI 应用一样。
AI 专属的量：MCP 工具调用 Q1 环比 4 倍、Q2 再 4 倍，比 2025 年 Q4 涨 22 倍；LLM Observability 的 span 6 个月涨 10 倍，1,000+ 客户；6,500+ 客户向 AI 集成发数据 (占客户数 20%、ARR 约 80%)。这些全是增速，没有一个绝对值或收入数字。
两道闸门：第一，客户决定送多少进来。一家大型信息服务平台的工程负责人说 agent 应用的遥测规模"小到不需要采样"；同一个客户把模型调用层的监控交给了更便宜的 BrainTrust，Datadog 只拿到应用层。第二，预算。前 Datadog 工程经理说观测预算约为云支出的 10%；独立顾问说环境扩大 3 倍预算不会跟着涨 3 倍。一位保险公司 IT 负责人说，公司上 AI 以后花的是账号和 token，没有专门加 Datadog 的采购。
结论：AI 对 Datadog 的拉动现在主要走"更多应用 → 更多主机和服务"这条老路径，量级大致跟着客户的云支出走 (渠道专家 B 估的弹性约 1.1 倍)。"每个 agent 步骤都要记录、Datadog 是 agent 电表"这条新路径还在早期，而且这部分的价格 (Agent Observability $3.50/万次模型调用) 远低于老产品。
+ Customers previously running 2-3 DDOG products are moving to at least 5; the land products are still Infra and APM, then Logs, followed by Synthetics and Cloud Security + The buying center is no longer just the CIO/CISO — SRE, QA, DevOps and platform engineering are now decision-makers, pulling through Synthetic Monitoring, CI Visibility and Error Tracking * 2H growth mix: traditional products (applications/infrastructure/logs) account for at least two-thirds; security contributes at least 15% and Dev-related products about 10%, together roughly one-quarter; AI-related products total 10-15%, which he concedes is on the optimistic side * The multiplier effect of AI on application count: before mid-2023, a typical enterprise had roughly 100 developers supporting 400 applications; today headcount is up only 5-10% (and skewed to senior engineers) while application count is close to 650, up more than 50%
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。
+ Customers onboarded per engineer went from 1-2 per week to 10-20 + **His day-to-day dashboard and DDOG usage intensity is up 4-5x, and he confirmed telemetry volume collected is up a similar 4-5x** (attributable to broad adoption of AI tools such as Claude and Gemini rather than to DDOG alone) * GPU observability is already in use for AI/ML workloads, monitoring utilization, memory and power draw; AI workflows carry higher observability intensity than traditional services, and what increases is the number of dimensions rather than simply more of the same telemetry
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。
The part that is, like right now, growing the fastest is Datadog for AI and how we observe basically the whole stack. So, a lot of the stack is actually the same as the non-AI. So, most of the AI companies are built on infrastructure that other companies also use. Most of the AI agents are actually spending the majority of their time using tools. The tools that the AI agents are using are standard applications, and those applications also need to be built and monitored. And so, a lot of it is the same. I would say, today, like 70%, 80% of the stack is the same. And then, there's a number of new layers. At the bottom end of the stack, GPUs are a new type of infrastructure that has been around for a long time, but was never really used at large scale in production environments. So, in a way that was going to be repeatable across a large number of customers.
If it can be an incremental growth driver for your infrastructure monitoring, have you seen any evidence of this yet? That's it for me. Thanks. Look, we do see an acceleration of consumption of our infrastructure products in general. And also that's – at a high level, we do see that across the customer base. I don't know that if we see specifically the CPUs that get attached to GPUs in the new build out. I think a lot of it has more to do with the fact that the AI agents are largely spending a good amount of their time, like sometimes the majority of their time calling tools and tools are just, applications that already existed. And those applications typically run on CPUs. And so we see quite a bit of that. Thank you. Our next question comes from the line of Brad Reback with Stifel.
As of Q2, over 750 AI customers use Datadog to monitor and improve their tech stacks. When we look at the largest companies driving AI, all 10 of the top 10 AI leaders are Datadog customers. Beyond AI natives, we see AI activity growing across our broader customer base. We're also seeing signs of rapid growth in agentic activity with a number of MCP tool calls quadrupling again quarter-over-quarter and growing more than 22x when compared to Q4, 2025. Second, we are delivering AI for Datadog to deliver more value and greater platform capability to our customers. This includes our Bits AI products; chat, investigation, detection, code, testing, release and many, many others. Third, next-gen AI introduces new complexity and observability challenges.
Look, I think one thing I'd say is it's hard to tell what we're going to be in four or five years. If you had told me two years ago that most engineers would go back to coding in the console, I wouldn't have believed you. And yet, that's one of the winning modalities today. Look, as far as we're concerned, we don't care whether most of the usage is humans. Most of the usage is agents, our business model lends itself to it pretty well like we're usage based then it doesn't really matter where the user is coming from that perspective. The way we see trends from up right now is we see both a stratospheric increase of agent usage. So, we have a ton of usage on our MCP server. We see customers trying to automate a lot with their own agents, using our agents, using a combination of those. But we also see an increase of usage of the web interfaces by humans. So, right now, the two work hand in hand, and we keep developing and pushing on both fronts.
+ Token usage, LLM calls, experiments, evaluation and prompt versioning all run through BrainTrust, not DDOG + He was explicit that the company has **no conversion formula or fixed budget ratio from AI spend to observability spend**, and no cap on AI-side cost at present * Agentic applications have not yet created a telemetry scale shock — they are small enough that no sampling or offloading of data out of DDOG is required; some non-LLM applications are in fact harder to trace
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。
回 答 者：我们并未专门增加对这些工具的采购，但扩大了围绕AI的各类套餐和使用机会。最初，我们从较为基础的用户类型和模型入手，这些模型在Token方面会产生一定支出。随着使用频率提高并向越来越多的人员开放，我们会评估哪些人员确实需要获得额外的使用额度，例如使用更复杂的模型。
来源背景：未具名专家，现任某保险公司 (金融业) IT 负责人，此前在软件公司工作 (原文 'I worked in software companies and also uh now I'm working in ... the financial sector for an insur...')。视角局限：单一保险企业样本，AI 支出以账号和 Token 为主。
展开本节全部 102 条原文 (支持 76 / 反对 9 / 其他 17，按时间)
- 支持 LLM spans 一个月内翻两倍以上 (四倍多)；MCP server 受关注
In LLM Observability, we recently launched LLM Experiments and Playgrounds for general availability, helping teams to rapidly iterate on LLM applications and AI agents. We also launched custom LLM-as-a-Judge evaluations for General Availability, which lets customers write evaluation prompts to access application quality and safety. As an illustration of both and adoption in the past month, the number of LLM spans customers attending Datadog has more than quadrupled. And we are seeing a lot of interest in the Datadog MCP servers. Our MCP server acts as a bridge between Datadog and AI agents, such as Codex, OpenAI, Claude Anthropic, Cursor, GitHub Copilot, Goose by Block, and many more.
- 反对 GPU 监控尚未产生显著收入，加速与 GPU 无关
Yes. So we have products that are getting into the market now for GPU monitoring. But these don't generate any significant revenue yet. So all the revenues we've shared, like the acceleration, et cetera, that's not related to us capitalizing more on GPUs, that's a future opportunity. .
- 支持 AI 让客户系统更复杂、需要更多监控，同时可观测性本身也在用 AI
A:这是双重驱动：客户总体上正在采用更多 AI，创造出更复杂、融入 AI 的系统需要监控；同时，可观测性 AI 也越来越被采用，因此 AI 在强化这一趋势。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 反对 客户的生成式 AI 应用仍在试验阶段，还没有成为大额开支
A:这与我们在所有生成式 AI 支持平台上看到的类似。客户几乎不可避免地感兴趣，但他们还没有大规模采用。生成式 AI 功能尚未成为巨额开支。它们正在被试验并逐渐被采用，但还没有在客户系统中大规模部署。我认为关注点更多放在其他领域。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 中性 AI 让被监控的应用数量每年多约 10%，这是自上而下估的，没有实证
基于需要被观察和监控的应用数量增加，潜在上它每年可能增加大约 10%，持续几年。我认为那至少是额外的 10%。可能甚至超过 13%，但我不愿以此作为假设。 A:这是一个自上而下的估计，基于在 AI 出现后特定企业中应用数量每年可能增加的程度，相对于之前。它更多是定性的，但提供了方向性视角。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 支持 保险理赔的 agentic 自动化流程上线后，客户的 Datadog 用量增加了
A:我们与使用生成式 AI 平台（底层为 LLM）的保险公司合作，用于自动化理赔流程。这把一个大部分是手工的流程转变为生成式流程，而该流程需要监控。这是增加 Datadog 需求的一个很好的例子。他们已经使用 Datadog，我们为这个代理式自动化流程实施了 Datadog 作为可观测性平台。 它正在提升他们的 Datadog 使用量。这是一个正在进行的推广。我们正在帮助他们在车险、旅游保险等产品领域以及随其进入新地域时在各国推广。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 支持 多步执行的 agent 产生大量日志，客户还在适应
Logging as an observability attribute is assuming more importance than ever. Applications are generating increasingly complex data in the form of logs, particularly where customers have deployed AI, including LLM. Agents that perform multi-step actions generate a huge volume of logs. Customers are still adjusting to this new reality in terms of how to handle it.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 支持 LLM Observability 驱动了部分 4Q 大单；客户每投 100 万美元做 AI 应用，约需 10% 的 DDOG 增量预算
+ **增长非常快，因为企业客户在使用AI应用前需要前置上线该功能以确认数据安全，4Q有些大单是由该功能驱动的** - 少部分案例来看，当客户投入$1m开发一个AI应用（包括token和人力等各种成本），往往需要provision约10%的DDOG增量budget
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 支持 LLM 可观测性带来的初始收入是几笔、每笔几十万美元，规模受限于生产级 LLM 负载
A:LLM 可观测性的规模受限于生产级 LLM 工作负载的规模，所以它推动的初始收入通常为几笔交易，每笔几十万美元。当前这不是一个巨额项目，但随着这些工作负载的规模及其影响力在多种 AI/生成式 AI 解决方案中增加，它有可能成为更大的票额项目。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 口径不齐 AI 应用要有数十万、数百万用户上线后才会带来真正的 LLM 可观测性收入
然后，一旦这些应用上线并有数十万或数百万用户开始使用它们，我们才会真正看到来自 LLM 可观测性的额外收入。这需要规模。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 支持 新 AI 应用总支出中约 8–10% 用于监控和可观测性
在跨所有这些项目的支出中，大约 8–10% 用于新 AI 应用的监控和可观测性，取决于当时市场的优先级。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 支持 MCP tool calls Q4 环比增长 11 倍；LLM Observability spans 6 个月增 10 倍，1000+ 客户
The Datadog MCP server is being used by thousands of customers in preview. Our MCP server responds to AI agent and user prompts and uses real-time production data and rich Datadog context to drive troubleshooting, root cause analysis and automation. And we're seeing explosive growth in MCP usage, with the number of tool calls growing 11-fold in Q4 compared to Q3. Second, let's talk about Datadog for AI. This includes capabilities that deliver end-to-end observability and security across the AI stack.
- 支持 LLM Observability 1000+ 客户，spans 6 个月 10 倍
This includes capabilities that deliver end-to-end observability and security across the AI stack. We are seeing an acceleration in growth for LLM Observability. Over 1,000 customers are using the product and the number of spans since has increased 10x over the last 6 months. In 2025, we broadened the product to better support rapid agent development and iteration adding capabilities such as LLM Experiments and LLM Playground, LLM prompt analysis and custom LLM-as-a- judge. And we will soon release our AI Agents console to monitor usage and adoption of AI agents and coding assistants.
- 支持 CEO 论述：AI 让应用更多、复杂度更高，价值从写代码转到验证/观测
Like people are building much more and they are building much faster. We covered that in previous calls, but we think that the -- this is nothing, but an acceleration of the increase of productivity for developers in general, so you can build a lot faster. As a result, you create a lot more complexity because you build more than you can understand at any point in time. And you move a lot of the value from the act of writing the code, which now you actually don't do yourself anymore to validating, testing, making sure it works in production, making sure it's safe, making sure it interacts well with the rest of the world, with end users, make sure it does what it's supposed to do for the business, which is what we do with observability. So, we see a lot more volume there, and we see that as what we do basically where observability can help.
- 支持 APM 增速到 mid-30%，反映 AI 原生客户对 LLM Monitoring/RUM 的需求和传统客户 AI 应用上量
* APM产品增速达到mid-30%，体现了AI native客户们对LLM Monitoring和Real User Monitoring等产品的强劲需求，以及传统客户们AI应用逐渐上量的趋势 * 我们此前指出，OpenAI尽管完成了续约，但由于Chronosphere的使用体验提升，其对DDOG的用量在2026年仍将下降，部分抵消了core cohort的强劲增长 * 与此同时，其他AI native客户的使用进展良好；公司提到与某大模型公司签下了八位数订单（预计为Anthropic), 尽管该用量短期内无法填补OpenAI流失的空缺，但有助于客户集中度的改善
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 中性 经历 2025 年的试验后，企业开始关注 token 消耗的性价比，要求更严格的成本控制
企业客户在经历了2025年广泛的项目试验后，开始关注Token消耗的性价比问题，并寻求更严格的成本控制。埃森哲提供的服务中包含一个帮助客户约束无效Token浪费的节约框架，该框架与风控体系整合在一起。具体而言，埃森哲会先评估业务的性价比并对 业务优先级进行分类，然后根据场景的复杂程度决定调用资源的级别。例如，对于流程数量、环节和角色超过特定阈值的复杂多级联动场景，埃森哲会建议使用更高级别的计算资源以保证模型精度，同时在训练前通过预设模型精度和控制数据维度来防止无效训练。这项服务 与DataDog等产品的区别在于，埃森哲不仅提供监控工具，更侧重于在项目立项阶段就介入，利用立项规则帮助客户迅速计算出项目的成本和效益比。在项目实际应用过程中，埃森哲倾向于使用自有的工具。虽然埃森哲的一些自有工具在技术门槛上可能与传统的应用 监控工具类似，但它们已经集成到埃森哲的工程化解决方案中。
来源背景：未具名；正文明确是埃森哲的人 ('埃森哲2026财年Q2'、'埃森哲倾向于使用自有的工具')，举的案例 (美孚中国区验证) 和国内模型对比说明主要是中国/亚太视角，职位不详。局限：咨询公司有推销自身方案的利益偏向；Datadog 只是被提问者点名，回答没有具体数据。 - 支持 数字员工操作已有软件的频率远高于人
首先，对于企业已投资的软件（如ServiceNow），数字员工的使用会使其应用频率变得更高。因为数字员工操作软件的效率远高于自然人，能够更快地处理大量订单和任务。例如，在一个日活处理70万张订单的模块中，原先需要几十名员工使用软件，而现在一 个数字员工就能完成十倍以上的工作量。
来源背景：未具名；正文明确是埃森哲的人 ('埃森哲2026财年Q2'、'埃森哲倾向于使用自有的工具')，举的案例 (美孚中国区验证) 和国内模型对比说明主要是中国/亚太视角，职位不详。局限：咨询公司有推销自身方案的利益偏向；Datadog 只是被提问者点名，回答没有具体数据。 - 支持 JPM 联系人称 AI 带来遥测量 '阶跃式' 增长，但增量多用于平台内 AI 功能
The firm highlights a positive channel check pointing to accelerating telemetry ingestion into 1Q, driven by AI-native capabilities embedded in observability platforms, with consumption trending up sequentially from Q4. An industry contact describes AI as driving a “step-function” increase in telemetry volume (analogous to cloud migration), with incremental ingestion largely tied to powering in-platform AI features (vs. third-party workloads). JPM sees this as a positive read-through for Datadog and Dynatrace fundamentals over the next 12–18 months.
来源背景：TMT Breakout (TMTB) 日报，转述 JPM (渠道检查及一位行业联系人) 与 UBS (客户检查) 的报告摘要，二手转述，看不到底稿。 - 支持 AI 使用数字：Bits AI SRE 调查量 12 月→3 月翻倍以上；LLM Observability span 环比近 3 倍；MCP tool calls 环比 4 倍；Bits Assistant 消息 12 倍 (均为增速，无绝对值、无收入)
We now have over 6500 customers sending data for one or more of our AI integrations. Though this is only 20% of total customers, they represent about 80% of our ARR, and our customers' usage of AI within Data platform continues to grow rapidly. Bits AI SRE Agent investigations have more than doubled from December to March. The number of spans sent to our LLM Observability product nearly tripled quarter-over-quarter. The number of Datadog MCP server tool calls quadrupled quarter-over-quarter, and the number of Bits Assistant messages increased by a factor of 12 in that period. While we are aggressively building with and for AI, we also continued to expand the Datadog platform to deliver against our customers' increasingly complex needs. To speak to a few of these efforts, last month we launched experiments for general availability.
- 支持 6,500+ 客户向 AI 集成发数据，占客户数 20%、占 ARR 约 80%
This drives higher GPU ROI and operational reliability. Our customers continued to move forward with their AI activities, and we can see that in their usage of the Datadog platform. We now have over 6500 customers sending data for one or more of our AI integrations. Though this is only 20% of total customers, they represent about 80% of our ARR, and our customers' usage of AI within Data platform continues to grow rapidly. Bits AI SRE Agent investigations have more than doubled from December to March. The number of spans sent to our LLM Observability product nearly tripled quarter-over-quarter.
- 支持 AI 写代码 → 更多应用进入生产，消费出现拐点，AI native 与非 AI 都有 (定性，无量化)
I'm wondering how much of that code is going into production and therefore driving activity for Datadog. Well, we definitely think and see that there's many more applications being created. There's going to be way more complexity in production. We see some of that happening already today. Some of those new applications are getting into production. They are finding users. We see some signs of that at every layer of our platform. We created a few stats on the increasing data volumes we see, AI products. That's definitely a reflection of that. So, we see an inflection point there in consumption from customers. We see a move to production that is very real. And we see that across both AI native and non AI companies. Okay. Thank you.
- 支持 按用量计费，使用者是人还是 agent 无所谓；agent 用量 '平流层式' 增长，人类网页使用也在增加
If you had told me two years ago that most engineers would go back to coding in the console, I wouldn't have believed you. And yet, that's one of the winning modalities today. Look, as far as we're concerned, we don't care whether most of the usage is humans. Most of the usage is agents, our business model lends itself to it pretty well like we're usage based then it doesn't really matter where the user is coming from that perspective. The way we see trends from up right now is we see both a stratospheric increase of agent usage. So, we have a ton of usage on our MCP server. We see customers trying to automate a lot with their own agents, using our agents, using a combination of those. But we also see an increase of usage of the web interfaces by humans.
- 支持 训练从 '不是市场' 变成市场：hyperscaler 的超级智能实验室用 Datadog 监控训练和 GPU
There's only, when you think of who actually has heterogeneous environments today, that is still a very small number of companies, Google, barely just started selling their TPUs, to the outside. So, I think it's still a small number of companies that are there, but we see a growing opportunity there. Interestingly, last year when we reported earnings, we said we're mostly interested in inference workloads and training is not really a market for us yet. Now we actually see training becoming a market. we started landing customers that are actually, hyperscalers that have, a whole host of, homegrown technologies, and that are using us specifically in their superintelligence labs to help monitor their workloads, etc., the training runs, monitor the GPUs also. So, we see that as a point of validation that there's going to be a great market for us.
- 口径不齐 CFO 说训练还早，AI 渗透目前主要来自推理
Yeah. I think going back to the metrics that Oli talked about, in terms of attach, we said that 6,500 customers are using our integrations, and that's 20% of the customers and 80% of the ARR. So, there is attach, I think it's earlier days for the training. It looks like it will be a contributor. But I think, that's early and I would sort of look at the larger attachment at this point as the evidence of inference, but also some training. Thank you both. And congratulations.
- 中性 芯片异构有利于 Datadog，但目前真正异构的公司很少
So, it becomes more of a viable category for service providers. So, they provide like us basically. I think the heterogeneity of the silicon is definitely a trend that plays in our favor there. The more heterogeneous, the more you need someone else to make sense of everything for you and tie it all together and also tie it all with the non-GPU aspects and the rest of the infrastructure and the application and the users and the developers like basically everything we do for you. There's only, when you think of who actually has heterogeneous environments today, that is still a very small number of companies, Google, barely just started selling their TPUs, to the outside. So, I think it's still a small number of companies that are there, but we see a growing opportunity there. Interestingly, last year when we reported earnings, we said we're mostly interested in inference workloads and training is not really a market for us yet.
- 支持 Claude Code 与 OpenAI Codex 都用 DDOG 做遥测 trace，coding agent 可能大幅推高 APM/trace 用量
+ Both Anthropic Claude Code and OpenAI Codex rely on DDOG for telemetry traces, with coding agents likely boosting significant APM/trace layer usage. * DDOG now has 22 AI-native customers with ARR over $1m (up from 19 in 4Q25), including 5 with ARR over $10m. * Most interestingly, DDOG secured two large ARR deals with AI research divisions at leading technology companies, showcasing its technology advantage and expected meaningful revenue contribution in 2H26**.**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 机制：AI 让软件产出更快，开发者不理解自己产出的东西，复杂度上升带来监控需求
Yeah. And so that tells you about the very specific need there is in there. Look, at the end of the day, we see there's so much more stuff that is being built. It's being produced so much faster that, by definition, the folks who produce all of that have no idea it actually works. Now, that's the least understood thing about productivity is that the more you increase productivity, the more complexity you create because folks manipulate way more things in way less time, and as a result, doesn't go through their brains and they don't understand what's going on. So they need help to actually understand what happened, to actually make sure it works properly, to make sure it delivers what it's supposed to deliver for the business in the end, to make sure it keeps working when everything keeps changing around it, to make sure it's secure and that's what we do.
- 支持 训练 (post-training) 开始成为市场：多个超大厂几乎同时找来，但相对推理业务仍小
And then we saw, instead of having 5 to 10 companies doing that, there were 50 to 100. So intuitively, it made sense to us that something might be more interesting there. We were surprised, though, to see a number of different customers, including multiple Hyperscalers, come to us for training at about the same time. Yeah. And which tells us like there's something happening and there's potentially like the emergence of a new market there.
- 口径不齐 训练业务仍很小，未量化
So we are hopeful, interested in that. Yeah. It's not the core, like, it's still small compared to the inference business compared to everything else that happens in the (00:13:34) our customers are running. But it's an interesting new greenshoot, I would say. Yeah.
- 口径不齐 企业客户 MCP 流量“增长很强”但基数小，本场未给数字
Yeah. I mean, look, so we do have some proof points that those customers are starting to adopt some AI in production. It's still small relative to the size of those customers and still a small driver of our growth there like we see traffic, for example, MCPs and traffic also to our (00:19:26) product. We see a very strong growth. I think we released some numbers in the earnings call on that, but that's still a small number. Just give us an idea that these customers are moving there, but the bulk of it for these customers is that they are still modernizing, moving to the cloud. We're getting to more of them because we have successfully built up sales capacity.
- 支持 训练成为新市场：做后训练的公司从 5–10 家增到 50–100 家，多家 hyperscaler 同时来用
*“Instead of having five to ten companies doing that, now there were 50 to 100. We were surprised to see a number of different customers, including multiple hyperscalers, come to us for training at about the same time. That tells us there’s something happening and there’s potentially the emergence of a new market. If we go from 50 to 100 of these companies to 500 to 1,000 to maybe more, then that becomes a really interesting market and there’s a real problem to
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 支持 生产力越高复杂度越高，需要 DDOG 帮助理解系统
Pomel: *“There’s so much more stuff being built. It’s being produced so much faster that, by definition, the folks who produce all of that have no idea how it actually works. That’s the least understood thing about productivity — the more you increase productivity, the more complexity you create, because folks manipulate way more things in way less time, and as a result, it doesn’t go through their brains and they don’t understand what’s going on.”*
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 支持 BofA：AI 推高应用复杂度和企业现代化，带来更强的监控需求
BofA raised PTs on $DDOG ($260 from $225) and $FROG ($85 from $72) following channel checks with Cognizant that pointed to improving demand trends across observability, security, and software supply-chain tooling. The firm said increasing AI-driven application complexity and enterprise modernization are driving stronger need for best-of-breed monitoring/security platforms, with Datadog benefiting from rising observability requirements and JFrog seeing healthy traction around software supply-chain/security workflows. BofA also believes both companies could continue delivering upside to Street estimates given improving execution and resilient infrastructure software spending trends.
来源背景：TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (渠道检查对象为 Cognizant)。二手转述，看不到底稿。 - 支持 非 AI 客户也开始用 AI：MCP 服务流量、LLM Observability 中 agent 产生的 trace 大幅增长 (本场未量化)
So on the AI side, obviously, the AI companies are consuming a lot of infrastructure, building a lot of applications, running a lot of applications. And that's what we see on a day-to-day basis. On the rest of the market, we see also more applications, but we also see some growing adoption of AI itself. So we disclosed in the last quarter that we've seen very large increases in volumes of our AI-specific products whether that's the traffic we see to our MCP services, or whether that's the traces we get into our LLM Observability product that really gets traces produced by AI agent. So we see inflection across all those points. Oli, I think you've done a really nice job kind of helping people kind of understand, like how you get paid and how you kind of ride the cloud wave.
- 支持 机制：coding agent 让工程师几分钟拼出应用却不读代码，生产力提升带来复杂度暴增
What it means when you get higher productivity is that you produce less or you produce more in less time, which means that you don't really get to spend as much time on it. You don't understand what's actually going on. And you see an even bigger acceleration of that today with the coding agents where engineers can, in a few minutes, we put (06:14) an application. They just have no idea what's going on, they probably won't even read the code. But they combine all sorts of different components into an application. So that increase of productivity creates a dramatic increase in complexity. And the problem we saw for our customers is we actually understand that complexity, we manage it for them. And in the future, we want to do even more and automate, basically, the way the applications run after they're integrating them.
- 支持 Datadog 支出相对客户总 IT 支出很小：$1 Datadog 对 $10-20 云、$20-100 工程团队 (管理层估算)
I mean, it doesn't make economic sense for you to do it yourself typically. And the reason for that is this is – what we provide gives enormous leverage. So typically, for like any dollars customer spend on us, they're going to spend $10 or $20 on their cloud provider. They're going to spend $20 to $100 on their engineering team, with the rise of coding agents and AI models and maybe some of that $20 to $100 just spending on their engineering team is actually going to go to these AI models instead. So we're talking about huge amount of leverage. So when the – and by the way, we help our customers make the most to optimize their spend on their cloud provider, on their engineering team, on their AI models and on all of that.
- 支持 agent 用工具，工具就是应用；生产环境监控 agent 本质上是监控应用，不该分开工具
There's one side which is everything that's in production. And when you talk about things that are in production, it doesn't make any sense to separate the AI from the non-AI components. For one thing, what the agents do is, so they have a lot of reasoning, but also they use a lot of tools. And the tools are actually apps. So when you monitor the whole thing, you basically end up monitoring like largely apps with a little bit of AI sprinkle on top. So when you're on production, it doesn't make any sense to have different tooling. And that's something that we develop.
- 支持 AI 应用的遥测量是传统应用的数十倍，按摄入量计费的观测厂商有扩张空间
Um, and if we look at Chronosphere, right? Uh, it's uh priced driven by the telemetry that you're ingesting, collection of metrics, log volume, trace volume. So traditionally application would have 100 gigs of uh telemetry daily. A application would easily have, you know, 4 to 5 terabyte. So that creates a lot of expansion in opportunities. So I think, um, the these companies like, you know, Cyberark, Polo and Chronosphere, their fastest growing products are um, the growth is directly tied to the AI scaling metrics. It it is not tied to the employee counts, right? So historically, you would see as more employees is more licenses, but now going forward, more AI agents, machine identities, data, workload means more revenue for the companies. Okay, got it. So could you share specifically how agents is priced uh by Paulo Alto Lab? Uh for example, is it that that's each agent charge about $10 per month or something like like that? Yeah, I mean, um pricing an agent, um, uh Palo Alto actually does not price AI security as simple like you're saying the $10,000 per month. I think uh in a customer deal, pricing is usually bundled into a platform contract and then tied to uh one or my one or more metrics. So, uh, these metrics examples are number of AI applications, API calls, data volume inspected. The users of AI service, platform consumption.
来源背景：未具名专家，Accenture (埃森哲) 前任人员 (身份由同场 MeritForum 06-04 标题'埃森哲离职专家'推得；本篇无自我介绍)。以'my portfolio'口吻谈客户，推测为咨询/集成侧的安全业务负责人。视角局限：咨询商视角，只见自己客户组合；主题是 Palo Alto/CyberArk 安全，DDOG 为顺带；Chronosphere 数字为举例性量级。 - 支持 AI 工作负载产生大量日志/事件/指标，客户遥测量大幅增长
Chronosphere 我认为是一个间接的 AI 赢家，因为它本身并非严格意义上的网络安全公司，其业务领域主要在可观测性、监控、遥测和云原生运营。AI 会产生大量的日志、事件、指标等数据，运行 AI 工作负载的客户会发现其遥测数据量大幅增长。因此，随着 AI 部署规模的扩大，Chronosphere 的价值主张变得更具吸引力，因为它能降低客户的可观测性成本，提高遥测效率，并提供对 AI 工作负载的全面可见性。
来源背景：未具名专家，Accenture (埃森哲) 前任人员 (身份由同场 MeritForum 06-04 标题'埃森哲离职专家'推得；本篇无自我介绍)。以'my portfolio'口吻谈客户，推测为咨询/集成侧的安全业务负责人。视角局限：咨询商视角，只见自己客户组合；主题是 Palo Alto/CyberArk 安全，DDOG 为顺带；Chronosphere 数字为举例性量级。 - 反对 非模型公司的 token 消耗量还不大，幻觉等 AI 监控还没真正用起来
当然，如果用 AI 相关的功能，费用肯定会上升。但目前新出的 AI 功能，比如 token 相关的监控，公司毕竟不是大模型公司，token 的消耗量还不算多，幻觉等监控暂时还没有完全用起来。未来可能会用，但目前公司还是鼓励大家多用 AI，还没到需要重点投入监控的阶段。等到那个阶段，肯定会发展起来。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 支持 BNP：多位客户把 AI 视为可观测性支出增加的催化剂而非替代
Following DASH 2026 and discussions with customers, BNP Paribas came away more constructive on DDOG’s growth durability, noting customer enthusiasm was stronger than expected across observability, security, and AI-related offerings. The firm highlighted Bits AI, AI Guard, Database Monitoring, Network Monitoring, and Query Optimization as products generating particularly strong interest, with several customers viewing AI as a catalyst for increased observability spend rather than a substitute for it. Customer conversations also suggested Datadog is benefiting from growing demand for AI workload monitoring, application tracing, security safeguards, and full-stack visibility across increasingly complex cloud environments.
来源背景：TMT Breakout (TMTB) 日报，转述 BNP Paribas DASH 2026 后客户调研摘要。二手转述，看不到底稿。 - 支持 AI 工作负载产生更多遥测：GPU、推理追踪、向量库、agent 编排层
From what I have observed, the AI native customers see data dog consumption probably growth in somewhere from 35 to more than 50% year-over-year. Whereas, uh, if you're looking at traditional customers, you know, they are growing in 10 to 20% range. I think the gap is driven by um, very high, uh, telemetry generation from the AI workload, which includes stuff like GPU, infra monitoring, model inference tracking, vector DBs, agent orchestration layers and the need, you know, for end-to-end observability across the application. application stacks which are becoming more and more complex. And having said that, the exact split uh does vary by customer maturity.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 支持 AI 原生公司 70–90% 已部署 Datadog AI 可观测性，传统客户 20–40%；部署集中在推理而非训练
And do you think in the future they there will be more configuring the observation layer? Yeah, um I think the adoption rate is way higher, like I said, among the AI native companies. I would say probably 70 to 90% of those companies have deployed Datadog's AI observability capabilities. Um compared to around 20 to 40% of traditional customers who are still experimenting with AI use cases. Um most of the current deployments are focused on inference rather than training. Because inference is where production reliability, latency, cost, model quality, and the user experience become very business critical. Training workloads um are typically monitored using a combination of cloud native tools, um ML platform tools and infrastructure monitoring, whereas Datadog's value proposition is stronger around production operations, model serving, agent workflows and end user interactions.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 支持 若 AI 带来的运营复杂度增长快于客户管理能力，Datadog 可能不成比例受益；但长期价值排序 Snowflake/Databricks 第一、Datadog 第二
The company that controls the data, governance, lineage, retrieval, I think they occupy the most defensible position. However, um I still see a one important nuance. Um, DataDog could benefit disproportionately, you know, if AI adoption drives operational complexity faster than customers can manage it. A company might deploy thousands of agent across customer service, software development, cybersecurity, business operations. At that scale, understanding what each agent is doing, how much it cost, and whether it is performing things correctly and where failures happen is a big challenge.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 支持 AI 工作负载 '吃遥测'：每次模型调用、每个 agent 步骤、每个 GPU 小时都产生需监控的 trace 和指标
But I would say say 40% for usage of of of observability. Uh and then, uh I think a if you look at AI and AI workloads, they are actually increasing the need for observability. So so AI workloads are very sort of telemetry hungry in a way traditional apps are not. So every model call you do, every agent step, every GPU hour generates sort of traces and metrics that you that you need to monitor, right? And that's that's the uh basically the story behind observability. So so as enterprise customer adopt AI more and more, uh you know, the need for observability will also increase, right? And we saw that with uh you know, um their dash 2026 uh sort of releases, uh you know, where they have uh they're highlighting out of the uh the sort of the new capabilities that Data Dog came out with. A lot of them resolves around uh you know, bits AI for example, agent console, AI guard, etcetera, right? So I think that's um that that shows the direction. Uh and they have also, I think if you look at data dog themselves, they say that, you know, AI integrated customers are driving a lot of the growth, right? So I think that's that's probably also where you can find some uh uh you know, tidbits of information uh to complement what I you said. 嗯。 OK,因为专家这边的客户都是企业级客户,所以没有什么AI native的公司,然后企业级客户里面40%的客户,他猜的就是40%左右的客户现在都在用这些关注性的工具,然后他说除了这些40%的客户之外,就是剩下的有一些客户其实也接触到了就是观测相关的一些功能吧,因为有一些CSP它其实也有类似的这个产品嘛。 然后他说这个随着AI就是workload的增加的话,其实企业对于观测工具都是有需求的,因为这些就是API的调用啊,然后agent呀,还有GPU小时啊等等,就这些都是一些遥测的数据嘛,然后都涉及到一些指标啊,还有一些链路的观测,所以他觉得在未来就是随着企业越来越多的使用AI,这个需求也会越来越旺盛。
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 支持 AI 增长对 Datadog 的弹性约 1.1，Datadog 应比底层 AI 负载增长更快 (多云、安全也在叠加)
So the AI, the growth in AI workload is also is also means that there's a growth in typically in multi cloud for example, which is also uh positive for uh for Data Dog. Uh and there's also uh, you know, security aspects coming from AI and so on, right? Um, and uh so I would actually say maybe a sort of uh if I put it the linear growth would be one, I would I would arguably say it's like 1.1 or something. So so uh you know, Data Dog should be able to grow, you know, faster than the underlying AI workload growth. Mhm. Okay.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 支持 约 50% 企业客户已有少量 agent 上生产 (通常少于 20 个)，正在加速规模化
And and that that I think is is a good bet because there will be uh, you know, a uh uh uh uh. vast number of agents, you know, running and performing autonomous actions, you know, in the in the middle and back office of enterprises. And it's still early days, but but I think it's like when we see agents, AI agents, it's like maybe 50% of our enterprise customers have a few agents in production. It's usually less than 20, but it's uh it's actually being adopted now increasingly at scale. Uh and then this centralized monitoring will be uh important to to keep track of what's happening, uh what's uh how are they doing, etc. And of course, if we look at uh AI assisted software development, like, you know, the the the coding to platforms.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 反对 token 成本长期会大幅下降，专家不确定大规模 token 支出 5–10 年后还存在
我也观察到 token 成本下降是一个长期趋势。使用量在上升，所以存在张力。我不确定最终如何发展，但随着足够的数据中心建设，token 成本可能会大幅下降。后端更高效的算法也可能显著改变局面。这有点像早期的互联网流量：你不能压缩数据包，所以需要铺设很多光纤，然后几乎一夜之间发明了能将流量压缩百倍的算法。我并不是说这里会完全相同，但我并不完全确定大规模 token 支出的故事会在五到十年后仍然存在。
来源背景：标题写明 '北美软件渠道商离职专家' ，作者 Lantian；内容与 Merit Datadog 产品进展调研 II (2026-06-17 音频) 逐段对应，判断为同一专家同一场。专家自述 '我们主要与企业客户合作……后者通常不需要系统集成商的支持' ，实为系统集成商视角；'离职' 与原话现在时矛盾。 - 支持 agent 带来大量复杂度，日志、指标、整体用量爆炸 (未量化)
Yeah, there is. Um, because I think what the AI um agents are are creating at the moment is an an enormous amount of um complexity within the environments. Um, whether it is, um, you know, through the entire infrastructure, I think we're we're seeing a lot more uh logs, uh matrix, um, we're seeing a lot more um, higher usage overall, I would say, more data just essentially the the data is just exploding within the environments. Um, so being able to capture um, the usage, again, all all the matrix and being able to unify this data, um, is so important. But from an AI perspective, this is really accelerating in terms of the data usage, um, and the activities within the environment.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 支持 存量客户的 AI 日志和安全日志增加，并且在向单一平台整合
A:会的。第一季度大约是 20%，第二季度确实加速了。我们在新客户和现有客户中都看到了增长，所以加速来自两方面。例如，在现有客户中我们看到了 AI 和安全日志的增加。我们也看到客户倾向于在一个单一平台上进行整合，而不是在环境中使用许多不同的平台，因为这能为他们产生节省。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 AI agent 让基础设施更复杂，日志和指标用量大幅上升
A:是的，二者之间存在关系。AI 代理在整个基础设施中创造了大量的复杂性。我们也看到日志和指标的使用大幅上升——总体上数据更多了。环境中的数据正在爆炸。捕获所有度量并统一数据至关重要。从 AI 的角度看，数据使用确实在加速，环境内的活动也在增加。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 部署的 agent 越多成本越高；他估计 agent 至少带来 20% 的增长，定价机制还不清楚
A:挑战在于你部署的代理越多，成本可能就越高。Datadog 在商业条款和定价上竞争力很强，目前在这方面击败了 Dynatrace 和其他竞争对手。它肯定会扩大机会；我们估计代理带来至少 20% 的增长。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 80–85% 的企业环境已经有可观测性，下一步的增长靠 agent 产生的流量
A:这是一个巨大的机会。作为一个组织，我们的全部重点都放在 AI 能力和解决方案上；增长基于 AI。如果你看可观测性市场，大约 80–85% 的企业环境中已经以某种形式存在可观测性。重点并不是完全赢得没有可观测性的全新客户；下一个挑战是流量。 这些流量将由代理产生。基本上，它们的增长和受益是因为 AI 与代理。我认为 Datadog 和其他竞争对手绝对会从中受益。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 传统客户用 APM 监控 AI 构建/接入 AI 的新应用，这部分公司不单独追踪，AI 对增长的实际拉动被低估
But uh, uh for the for the growth itself, the contribution is significant and the contribution is actually uh increasing for uh uh AI native companies specifically, right? Given they're going through a massive uh user growth as well, if you look at the uh major AI native companies right now. And in addition, there's also the fact that a lot of companies are increasingly using their traditional products like application performance monitoring in order to monitor applications that they might have created with uh with uh uh AI or created using AI models or which have AI models in the backend, right? So, given a lot of new applications are also being written as a result of AI and using the AI models, that is also driving uh further acceleration of the revenue, which the company is not tracking separately. Like they're tracking the revenue that they're getting from uh the AI company cohort, the AI native cohort companies, but uh you know, they they're not specifically uh monitoring the revenue that they're getting from traditional products like application performance monitoring, but which are driven by, let's say applications that are actually uh with uh you know, the the uh fundamental AI models in the uh in the backend. Hmm, okay.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 AI 让企业写更多软件，就要监控更多基础设施和应用，这已是过去几个季度的大驱动，今明两年会持续
So uh, it's some of the revenue revenue growth that Data Dog has seen, as we already discussed, is driven by these factors. The fact that there is increased usage of AI by uh, you know, every every single company uh in order to make new applications and also accelerate the production of any applications in general uh using coding use insights. So uh the pace at which you write software fundamentally has changed and uh uh data dog is definitely seeing the benefits of that because as people write more software, they're also looking to monitor more infrastructure, more uh uh applications, right? So uh this this is already been a big driver of the growth over the last couple of quarters uh for data dog and it will only continue to accelerate in the in the near term, right? So you will you will continue to see some of the momentum, the growth momentum that you've already been seeing associated with Data Dog revenue and 100% it continues through this year and the next one uh as the quality of the models increases, so there's less hallucination. So uh that that gives people more confidence to put AI in front of the customer as well. And uh uh, yeah, just more broadly. Uh, OK. 他说就是written which can then be monitored by。
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 AI 让老产品 APM 的需求显著上升 (AI 应用增加复杂度)
So there's also a lot of uh attention on these replacement programs for companies that uh are not in a good shape but hold a significant market share in the observability uh market, right? So sure, there are some observability products like uh GPU monitoring, LLM monitoring with AI uh, which which uh have a lot of focus, but there are also these pieces. Also, I want to just clarify that, you know, one of the biggest unlocks from the whole AI frenzy is also application performance monitoring. Because even though the product has always existed, if you have a product which you have deployed and is facing customers and is using AI, that's an additional layer of complication. So, you're going to want to uh do application performance monitoring for that specific product.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 企业的业务日志全量存储和分析一般用 Splunk/Datadog，Cloudflare 只是其中一个分片 (通过 log push 推送)
Cloudflare覆盖领域和他们不是特别一致，Cloudflare的日志主要是代码端运行日志，包括debug日志，需要观测能力。实际应用业务的日志存储和分析一般用Splunk或Datadog，Cloudflare只是他们日志中的一个分片。客户会用这些工具做全量日志保存、中心化分析、调度等，但Cloudflare只关注SaaS服务这一块。Cloudflare能提供的是Workers日志，包括最原始的代码运行日志、错误日志、访问日志、安全日志（哪些被拦住、哪些放行）、AI gateway的大模型调用和费用、Cloudflare运维（配置变更、访问记录）等，通过log push连接外部系统，把这些日志推送到Splunk、Datadog等。
来源背景：未具名；用 Cloudflare 内部视角回答 ('Cloudflare内部'、'我们能看到merge request的数量')，推测为 Cloudflare 现任员工，职位不详。局限：对 Datadog 只是顺带一提。 - 中性 agentic 开发让 MR 数几乎翻倍，代码要经过多个 agent review
主要的内部场景第一就是代码相关，比如代码生成、代码修改，写代码这块的工作量大幅增加。我们能看到merge request的数量几乎翻了一倍。
来源背景：未具名；用 Cloudflare 内部视角回答 ('Cloudflare内部'、'我们能看到merge request的数量')，推测为 Cloudflare 现任员工，职位不详。局限：对 Datadog 只是顺带一提。 - 支持 增长最快的是日志分析：要记录的东西更多、更深更广
If we think about the subsegments of observability, like application performance management... always a key focus area for myself and my customers... as well as infrastructure monitoring, I would say, not dissimilar to what I said before, that even last year and this year, the strongest growth that we're seeing is in the area of log analytics. Semi-structured, unstructured data observability is a core focus for customers. It's the area that's growing the most.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 企业用前沿模型 API 增强应用，会产生额外的 trace 和日志遥测
Sid Roy: So Agentic, when we think about Agentic, I know in the last few quarters we've talked about, it's coming, it's coming, but we're now at that point where there are 2 aspects to it. So you have customers that are now augmenting their applications and systems in particular leveraging the frontier model APIs. And as they augment their application using an API, that basically in essence is generating additional telemetry. And when we think about Dynatrace, the telemetry is basically coming in through traces as well as logging. So that's going to be expanding the overall sales growth as well. From the... so you've got... we're monitoring the applications that are now AI-enabled. On top of that, we also need to actually monitor the models themselves. So things like, for instance, token usage and performance, response times, safe AI, things of that nature. So all of these things are starting to become very relevant. Customers have allocated budget for this in terms of a governance structure. It's critical for these organizations to be able to, in particular around logging, if we think about an AI-enabled application, tracing is probably the most core component, but we think about it from a governance and a performance and a model performance. From a human standpoint, it would be the logging aspect associated with it. Hence, that's why we have a very robust forecast for Grail for logs in the second half.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 最普遍的新场景是部署 MCP 服务器做 agent 间通信，需要监控其性能
Sid Roy: It's growing. We definitely haven't hit our stride yet. But when you talk to the average enterprise, what are the things that are important for them from an AI monitoring standpoint? The concept of being able to monitor just the basic agent, being the agent, the frontier models. For organizations that are deploying coding assistance or co-worker assistance, being able to monitor those interactions. What I would say probably the most prevalent thing that we're seeing is the deployment of MCP servers, for agent-to-agent communication. So monitoring the performance of the MCP servers. For organizations, and I'm sure everyone has heard this, that especially for large organizations, when they look at their cost-benefit dynamic around, where do we host our models? Do we host it in the cloud? Do we put it in something like an AWS Bedrock or an Azure? Or do we actually just bring that in-house, so that we have lower costs for compute, CPU, GPU compute, and storage as well, these are also things that need to be monitored and will, in essence, drive up observability revenues. In addition to that, what are the interactions with external services that support the AI modeling? So like Git operations, GitHub, GitLab, all of these things like when you're commingling together, plus with the things I mentioned earlier around just the basic interactions of an AI-enabled application and then base monitoring for AI governance, all of these things pulled together, it's a relatively bright story for the observability vendors at this point.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 口径不齐 AI 观测在增长但 '还没进入状态'
Sid Roy: It's growing. We definitely haven't hit our stride yet. But when you talk to the average enterprise, what are the things that are important for them from an AI monitoring standpoint? The concept of being able to monitor just the basic agent, being the agent, the frontier models. For organizations that are deploying coding assistance or co-worker assistance, being able to monitor those interactions. What I would say probably the most prevalent thing that we're seeing is the deployment of MCP servers, for agent-to-agent communication. So monitoring the performance of the MCP servers. For organizations, and I'm sure everyone has heard this, that especially for large organizations, when they look at their cost-benefit dynamic around, where do we host our models? Do we host it in the cloud? Do we put it in something like an AWS Bedrock or an Azure? Or do we actually just bring that in-house, so that we have lower costs for compute, CPU, GPU compute, and storage as well, these are also things that need to be monitored and will, in essence, drive up observability revenues. In addition to that, what are the interactions with external services that support the AI modeling? So like Git operations, GitHub, GitLab, all of these things like when you're commingling together, plus with the things I mentioned earlier around just the basic interactions of an AI-enabled application and then base monitoring for AI governance, all of these things pulled together, it's a relatively bright story for the observability vendors at this point.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 BTIG：AI 推理是最明确的需求驱动，利好基础设施和日志监控、LLM 可观测性、网络性能监控
BTIG reiterates Buy on both names and raises DDOG’s price target to $289 from $255 after partner checks showed broad acceleration, with its largest contact seeing growth improve to roughly 35% in 2Q from 32% in 1Q and about 20% in 2025. BTIG says AI inference is the clearest demand driver, benefiting infrastructure and log monitoring, LLM observability and network performance monitoring, while consolidation away from SolarWinds is adding further upside. BTIG also says DT checks were consistently positive, with partners expecting high-teens to 20% growth as AI expands observability demand, supporting its Buy rating and $47 price target.
来源背景：TMT Breakout (TMTB) 日报，转述 BTIG (伙伴检查) 与 Morgan Stanley (企业检查) 报告摘要。二手转述，看不到底稿。 - 支持 BofA：近期 OpenAI 事件强化企业对 agent 可观测性的需求
BofA keeps DDOG as a top 2Q26 pick, saying its AWS revenue analysis suggests current estimates are conservative and leaves room for another acceleration from the 32% y/y growth reported in 1Q. BofA says additional frontier-lab wins, including two hyperscaler AI-research deals and five $10M+ AI customers, diversify the customer base and make renewal volatility less concerning. BofA also argues the recent OpenAI incident reinforces the need for agent observability as enterprises require greater visibility into autonomous AI behavior, expanding DDOG’s long-term opportunity.
来源背景：TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (AWS 收入分析)。二手转述，看不到底稿。 - 支持 AI 网关可与 Datadog 集成导出遥测/日志/链路，使用网关会带动观测类采购
回 答 者：我认为会。现在许多 AI 网关都提供相关集成。例如，它们可以与 Datadog 集成，用于导出遥测数据、日志和链路追踪信息。因此，企业使用 AI 网关后，自然会增加对这类解决方案的采购。
来源背景：未具名专家，LiteLLM 前任人员 (由同场 MeritForum 07-31 标题'LiteLLM 离职专家'推得)。视角局限：AI 网关/路由厂商视角，对可观测性采购为推测性判断，没有具体账单数据。 - 支持 AI 网关与 Datadog 集成，带动观测采购
A:我认为会，因为许多 AI 网关现在都有集成——例如与 Datadog 的遥测、日志和追踪导出集成。使用 AI 网关自然会导致更多此类解决方案的采购。
来源背景：LiteLLM 离职专家 (标题明示)，具体职位未说明。同 Merit 07-31 那篇。 - 支持 AI 让应用数量变多：开发人数只增 5–10%，应用数从约 400 增到约 650 (+50%)
+ Customers previously running 2-3 DDOG products are moving to at least 5; the land products are still Infra and APM, then Logs, followed by Synthetics and Cloud Security + The buying center is no longer just the CIO/CISO — SRE, QA, DevOps and platform engineering are now decision-makers, pulling through Synthetic Monitoring, CI Visibility and Error Tracking * 2H growth mix: traditional products (applications/infrastructure/logs) account for at least two-thirds; security contributes at least 15% and Dev-related products about 10%, together roughly one-quarter; AI-related products total 10-15%, which he concedes is on the optimistic side * The multiplier effect of AI on application count: before mid-2023, a typical enterprise had roughly 100 developers supporting 400 applications; today headcount is up only 5-10% (and skewed to senior engineers) while application count is close to 650, up more than 50%
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 主要驱动是 agent：客户 agent 消耗仅约 10% 已带来云支出跃升
+ **The main driver is agents** — on a 0-100 scale, customer agent consumption is still only around 10%, yet it has already produced a step-change in cloud spend; customers cannot control that consumption themselves and their tooling is fragmented, which is why demand for standardized observability is so strong + Banking and high-tech are the two fastest-moving industries on AI investment, but most customers still lack the governance framework to support AI in production, so the real ramp is a next-year story * **DDOG net-new customer count is up about 30% y/y versus 20-25% for Dynatrace**; DDOG is stronger in mid-market, and its enterprise new-logo opportunity is widening
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 客户 E 2026 DDOG 支出预计 +30%，约 15pp 来自 AI 相关、15pp 来自存量应用
* **2026 DDOG spend is expected to grow 30%, split roughly 15 percentage points from AI-related usage and 15 points from existing and legacy applications**
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 口径不齐 AI 带来的 DDOG 增长几乎全在应用层，不在模型层；token/LLM 调用等走 BrainTrust
* An important nuance: **essentially all of the AI-driven DDOG growth comes from the application layer rather than the model layer**
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 agentic 应用尚未造成遥测规模冲击，规模小到不需要采样
+ Token usage, LLM calls, experiments, evaluation and prompt versioning all run through BrainTrust, not DDOG + He was explicit that the company has **no conversion formula or fixed budget ratio from AI spend to observability spend**, and no cap on AI-side cost at present * Agentic applications have not yet created a telemetry scale shock — they are small enough that no sampling or offloading of data out of DDOG is required; some non-LLM applications are in fact harder to trace
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 客户 F 用 Claude 自动化后 dashboard 和 DDOG 使用强度升 4–5 倍，遥测量也升 4–5 倍
+ Customers onboarded per engineer went from 1-2 per week to 10-20 + **His day-to-day dashboard and DDOG usage intensity is up 4-5x, and he confirmed telemetry volume collected is up a similar 4-5x** (attributable to broad adoption of AI tools such as Claude and Gemini rather than to DDOG alone) * GPU observability is already in use for AI/ML workloads, monitoring utilization, memory and power draw; AI workflows carry higher observability intensity than traditional services, and what increases is the number of dimensions rather than simply more of the same telemetry
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 模型准确度提升后，调试和修 prompt 的近期需求反而下降
+ Risk on the other side: as model accuracy improves, near-term demand for debugging and prompt correction actually falls; what genuinely benefits observability is adoption at scale rather than better models per se
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 MCP tool calls 环比再翻 4 倍，较 2025Q4 增长 22 倍以上
When we look at the largest companies driving AI, all 10 of the top 10 AI leaders are Datadog customers. Beyond AI natives, we see AI activity growing across our broader customer base. We're also seeing signs of rapid growth in agentic activity with a number of MCP tool calls quadrupling again quarter-over-quarter and growing more than 22x when compared to Q4, 2025. Second, we are delivering AI for Datadog to deliver more value and greater platform capability to our customers. This includes our Bits AI products; chat, investigation, detection, code, testing, release and many, many others.
- 支持 LLM 和 tool call 流量 '爆炸'，过去两季 MCP 端点调用激增 (定性)
So it's a very positive trend there. That's the first thing we see. The second thing we see is a very rapid increase in the usage of all of our AI-first services, so that would be the products that measure agents and LLMs. We see an explosion of traffic in terms of the LLM and tool calls we're getting. That would be the amount of calls we're getting to our MCP endpoints. So we see that explode completely over the past two quarters.
- 支持 agent 大量时间在调用工具，工具就是已存在的、跑在 CPU 上的应用，带动基础设施产品消费加速
That's it for me. Thanks. Look, we do see an acceleration of consumption of our infrastructure products in general. And also that's – at a high level, we do see that across the customer base. I don't know that if we see specifically the CPUs that get attached to GPUs in the new build out. I think a lot of it has more to do with the fact that the AI agents are largely spending a good amount of their time, like sometimes the majority of their time calling tools and tools are just, applications that already existed. And those applications typically run on CPUs. And so we see quite a bit of that. Thank you.
- 支持 Bits AI 不会减少用量：用 Bits 的客户用更多产品、建更多 dashboard/告警、更多用户 (定性，无数据)
And we're super confident that this will yield great business outcomes for us in the end. And that's what we see from customers in the market. Like when they use Bits AI, they use more of our product, they deploy more of it, they create more dashboards and alerts and everything else. They have more users inside of our product, like, it's not a zero-sum game. Thank you. Thank you.
- 口径不齐 访谈：AI 加速工程流程使可观测性需求上升，但 Grafana 根深蒂固阻碍 DDOG 整合
Observability demand is rising as AI accelerates engineering workflows, while entrenched Grafana usage complicates Datadog’s platform consolidation push.
来源背景：FundaAI 周报：DDOG 段落为 2Q review 摘要 + Funda 当周两次专家访谈的一句话摘要 (某全球云通信平台的工程高管——疑似 Twilio，因标题标 TWLO；某全球金融信息服务商经理)。二手转述，看不到底稿，访谈全文在 Funda 平台未附。 - 口径不齐 agent 监控和 MCP 调用增长很快，但本场未给数字，仍早期
We're starting to do more in training. So, we basically set that up and we've been seeing very good growth in that area. So, we have to do the investment to set it up and then as clients introduce them to production environments, there are a lot of metrics we've been giving out on the growth of agent monitoring, the growth of MCP calls, lots of metrics. If you read the scripts, etcetera, you'll see these things are growing at a very high rate. It's still early on. So, if they're training models and they're not putting in production and it's in-house or they're using it for their marketing collateral, internal, it doesn't tend to be our market, but more and more of it is being our market. So, we're monetizing it through pricing.
- 支持 AI 让技术运营全面提速，带来更多复杂度和遥测
Then there's obviously the AI trend. Every company is racing to build with AI so that they can scale faster or be more differentiated and be ahead of their competition. I think that just speeds up everything across the technical operations side of things. It's using more cloud resources or new types of resources like GPUs. It just accelerates the application development timeline from development to production with agents writing code themselves on behalf of developers and taking actions on behalf of developers and then also taking actions as they run in production, the agents that the companies build. I think that just creates a ton more complexity, a ton more telemetry that is hard to sift through and can obviously potentially just create more errors and more slowness and more issues and higher costs. Datadog sits on top of all of that, can see all of that and correlate all of that so that companies can move their products and run them confidently.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 反对 AI 写出更干净的代码可能让传统 APM 的问题变少，但 agent 会带来新问题
**EE** [00:32:27]: That's a great question. Maybe. I mean we need to see how good AI is. I'm sure AI is just going to get better and better. I think today, at least I'm following a lot of developers talking about how they use AI and all that. I think they are still struggling in terms of how to use it in the right way. It's still creating its own bugs and issues and doesn't always understand what they want to achieve. I don't think it's perfect. Issues are still happening and going to happen and maybe in other types or forms. I think also the AI that's running in production for companies is just going to generate other types of issues. It's calling tools. It's doing a bunch of things that no one really knows autonomously and no one has visibility. Maybe APM, traditional monitoring, will see less problems, maybe new problems will arise, but at least from the agent monitoring or agent performance monitoring side of things, I think we'll continue to see efficiency issues because a lot of these things are just unpredictable, and I think there's life of their own.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 反对 传统企业 (保险) 部署 AI 并未专门增加 Datadog 等观测/数据库采购
回 答 者：我们并未专门增加对这些工具的采购，但扩大了围绕AI的各类套餐和使用机会。最初，我们从较为基础的用户类型和模型入手，这些模型在Token方面会产生一定支出。随着使用频率提高并向越来越多的人员开放，我们会评估哪些人员确实需要获得额外的使用额度，例如使用更复杂的模型。
来源背景：未具名专家，现任某保险公司 (金融业) IT 负责人，此前在软件公司工作 (原文 'I worked in software companies and also uh now I'm working in ... the financial sector for an insur...')。视角局限：单一保险企业样本，AI 支出以账号和 Token 为主。 - 支持 Agent 的每一步调用、延迟、错误、重试都要记录，Datadog 最像'Agent 电表'
Agent 运行以后，每一步调用、延迟、错误、重试和工具结果都需要被记录。Q2 2026 Datadog 收入 11.2 亿美元，同比增长 36%；年化收入超过 10 万美元的客户达到 4,720 家，自由现金流 2.79 亿美元。公司同时推出 Bits Code、Bits Chat、Agent Builder、自治 Bits AI 与 AI Guard。[5](https://investors.datadoghq.com/news-releases/news-release-details/datadog-announces-second-quarter-2026-financial-results/)公开电话会转录显示，31
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 - 反对 厂商的收费单位不等于客户的经济分母，机器动作多不等于有效结果多
1、第八篇找到了成本的新重心，本篇开始计算利润落在哪里。 Token 越便宜，数据接入、权限、业务规则、验证和写回等非 Token 成本越重要；本篇进一步用 ARR、ACV、自由现金流和 EV/Sales，检验这些成本能否变成上市公司的收入和估值。 2、同一笔非 Token 收费，需要同时进入供应商收入并改善客户财务结果。 在软件公司账上，它要进入净新增 ARR、毛利和自由现金流；在客户账上，它要减少人工、外包和事故损失，最终改善增量 EBITDA、自由现金流或项目回收期。如果只进入供应商收入，耐久性仍待验证。 3、厂商的收费单位，与客户的经济分母不是一回事。 信用点、工作单元、受保护身份和数据用量可以用来出账；企业最终计算的是每个被业务接受结果的总成本。机器动作越多，未必代表客户得到的有效结果越多。 4、非 Token 成本成为平台收入，需要增加一项客户价值条件。 除了跨客户复用、持续发生、准确计量和卡住控制点，产品还要降低客户每个成功结果的总成本，或者提高结果价值。否则这张账单更像暂时的 AI 税。 5、受益方向已经可以落到公司，但价值创造、商业兑现与投资赔率要分别排序。 Palo Alto Networks、CrowdStrike、Datadog 的供应商侧财务证据走在前面；Okta、Snowflake、HubSpot 分别卡住身份、企业上下文和业务结果。下一步还要看客户侧效率改善，能否支持这些收入续费和扩围。
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 - 支持 agent 大部分时间在调用工具，工具是标准应用，70-80% 技术栈与非 AI 相同
So, a lot of the stack is actually the same as the non-AI. So, most of the AI companies are built on infrastructure that other companies also use. Most of the AI agents are actually spending the majority of their time using tools. The tools that the AI agents are using are standard applications, and those applications also need to be built and monitored. And so, a lot of it is the same. I would say, today, like 70%, 80% of the stack is the same. And then, there's a number of new layers. At the bottom end of the stack, GPUs are a new type of infrastructure that has been around for a long time, but was never really used at large scale in production environments.
- 口径不齐 agent 带来的 trace 流量“爆发”，但 CEO 说这只指向未来，还不是当前主要驱动；增长最快的是基础设施和应用层
So, we see all that. This – today, the part of the business growing the fastest is the part that has to do with the build-out at the infrastructure and the application layer. We also see an explosion of traffic in all of the AI surfaces like – so the agents and the number of traces we get from that and things like that. But I think it's more of a – it's showing us where the world is going, rather than already today being the driving part of the business. Bits AI has been a fairly momentous set of product introductions. That portfolio of capabilities has expanded very swiftly.
- 支持 全面渗透时客户在 Datadog 上花云账单的 10-20%，CEO 认为 AI 时代会类似
Our job is to help customers deal with that complexity. And so, that's why they're going to spend in the end of the day. In the cloud, when we're fully penetrated (00:20:38) with the customer, they spend between, let's call it, around 10%, maybe between 10% and 20% of their cloud bill on their – on observability and security and everything else with us. And I think some version of that would remain true in the AI world, where the spend on us is going to be this percentage of their total bill that they actually justify that pays for itself basically, not only in the savings you make from the usage of the underlying compute or applications you pay for, but also in your ability to deliver the outcomes you want and their trust in the quality of the outcomes you can deliver. So, we think that this is going to remain true and this is why this is such a great opportunity in the long term.
- 支持 Agent 工作流的调用树和上下文远比传统服务深、密，MCP 之间互相调用
JC: I don't know. I think the same concepts certainly apply. I think what is different is just the explosion of context and the tree of calls that are happening, you know, for a traditional backend service, it's reaching out to a couple microservices, maybe a couple of APIs or something like that, in a relatively generally shallow fashion. Whereas for these agentic workflows, there's many MCPs that are hooked up. They can all be calling each other, calling things that the end user won't even know about or realize. So kind of like that call stack can be much larger and denser.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 支持 Agent 场景下 MCP 互相调用、调用栈更大更密，意味着更多 trace/span
JC: I don't know. I think the same concepts certainly apply. I think what is different is just the explosion of context and the tree of calls that are happening, you know, for a traditional backend service, it's reaching out to a couple microservices, maybe a couple of APIs or something like that, in a relatively generally shallow fashion. Whereas for these agentic workflows, there's many MCPs that are hooked up. They can all be calling each other, calling things that the end user won't even know about or realize. So kind of like that call stack can be much larger and denser.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 口径不齐 新工作负载上线的速度可能无法维持过去 1–2 个季度的节奏 (取决于 token 成本)
JC: I mean, by definition, it's tied to the AI spending cycle, but I don't expect it to go away. There are already a huge amount of AI and LLM-powered workflows that have been put into production and they need to be monitored. As long as they are in production and serving production traffic, I think, depending on what happens in terms of the cost of tokens over time and using these solutions, we might expect the rate of new workloads to perhaps not continue at the pace that it has over the past quarter, 2 quarters. But I don't see technology executives saying at some point in the future, we need to cut the observability over this huge swath of systems.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 口径不齐 观测预算经验法则约为云支出的 10%，LLM 支出单列后他仍按 10:1 管理 (支出锚定底层支出而非数据量)
JC: Yeah, I think the common rule of thumb that I'm familiar with from the old days is that if your cloud provider spend is X, then your observability spend should be about 10% of that. I think we now obviously have this whole category of LLM spend and then observability spend within that. And if we break those out into their own separate line items, that 10 to 1 relationship is still similar to what I'm trying to maintain in our organization.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 支持 agent observability 已有数千客户、开始有收入；MCP server 调用和 Bits 活动增加 (未量化)
And I think we talked about last night that early on, the first wave was essentially calling out to the model companies through APIs. So, the first way to monetize, this is Datadog for AI, meaning it's just like Datadog's monitoring code, Datadog's monitoring CPUs, Datadog's monitoring databases. This is something else that we've begun to monetize through our agent observability. And although early on, which we're seeing thousands of customers use this, and we're starting to get the revenue streams from it, and what we believe is happening is, if we look at some of the other metrics of building out AI-enabled applications through a combination of the outside models, open source, and inference, is that creates a whole another set of things to monitor that we're starting to see in our Bits products, in our MCP server calls, and other indications, a lot of activity which, again, will increase the complexity and we believe create more importance in observability, more to monitor, and more velocity in application creation, which should all benefit Datadog. There is a sovereign threat in here around. We've had actually a number of companies, if you think about Satya Nadella talking about frontier ecosystems.
- 支持 Agent 增加带动 Datadog、Snowflake、MongoDB 采购增长
另外，从数据基础设施的角度看，新的 AI Agent 模型带动了相关需求。正如我前面提到的 Datadog，以及 Snowflake、MongoDB 等公司，其基础设施产品和云数据库的采购量都在增长。基础数据规模的扩大，或者说基础数据底座的扩容，正在受到 AI 赋能的推动。
来源背景：未具名专家，LVMH 集团内部人员 (原文多次以'集团''我们'谈 LVMH 与 Google Cloud 共建 MAIA，具体职位未明示，推测是数字化/IT 方向)。视角局限：奢侈品集团单一样本；本人承认不清楚 Datadog 等的具体采购金额。 - 口径不齐 Datadog 采购混在整体 AI infra 采购里，金额不清楚
回 答 者：是的。平台在搭建数据底座的过程中会进行统一采购，采购内容并不只是品牌采购，例如 DataDog 或一系列 AI Infra，也包括部分代采。因此，这部分数据并不会非常清晰，可能是混合采购的情况。通常而言，这些采购都会被视为整体 AI Infra 采购的一部分。
来源背景：未具名专家，LVMH 集团内部人员 (原文多次以'集团''我们'谈 LVMH 与 Google Cloud 共建 MAIA，具体职位未明示，推测是数字化/IT 方向)。视角局限：奢侈品集团单一样本；本人承认不清楚 Datadog 等的具体采购金额。 - 支持 LLM 调用开始被当作 request trace 监控 (查询数、吞吐、响应性能)，这是以前不监控的新对象
MC: Yeah. I think 1 of the biggest is just observing and so LLMs, right? Your large language models, whether they're self-hosted, whether they're hosted by some other company and you're just utilizing them, those are currently being treated like request traces, right? So they're looking at the actual queries, how many are happening, what is the throughput, what is the performance of those responses and the outputs. That was something that, for example, was not being tracked, was not being monitored previously. And I think the second is just overall tokenization and token usage across the organization. It's another way to reduce costs from an AI perspective, but that is now being looked at as an observability kind of pillar, something that you want to have visibility on within your organization to just make sure things are running as they're intended.
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 支持 AI 写的代码 (vibe coding) 上线后人类不完全理解，必须加观测
Bernd Harzog: Customers are putting agents into production. Agents are pieces of software that can do work independently to do things for you. Well, guess what? They need to be monitored to make sure that not only are they working, but are they not doing something harmful. And that's a whole new and different problem that the vendors have to solve. Turns out you can use AI to write code for you. It's called vibe coding, right? And so as you as a customer use the AI to write software for you, guess what? You put software into production that none of your humans fully understand because it's written by AI, not you, not your humans.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 支持 现在是 LLM 监控，下一波是 agent，但 '是明年还是后年' 不确定
Bernd Harzog: Right now it's LLM monitoring.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 支持 Wolfe：agent 的每次 tool call、重试、交接都要 trace，trace 成本约是 metrics 2 倍
Wolfe says ecosystem checks show agents materially increasing observability intensity as every tool call, retry and handoff needs to be traced, with traces costing roughly 2x metrics and already driving higher consumption. AI coding tools are also accelerating software-development velocity by ~2–3x, creating more code/services/pipelines for DDOG to monitor, while Wolfe’s initial conversations suggest an increasing likelihood of DDOG becoming directly involved with Meta’s Muse. Wolfe’s upside case now implies ~32% FY25–28 revenue CAGR vs. ~25.5% consensus and sees DDOG as one of the best-positioned ways to play inference/agentic demand.
来源背景：TMT Breakout (TMTB) 日报，转述 Wolfe Research 生态检查报告摘要。二手转述，看不到底稿。 - 支持 Wolfe：AI 编程工具让开发速度提升约 2–3 倍，带来更多代码/服务/pipeline 要监控
Wolfe says ecosystem checks show agents materially increasing observability intensity as every tool call, retry and handoff needs to be traced, with traces costing roughly 2x metrics and already driving higher consumption. AI coding tools are also accelerating software-development velocity by ~2–3x, creating more code/services/pipelines for DDOG to monitor, while Wolfe’s initial conversations suggest an increasing likelihood of DDOG becoming directly involved with Meta’s Muse. Wolfe’s upside case now implies ~32% FY25–28 revenue CAGR vs. ~25.5% consensus and sees DDOG as one of the best-positioned ways to play inference/agentic demand.
来源背景：TMT Breakout (TMTB) 日报，转述 Wolfe Research 生态检查报告摘要。二手转述，看不到底稿。 - 口径不齐 客户要求 DDOG 提高自家平台 rate limit，意味 agent 以机器速度查询 DDOG API；可能带来按量变现，也可能带来成本压力
* One customer ran Datadog’s SRE investigation agent over a few weeks, averaging just over three minutes to root cause with no human and no dashboards opened, which is concrete evidence for paid AI-SRE adoption. * Li says customers keep asking Datadog to raise rate limits on its own platform, implying agents now query Datadog APIs at machine scale; that could open a usage-based monetization path or create cost pressure (our read). * Li frames incident war rooms of tens to hundreds of people as mostly teams proving their innocence; investigation agents remove that labor, which supports a labor-replacement pricing story rather than a seat-based one (our read). * Datadog builds its own models, and Li says a lot of AI’s lift to observability demand still comes from core metrics, traces and logs, so agents should boost the base business, not just new SKUs.
来源背景：TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。 - 支持 Li：AI 对可观测性需求的拉动很大部分仍来自核心 metrics/traces/logs
* One customer ran Datadog’s SRE investigation agent over a few weeks, averaging just over three minutes to root cause with no human and no dashboards opened, which is concrete evidence for paid AI-SRE adoption. * Li says customers keep asking Datadog to raise rate limits on its own platform, implying agents now query Datadog APIs at machine scale; that could open a usage-based monetization path or create cost pressure (our read). * Li frames incident war rooms of tens to hundreds of people as mostly teams proving their innocence; investigation agents remove that labor, which supports a labor-replacement pricing story rather than a seat-based one (our read). * Datadog builds its own models, and Li says a lot of AI’s lift to observability demand still comes from core metrics, traces and logs, so agents should boost the base business, not just new SKUs.
来源背景：TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。 - 支持 agent 仍是分布式系统，传统四支柱仍是基础，新旧两层需求同时增长
Li states that AI agents are still fundamentally distributed systems, so the traditional observability pillars — logs, metrics, traces, events — remain foundational and demand is growing for both the old and new layers simultaneously. What’s new is a layer of model-specific signals: prompts, responses, reasoning, tool calls, full agent trajectories, and token usage. She cites a concrete pattern from Datadog’s own agent-observability product data: a large chunk of agent errors trace back to hitting API rate limits, an old problem that becomes more pronounced because agents access data at much higher speed and scale, driving customer demand to raise rate limits.
来源背景：TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。
E04 单价被通缩吃掉了吗
判断：单价在下行，这一点管理层、渠道、竞品、毛利率都指向同一个方向；但 Datadog 的定价权仍然很高，客户很少整体迁走，降价更多是"为了留住量而主动让利"。
单价下行的证据：(1) 合同结构本身：CFO 说定价是"按量和按期限"的，客户扩大承诺就有价格优势。(2) 主动降价产品：Flex Logs (低价日志层，ARR 接近 $1 亿)、Infinite Cardinality、BYOC、Federated Logs。CFO 说 Infinite Cardinality 是"主动把客户转过去"，对留存和毛利有好处，没说对收入的影响。(3) 毛利率：81.7% (2024 Q4) → 79.6% (2026 Q2)，Funda 认为是大客户折扣所致，管理层坚持"80% 上下"。(4) 客户行为：Zendesk 说即使 88% 的日志不进索引，摄入费照样要付，所以他们在源头去重；Viber 只用 Flex Logs，因为"全量索引太贵"；一位北美渠道/SI 专家说已看到 5–7% 的日志支出转向客户自己的湖仓 (Snowflake/Databricks)。(5) 竞品价差：Grafana Cloud 约便宜 20%，Funda 估 Chronosphere 比 Datadog 便宜 30–40%。
定价权还在的证据：北美 SI 说客户找他们帮忙压价"通常不成功"，所见案例没有一家完全迁走；Elastic 前总监说很少从 Datadog 赢单；Snowflake 一方的专家说没看到客户从 Datadog 转到 Snowflake；北美 GSI 说 Datadog"没有系统性打折"。GRR 在 mid-to-high 90s。
怎么理解：Datadog 在用"单位价格换总量"的方式防守：与其让客户把数据搬到 ClickHouse，不如让客户在 Datadog 里用便宜的层存、在 Datadog 界面里查别处的数据。CFO 的原话是"用户越集中在 Datadog，平均每客户收入越高"。这个策略对留存是对的，但意味着未来新增的数据量会以更低的单价进来，"数据量涨 3 倍、收入涨 3 倍"不会发生。
And then for the most part, this is really about – it's the same thing with AWS and hyperscalers. There's essentially capacity planning we do together. And then for the most part, you can tell with the net retention that they use more than they've committed to. And so, the motion is we then work with them to figure out what their next consumption is going to be, and they have an incentive to do it because we're volume-based pricing and term pricing. So, most of the time we have within the contract – they have a set amount, but it's in their advantage to expand that. And so what we do is we have long histories by customer of what happens. We know because every day we see the usage. I can see the usage when I wake up in the morning, I can slice and dice it by customers.
And what we came to understand is, in some examples of how a customer is using it, it's much better to, I would say, curate or figure out how all of the metrics don't flow in and get processed, but they get sorted before. I'm using very simple language. But – and then that is good for the customer because they get more value in their metrics. That's also good for us because it doesn't help us to have metrics or even logs that flow in and are costly to us. So, there are many, many examples. I think it's a very good point of how we've been evolving the platform to meet the client at the value point. And what we've been doing is proactively converting those customers that can benefit from Infinite Cardinality to our SKU, which has essentially been retentive and also margin-enhancing for us, given the weight on our platform of all that – of large cardinality of our metrics. Yeah. The other pricing announcement, it's not really a pricing announcement, but I think we could frame it as driving value to customers, Federated Logs. Federated log.
And I don't know whether I'm reaching the right outcomes with that. And so that's one of the reasons we've invested in all those products we've mentioned earlier. And also we're seeing some of the great returns on that products already. In terms of Infinite Cardinality, that's a – I would say it's been one of the longest standing source of frustration for customers, when sometimes, they send more data or they send more fine grained tags with their data. And they get some unpredictability on the bills because of that, because it increases the cardinality of the data we're getting. And we've solved that from a technical perspective and from a commercial perspective by packaging our metrics a little bit differently. And we think it's particularly important and relevant as customers are building more applications with AI. And as they want to send basically more tags, more information and ask more complex questions and get more fine grained answers to those questions. And so that fits well within their plans, basically.
Now let's review some of the key income statement results. Unless otherwise noted, all metrics are non-GAAP. We have provided a reconciliation of GAAP to non-GAAP financials in our earnings release. Our Q2 gross profit was $892 million for gross margin of 79.6%. This compares to a gross margin of 80.2% last quarter and 80.9% in the year ago quarter. As we've discussed in the past, our gross margin varies from quarter-to-quarter with investments into innovations for our customers, offset by efficiency efforts. There's no change in our expectations for gross margin, which has been in the 80% plus or minus range historically. Q2 OpEx grew 26% year-over-year versus 31% last quarter and 36% in the year ago quarter. We held our DASH user conference in June, and as expected, the event cost about $15 million. Q2 operating income was $257 million for a 23% operating margin compared to 22% last quarter and 20% in the year ago quarter.
A:是的，你的 10% 到 15% 区间是正确的。我们已经看到 5% 到 7% 的转移。我们感到日志管理在可观测性供应商处的支出未来可能呈下降趋势，尤其是在企业领域。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。
DS: Oh, yeah, in this case, it's more what you're talking about. It's like a Datadog platform or an observability platform. And some of that data gets stored, but a lot of times it's the desire not to put all of the data in the observability platform itself, like not to put all the data into Datadog, but to run some of the higher volume data that maybe you could... higher volume less critical immediately or to peel off after you know you've used it for a week or 2 and put that data in a longer term store into AWS. So the Datadog might be in AWS alone, but not all the data you want there because you're paying Datadog, so.
来源背景：Dave Sidwell，Improving Holdings 芝加哥区总裁 (President for Chicago)，多云咨询/实施伙伴 (Industry Consultant，标题明示)；主持人 Brian Oh。视角局限：只看自己公司的客户，增速是他们自己 AWS 业务的数字，不是 AWS 整体。
MC: I think it typically varies, but you can see there's usually a reduction of around 20%. It's not as dramatic anymore as it was. I think overall that market is... there's a lot of competition. But I think overall, it's probably like a 20% savings if you're leveraging Grafana Cloud. Now, if you're using Grafana Enterprise in your own environment, the savings could be tremendous because you're then using your own resources, there's gear, whatever you want, and you don't need to use that. And there's a lot of advantages to that as well. So I think it really, if it's cloud versus cloud, I'd say... because you can't really deploy Datadog in your own infrastructure. I would say you'd see a 20% savings if it's Datadog, which is the cloud anyways, versus an on-premise deployment. I think it could be a lot bigger depending on the environment and how much data the metrics folks have.
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。
So that that I've seen uh being discussed and and that is technically. quickly more feasible uh compared to a full migration. And then of course, you know, even if they don't migrate fully, uh you know, customers try to negotiate and sometimes they use us uh uh as help here as well. Um but but uh Data Dog basically knows that it's really difficult to replace them uh fully. So so uh, um it it usually doesn't work. Uh I mean you you may get some discounts, etc, but they they have pretty high pricing power. 嗯,OK,对,他说是这个节级客户的话,一直都会去比价,他们的采购部门,因为现在变得也越来越成熟了嘛,但而且这个data dog的价格确实也就是逐年增长,但是他说data dog就是你要完全从它的平台上迁移走,不是一件很容易的事情,因为首先他们的生态系统是很全面的,然后这个 全面性的话,其他的玩家很难替代,第二个就是它是有客户年性的,因为 因为就是datadog本身的各种各样的工具,企业其实是围绕它的工具去做了一系列的工作方式的调整,去配合,比如说它的仪表盘呀,它的各种警报,警告啊等等,所以你要去换其他的工具的话,你也要调整自己的这个工作方式,所以它其实是一个很费劲的过程,还有就是你调整的过程中也会涉及到一些运营的风险。 所以就是有看到客户迁移,但是没有办法完全迁移走。那么客户现在出现的一个趋势就是说,他把一些数据高通量的,但是低价值的一些任务呢,他不用data dog去做观测,他用一些更加便宜的解决方案,比如说用elastic或是grafana等等,然后把这个比较高价值的就留在data dog上。
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。
**CM** [00:23:03]: We did not see it often with Datadog or Dynatrace. We did see it with Splunk, absolutely. Absolutely, we saw it. I think if we were in with a Datadog or a Dynatrace, I know what I would do today as of the new Gartner Magic Quadrant, which I found very interesting. We would win if it was just okay, we play up on… Custom search is critical. That was just one thing that we could obviously be very strong on and how important was that, or maybe it was just price to performance. Maybe a little bit of data retention mattered, we could win other than that. We didn't see as many of those competitive takeouts for sure with Datadog or Dynatrace. Datadog wins on ease of use. They have just a strong developer affinity. They have a really great product. I know that just from spending time at other organisations, New Relic, knowing the competition in Datadog is really pretty strong.
来源背景：Christine Malloy：Third Bridge 介绍为 Elastic 前 Director of National Strategic Partners ；页头写的是 Perforce Software 现任 Partner Ecosystem Manager；自述曾在 New Relic 工作数年。访谈日期 2026-06-10 (Analyst-led，主持人 Luke Fragnito)。视角局限：伙伴/联盟视角；离开 Elastic 的时间未说明；曾在 New Relic 工作可能影响她对 Datadog 的看法。
展开本节全部 116 条原文 (支持 18 / 反对 65 / 其他 33，按时间)
- 反对 CEO 明说：大客户续约拿到更低价格后，收入先降、之后靠用量再涨
Yes. Yes. I mean, look, this is -- we're not getting to the detail of any specific customer that this is the motion is always the same. Like customers grow they commit to more, they get better prices. So you see again, talking about customers in general, you see growth of usage, drops in revenue as customers renew and get higher commit and a better price and then unit growth after that for those customers. That's the motion that we've had. We have about 30,000 customers so far.
- 反对 用量模型下"更好经济性"意味着同样用量收入变少
Well, the bigger economics means you coming tomorrow, you get a better price. And as we -- remember, we have a usage model. So we charge people every month on what they use at the crises we agreed. So if you get better economics, and your usage is somewhat similar month to month less. But the overall backdrop of our business is increased consumption.
- 中性 毛利率 81.2% (Q2 80.9%)，管理层维持 "80% 上下" 框架，大客户折扣被大量小客户平衡
Yes, there's a couple of things. I think we've mentioned that we've been focused and have focused over the many years on the efficiency of our cloud platform. We have significant engineering efforts around cost of sales and delivery of value. And so we've been able to deliver on that. We also have a very broad customer base distributed in terms of volume. So as customers get larger and maybe get volume discounts, we have a number -- a lot of customers coming in, it's smaller, so that balance there. And then in terms of the sort of the future -- I'll repeat what we've always said that we've been running the company with a gross margin plus or minus 80%, we've given that range and not changed it, and we watch it. And it gives us signals in terms of efficiency, how we're operating, it gives us signals in pricing and things like that and I wouldn't change the comments we made over the many years about looking at that and then developing operations and strategies around that.
- 反对 Datadog 比竞争对手贵，客户对开源有兴趣，他怀疑 Datadog 能否长期保持地位
Datadog 比其他人更昂贵。它增长健康——比竞争对手快——但总体上人们对开源平台有兴趣，尽管那在市场增长中尚未明显。我也对 Datadog 是否能长期保持其地位持有猜测。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 口径不齐 客户担心锁定、价格陡峭，但切换成本高，目前还没看到迁出
存在对锁定的担忧。我有过关于替代方案的对话，并且 Datadog 的价格很陡峭，所以如果发生迁移我不会感到惊讶。但同时，切换成本相对较高，因为平台已经嵌入。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 口径不齐 Flex Logs 表现好，近 4 个月签了 3 个大单
We believe Datadog will garner the required traction. Focusing on logs-Flex Logs have been doing well. In the last four months, we have signed three large deals; they were in the works for a few months and finally materialised.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 反对 如果 OpenAI 这类公司在 Grafana 上补齐 LLM 观测能力，会改变格局
A:The difference now is that Grafana is open source. If OpenAI or a similar company decides to build on Grafana and expand its capabilities to address these other aspects, that could be a game-changer. It could happen because it is largely about configuration and plumbing.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 支持 Datadog 没有系统性打折，定价纪律稳固
A:It has remained fairly solid. Datadog has not, to my knowledge, provided any discounts. Yes, there have been some one-off discounts, such as a free usage period for this capability or one particular module, but pricing and discounting have been fairly solid.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 反对 客户更倾向在 Snowflake/Databricks 自建湖仓存日志，不愿为可观测性厂商的日志存储重复付费
- 客户在加速data modernization的同事，在更慎重考虑log management的部署方案，更多的倾向在Snowflake/Databricks自建datalake/lakehouse, 而不愿为observability厂商的log存储重复付费
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 反对 他预计 DDOG 后续定价有压力，需要多推成本管理工具
认为后续DDOG的定价可能会有压力，需要更多推广cost management工具
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 支持 Grafana 等开源工具用量上去后稳定性差、要招工程师调优，客户往往迁回 DDOG
* Grafana等开源工具一开始能吸引企业客户兴趣，但随着用量提升，客户发现稳定度不好，并且需要招聘更多工程师来内部调优，成本实际上并不低，往往会迁移去DDOG
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 反对 日志在客户自己的基础设施里已经存了一份，再推到 Datadog 等于被收两次钱
随着数据现代化计划的推进，客户在审查感知到的重复收费。日志在生成时先存储在客户的基础设施中，然后你将它们推送到 Datadog 进行日志分析或分布式跟踪。端到端评估时，日志费用可能几乎被重复收取。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 反对 客户在看 OpenTelemetry 供应商等替代方案
与成本相关的担忧越来越常见，客户正在探索替代方案——尤其是 OpenTelemetry 供应商，它们已经存在一段时间——并且随着数据现代化和数据流采纳，客户在考虑以不同方式处理整个数据集。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 反对 已看到 5–7% 的日志支出转移，他预计企业在可观测性厂商处的日志支出会下降
A:是的，你的 10% 到 15% 区间是正确的。我们已经看到 5% 到 7% 的转移。我们感到日志管理在可观测性供应商处的支出未来可能呈下降趋势，尤其是在企业领域。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 反对 基于 agent 的采集架构受到质疑，可能影响 agent 计费和存储计费两块收入
如今，我可以在数据所在处高效分析数据。这意味着 Datadog 的收入可能在几个方面受到影响：基于代理的许可成本和计费的存储成本。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 反对 定价能力下降 2025 年还没发生，他预计 2026 年会成为影响因素
定价能力下降是第二个因素。它已经发生了吗？它在 2025 年并未发生，但意识在增加，所以我们认为它将在 2026 年成为一个因素。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 中性 Datadog 用 Flex Logs、存储/索引解耦等产品应对成本压力 (自我蚕食)
A:正是如此。我们将拭目以待。Datadog 已经做了多项改变。引入了 Flex Logs 和存储/索引解耦，他们提出了一种通常称为 Cloud RAM 的索引能力，允许 Datadog 索引在客户的基础设施内运行以满足数据驻留要求。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 支持 Grafana 主要用在初期搭建阶段，到生产规模管理负担重
A:Grafana 不断出现，但我们看到它主要在初期构建阶段被使用。一旦客户进入生产级别，管理 Grafana 需要大量工作；它需要具备合适技能的内部资源。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 中性 毛利率 81.4% (Q3 81.2%，去年同期 81.7%)；AI 大客户按加权平均不稀释毛利，折扣只跟客户规模有关
On a weighted average, they're not. As we've always said, for larger customers, it isn't about the AI-natives or non-AI-natives, it has to do with the size of the customer. We have a highly differentiated -- diversified customer base. So, I would say we're essentially expecting a similar type of discount structure in terms of size of customer as we have going forward. And there are consistent ongoing investments in our gross margin, including data centers and development of the platform. So, I think it's more or less what we've seen over the past couple of years, not really affected by AI or non-AI native.
- 中性 Flex Logs (低价日志层) ARR 接近 $1 亿，日志总 ARR >$10 亿
Today, infrastructure monitoring contributes over $1.6 billion in ARR. This includes innovations to deliver visibility and insights across our customers' environments, whether they are on- prem, virtualized servers, containerized hosts, serverless deployments or parallelized GPU fleets. Meanwhile, Log Management is now over $1 billion in ARR. And this includes continued rapid growth with Flex Logs, which is nearing $100 million in ARR. And our third pillar, the end-to-end suite of APM and DEM products also crossed $1 billion in ARR. This includes an acceleration of our core APM product into the mid-30% year-over-year and currently our fastest-growing core pillar.
- 反对 埃森哲在项目中倾向用自有工具，功能与传统应用监控工具类似
企业客户在经历了2025年广泛的项目试验后，开始关注Token消耗的性价比问题，并寻求更严格的成本控制。埃森哲提供的服务中包含一个帮助客户约束无效Token浪费的节约框架，该框架与风控体系整合在一起。具体而言，埃森哲会先评估业务的性价比并对 业务优先级进行分类，然后根据场景的复杂程度决定调用资源的级别。例如，对于流程数量、环节和角色超过特定阈值的复杂多级联动场景，埃森哲会建议使用更高级别的计算资源以保证模型精度，同时在训练前通过预设模型精度和控制数据维度来防止无效训练。这项服务 与DataDog等产品的区别在于，埃森哲不仅提供监控工具，更侧重于在项目立项阶段就介入，利用立项规则帮助客户迅速计算出项目的成本和效益比。在项目实际应用过程中，埃森哲倾向于使用自有的工具。虽然埃森哲的一些自有工具在技术门槛上可能与传统的应用 监控工具类似，但它们已经集成到埃森哲的工程化解决方案中。
来源背景：未具名；正文明确是埃森哲的人 ('埃森哲2026财年Q2'、'埃森哲倾向于使用自有的工具')，举的案例 (美孚中国区验证) 和国内模型对比说明主要是中国/亚太视角，职位不详。局限：咨询公司有推销自身方案的利益偏向；Datadog 只是被提问者点名，回答没有具体数据。 - 反对 他称存在客户用埃森哲方案替代 DataDog 等工具的趋势 (没有数据)
存在这种趋势。当前客户在技术趋势把控上越来越依赖咨询方的专家意见。由于AI领域工具迭代迅速，传统IT视角下的工具正在被重塑和重构，技术栈之间甚至出现了断层。客户自身的专业度和精力有限，难以对整个技术更新路线进行全面考量，因此需要专业的咨询服 务来提供指导。
来源背景：未具名；正文明确是埃森哲的人 ('埃森哲2026财年Q2'、'埃森哲倾向于使用自有的工具')，举的案例 (美孚中国区验证) 和国内模型对比说明主要是中国/亚太视角，职位不详。局限：咨询公司有推销自身方案的利益偏向；Datadog 只是被提问者点名，回答没有具体数据。 - 反对 客户刻意不把全部数据放进 Datadog，把高量低优先级数据分流到 AWS 存储以省钱
DS: Oh, yeah, in this case, it's more what you're talking about. It's like a Datadog platform or an observability platform. And some of that data gets stored, but a lot of times it's the desire not to put all of the data in the observability platform itself, like not to put all the data into Datadog, but to run some of the higher volume data that maybe you could... higher volume less critical immediately or to peel off after you know you've used it for a week or 2 and put that data in a longer term store into AWS. So the Datadog might be in AWS alone, but not all the data you want there because you're paying Datadog, so.
来源背景：Dave Sidwell，Improving Holdings 芝加哥区总裁 (President for Chicago)，多云咨询/实施伙伴 (Industry Consultant，标题明示)；主持人 Brian Oh。视角局限：只看自己公司的客户，增速是他们自己 AWS 业务的数字，不是 AWS 整体。 - 反对 客户这样做的原因是 Datadog 按量收费
DS: Oh, yeah, in this case, it's more what you're talking about. It's like a Datadog platform or an observability platform. And some of that data gets stored, but a lot of times it's the desire not to put all of the data in the observability platform itself, like not to put all the data into Datadog, but to run some of the higher volume data that maybe you could... higher volume less critical immediately or to peel off after you know you've used it for a week or 2 and put that data in a longer term store into AWS. So the Datadog might be in AWS alone, but not all the data you want there because you're paying Datadog, so.
来源背景：Dave Sidwell，Improving Holdings 芝加哥区总裁 (President for Chicago)，多云咨询/实施伙伴 (Industry Consultant，标题明示)；主持人 Brian Oh。视角局限：只看自己公司的客户，增速是他们自己 AWS 业务的数字，不是 AWS 整体。 - 中性 大客户 D 年度涨价封顶 <5%，增长主要来自自然的调查量增长；2026 预算不含 Bits AI 约 +20–25%
* 2025 DDOG spend of ~US$1.2m, +10–20% y/y. Mid-2025 the customer expected potential overspend to ~US$1.5m driven by agentic AI deployments, but this did not materialize * 2026 budget: ~US$1.5m or below excluding Bits AI (+20–25% y/y); large-scale Bits AI deployment could push it to US$1.5–2m, but the customer does not currently expect a broad rollout * No new feature purchases this year; annual price increases capped at <5%, with growth coming primarily from organic investigation volume * Views DDOG as key infrastructure, and continues to consolidate observability onto DDOG
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 口径不齐 银行客户把剩余日志迁入 Datadog，卖点是 Flex Logs 的成本控制 (低价层产品在赢份额，同时可能压低单价)
This customer will grow to 16 Datadog products including Datadog MCP server. Next, we signed a seven-figure annualized expansion for an eight-figure annualized deal with a Fortune 500 bank. With this expansion, this customer will migrate the remaining log data into Datadog fully replacing their legacy log vendor. Most notably, our Flex Logs gives them granular control over costs while meeting strict compliance requirements. This customer uses 10 Datadog products including Bits AI SRE Agent to accelerate incident response with AI. Next, we signed a seven-figure annualized expansion with a leading global hedge fund.
- 中性 毛利率 80.2% (上季 81.4%、去年同期 80.3%)，解释为投入创新与效率提升相互抵消
Now, let's review some of the key income statement results. Unless otherwise noted, all metrics are non-GAAP, and we have provided a reconciliation of GAAP to non-GAAP financials in our earnings release. First, Q1 gross profit was $807 million with a gross margin of 80.2%. This compares to a gross margin of 81.4% last quarter and 80.3% in the year-ago quarter. As we've discussed in the past, our gross margin varies from quarter-to-quarter with investments into innovations for our customers offset by efficiency efforts. Our Q1 OpEx grew 31% year-over-year versus 29% last quarter and 29% in the year-ago quarter. As a reminder, we continue to grow our investments to pursue our long-term growth opportunities.
- 支持 开源普遍存在，但客户为统一平台和省钱而整合到 Datadog (回应 Lenschow 开源/噪音问题)
Can you speak what you see in real life there? Thank you. I mean, in real life, most companies have open source in some capacity somewhere. But, when it comes to having a platform that unifies everything, takes care of everything, does more of the problem solving for you, that's typically what customers use us, and the motion we see pretty much, everywhere. These customers have four, six, seven, 15, 25 different things and different pockets in the organizations and different business units. And it's a huge mess.
- 口径不齐 BYOC/云上私有部署：未来可能有相当部分市场倾向在客户自有环境运行；目前还不是市场主体
Well, definitely, we think, I think there was a question earlier on data residency and, living in customers' environments. We definitely see a great opportunity there. There's a chance that a good portion of the market, leans this way in the future. Today, it's not the largest part of the market, but we definitely see the potential for that. So, we're investing heavily in that side of our product. We are starting to see some interesting customer attraction there. So, we think this can be another growth lever definitely.
- 反对 毛利率下降，Funda 推断是给大客户的折扣所致
+ The “core” non-AI customer cohort **grew at a mid-20% year-over-year rate, up from 23% in 4Q25** and 19% in 3Q25. + This implies **AI-native customer growth accelerated to 17% QoQ**, after growth was impacted by OpenAI’s migration out in 4Q, demonstrating Anthropic and other AI native customers’ consumption picking up. * 4Q adj OPM beat expectation by 1ppt, demonstrating strong operating leverage despite GPM being lower, likely due to the discount offered to large customers * 2Q26 revenue guidance is set at 30% year-over-year growth, surpassing market expectations by 6-7 percentage points. **Based on a normal beat cadence of 4-5 percentage points, actual growth could reach 35% year-over-year, the highest** **since 2022.** * Full-year revenue guidance is now 26-27% year-over-year growth. Factoring in **typical beat cadence, FY26 revenue is projected to accelerate to low-to-mid 30% growth.**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 自建观测：管理层认为主要是文化原因，最终经济性更差
So if we think about it that way, like if even the hyperscalers conclude that they need to buy from you rather than use what they have, I mean, was that surprising to you at all? And what do you think that says about the value proposition that you have and even the moat that you've built? So, I mean, I wouldn't call it surprising, but it's definitely a very interesting proof point because the way we see it is the reasons to build it yourself are usually mostly cultural. It's because you want to build yourself. You want to have your teams do it or somebody on the team want it so bad and nobody else around them wants to prevent them from doing that.
- 反对 客户在数据爆炸下持续追求效率：成本反馈、更聪明的采样、更高效存储 (BYOC) 三大机制
Yeah. I mean, and just to zoom out, customers, in general, they face just an explosion of the amount of data and application and everything, right? And so, in observability, there's a perpetual quest for efficiency and making sure that what you spend on measuring, auditing, bringing your application remains predictable and within range of what you were expecting. For that, there are three core mechanisms. One is you need to have a feedback loop that lets you understand what you're consuming and what's valuable and what's not or what you need to do more or less of.
- 反对 BYOC：客户自己存数据、Datadog 管应用；大规模成本是原因之一
Another part is to be to get ever smarter on how much of a sample you need to understand the data, and that goes back to all the smarts, the building, the models, et cetera, et cetera, et cetera. And the third part is to keep innovating on how efficient you can be at storing and querying and managing the data and all the various topologies you can employ to make that work for various customers. So, in that last one, one thing we started doing last year and we see quite a bit of demand for and we are accelerating on is we are doing some what we call Bring Your Own Cloud options for our customers to deploy our product, meaning that they can store the data on infrastructure that they manage themselves even though we still run and update the application for them. And the reasons for that are, so we see cost at large scale is one of them, (44:54) petabytes of data. And that's something that you probably want. Other reasons would be that are (45:02) and we see that the world, in general, is fracturing into smaller zones as opposed to one big zone.
- 反对 CloudPrem 目标是“无所谓”：客户自己存数据也行，给 Datadog 也行，先做日志
Now, we're investing behind it. We have customers that want it and we've been successful. It still hasn't been the vast majority of the way our customers want to buy, but what we're working towards is indifference, meaning that if you want to keep the data, great; if you want to give us the data, great; if you want to have all the functionality go over to your side – and so that's taking time. To get complete product parity where we have exact functionality, that's taking time and that's what CloudPrem means. We're working on that.
- 中性 按信用额度 (credit) 卖，客户在平台内任意使用，不区分产品
There's proliferation, there's a lot of observability point solutions out there, like I said, all of the decision-making is leading towards consolidation in the single platform so we think we're pretty early on in that trend. Do your buyers still think about – and we'll talk about – in the three core applications, infra, APM, log analytics, do they still think about that as three separate products or are there – are customers beginning to come to you and saying, please solve this... (00:33:47) No, they've – for a long time, we've been selling as credit. So you buy $2 billion of Datadog, you go in and you use all of this. No, they've never thought of it as different products. We're doing that in order to provide transparency to you all.
- 中性 少数公司 (如 Google) 倾向自建，不是核心市场
Is there any limitation to the type of company or size of company that might not look at Datadog anymore, or is Datadog available to all types of companies? Well, the – so there are some companies that have in their past wanting to – want to do this themselves. They're very few, I mean, maybe I think – Google's a good example of trying to do things themselves. So it's not – that's not really core to our business, although – and we'll talk about it, there's examples where we've gotten that – those types of businesses, and we'll talk about that. I think the other thing is if we generally are delivering our product through the cloud, so if you require an on-premise solution for regulatory or other reasons, that has by choice not been where we've concentrated our R&D.
- 反对 Chronosphere 的卖点是帮客户降低可观测性成本，遥测越多这一主张越有吸引力
Chronosphere 我认为是一个间接的 AI 赢家，因为它本身并非严格意义上的网络安全公司，其业务领域主要在可观测性、监控、遥测和云原生运营。AI 会产生大量的日志、事件、指标等数据，运行 AI 工作负载的客户会发现其遥测数据量大幅增长。因此，随着 AI 部署规模的扩大，Chronosphere 的价值主张变得更具吸引力，因为它能降低客户的可观测性成本，提高遥测效率，并提供对 AI 工作负载的全面可见性。
来源背景：未具名专家，Accenture (埃森哲) 前任人员 (身份由同场 MeritForum 06-04 标题'埃森哲离职专家'推得；本篇无自我介绍)。以'my portfolio'口吻谈客户，推测为咨询/集成侧的安全业务负责人。视角局限：咨询商视角，只见自己客户组合；主题是 Palo Alto/CyberArk 安全，DDOG 为顺带；Chronosphere 数字为举例性量级。 - 中性 AI 技术栈成本结构中监控与运营约占 10-15%
展望未来的技术栈成本结构，API 成本可能占 25% 至 40%，GPU 成本占 20% 至 30%，数据与集成成本约 25%，安全与治理占 20%，而监控与运营则占 10% 至 15%。当然，确切的百分比会因垂直行业而异，但不会改变的是，安全将成为 AI 成本结构中的一个永久组成部分。安全之所以会成为结构性成本，是因为当 AI 规模化应用时，客户若没有身份安全、数据安全、AI 运行时安全及合规性保障，将无法正常运营，这些都不是可选项，而是业务关键环节。
来源背景：未具名专家，Accenture (埃森哲) 前任人员 (身份由同场 MeritForum 06-04 标题'埃森哲离职专家'推得；本篇无自我介绍)。以'my portfolio'口吻谈客户，推测为咨询/集成侧的安全业务负责人。视角局限：咨询商视角，只见自己客户组合；主题是 Palo Alto/CyberArk 安全，DDOG 为顺带；Chronosphere 数字为举例性量级。 - 反对 Chronosphere 能在数据入库前过滤、聚合、采样，降低数据量
它有两个优势：第一是降本增效能力，包括对数据点做采样，减少数据量。很多遥测数据需要采样，比如秒级、毫秒级采集的数据要上卷到分钟级、五分钟级，去看更粗的趋势，历史数据会丢弃，这里需要采样算法帮助有效上卷，减轻存储压力。Chronosphere 在数据进入存储前就能做过滤、聚合、塑形、采样，这些能力很强。第二，它原生做GPU解决方案，非常了解不同GPU型号及相关外设，包括铜缆、连接、交换机等，整套指标体系都做了，知道哪些指标重要，代表什么意义，并制定了对应规则，做了GPU级的解决方 案，所以能得到Anthropic、OpenAI等公司的认可。ClickHouse只是一个OLAP数据库，强调数据分析能力。
来源背景：未具名；以 Elastic (ES) 内部视角回答 (清楚 ES 组织架构调整、API 文档策略、Jina 整合)，推测为 Elastic 现任或前任员工，带明显中国区视角 (国内定价、阿里/腾讯 ES 服务)。局限：竞争对手视角；对 Datadog 的判断是间接的。 - 反对 续约谈判中客户掌握杠杆；集中在一两家观测厂商时可随量增压价
Expert: This will depend upon your landscape. Also, are you heavily consolidated already, or are you uniquely distinct? Meaning if you have multiple point solution, then you might lose the leverage so on the negotiation, but if you are actually heavily concentrated in one or two key vendors in this space since we're talking about observability, then actually you have the leverage to negotiate as your volume goes. More or less, again, you got to look into your what is your strategy, what is your approach, your best of breed or a platform. That will be your trigger point, but more or less depending, as I was talking about, here, within my space, we always work with the best of breed. There are multiple solutions in the observability space, and that's definitely a challenge, but also there's a plus. The advantage of having because we have the unique visibility.
来源背景：匿名专家 ([Expert])。Third Bridge 主持人介绍为 'former Head of CCoE [Cloud Centre of Enablement] and Platform Engineering at PNC Financial Services'；但专家自述 '我是' PNC Cloud CoE 和平台工程负责人、在该行两年 。负责容器、中间件、数据平台及平台安全 (本地和云)；上一家机构曾把 Splunk 整合迁移到 Google Chronicle 。视角局限：受严格监管的大型银行买方，偏 '最佳单品 + 多厂商冗余' 策略，不代表一般企业；角色偏平台/安全，不直接管观测账单细节；对 Datadog 的评价较笼统、无数字。 - 反对 日志量随规模、并购、监管保留无限增长，按日志计价伤大企业，客户更想要按席位计价
If you're highly regulated, then there is some laws you have to mention about the logs that are there over seven years depending upon if you're financial and then if there's any litigation, then logs goes indefinite. Some players, they did not change their model because they're still capitalising and this where customers are looking into offsetting because if they cannot change the model, because log, there is no stoppage to it. You cannot control the logs. Logs is going to grow. That's why we're looking for that per seat basis because that is a named account. They can control that. If there's an onboarding happens, someone new comes in, we can remove those licensing. There are multiple factors here engaged, but I would say it's per seat, named account actually works in our favour than actually the size of the log because that is never ending.
来源背景：匿名专家 ([Expert])。Third Bridge 主持人介绍为 'former Head of CCoE [Cloud Centre of Enablement] and Platform Engineering at PNC Financial Services'；但专家自述 '我是' PNC Cloud CoE 和平台工程负责人、在该行两年 。负责容器、中间件、数据平台及平台安全 (本地和云)；上一家机构曾把 Splunk 整合迁移到 Google Chronicle 。视角局限：受严格监管的大型银行买方，偏 '最佳单品 + 多厂商冗余' 策略，不代表一般企业；角色偏平台/安全，不直接管观测账单细节；对 Datadog 的评价较笼统、无数字。 - 反对 只用 Datadog 成本高，计费复杂，看不清月度账单为什么变
这两种趋势其实可以共存，并不会出现谁完全取代谁的情况。比如公司现在既用Datadog，也用Prometheus、Grafana等工具，都是混合使用。之所以这样做，首先，如果只用Datadog，成本会很高。Datadog的计费逻辑比较复杂，有 时候按日志、指标、定制化功能等多项收费，账单拉出来后很难完全搞清楚每月的费用变化，因为指标采集的多少都会影响账单。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 反对 AI 降低了维护开源监控的门槛：写 YAML、定制 dashboard 都能靠 AI 辅助
AI的出现带来了新的变化。以前在Prometheus或Grafana上做一些专业运维的事情，需要花几个月时间去学习和掌握相关知识。现在有了AI，很多知识和配置都可以通过AI辅助快速解决，比如写YAML文件、定制dashboard监控等，都能 找到很多解决方案，效率提升很大。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 反对 公司规划把非核心监控逐步迁到 Prometheus，整体趋势是向开源倾斜
公司的规划是，很多非核心、非重要的监控项会逐步迁移到Prometheus等开源方案。专业定制的报表也可以通过AI生成，这样可以逐步接近只用开源方案的目标。以前Datadog在dashboard监控和界面上确实比Grafana等好看、易用，但 现在AI的帮助下，开源方案的界面也越来越好看了。个人感觉未来如果发展到一定程度，维护Prometheus其实就足够了，Datadog可以不用了。但目前公司还是两者并用，整体趋势是向开源方向倾斜。不过开源方案也不太可能完全取代商业版的Data dog，因为Datadog有自己的优势，比如它的Agent采集数据非常精准，时延控制得好。Prometheus只是数据库，需要在每台机器上安装Agent，国内的开源Agent在采集效率和指标量上和Datadog的Agent还是有差距。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 支持 开源也很难完全取代 Datadog，因为 Datadog 的 agent 采集精准、时延低
公司的规划是，很多非核心、非重要的监控项会逐步迁移到Prometheus等开源方案。专业定制的报表也可以通过AI生成，这样可以逐步接近只用开源方案的目标。以前Datadog在dashboard监控和界面上确实比Grafana等好看、易用，但 现在AI的帮助下，开源方案的界面也越来越好看了。个人感觉未来如果发展到一定程度，维护Prometheus其实就足够了，Datadog可以不用了。但目前公司还是两者并用，整体趋势是向开源方向倾斜。不过开源方案也不太可能完全取代商业版的Data dog，因为Datadog有自己的优势，比如它的Agent采集数据非常精准，时延控制得好。Prometheus只是数据库，需要在每台机器上安装Agent，国内的开源Agent在采集效率和指标量上和Datadog的Agent还是有差距。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 支持 不考虑成本的话，Datadog 是非常优秀的产品，在安全和全链路追踪上有不可替代的优势
采集数据Agent的壁垒主要体现在合规、安全等多维度。如果公司不在乎成本，Datadog是非常优秀的产品。从安全、合规、后续成本、业务合作等角度综合考虑，Datadog有不可替代的优势，尤其是在安全和全链路追踪方面。开源方案虽然可以实现类似 功能，但对运维人员的素质要求很高，即使有AI辅助，短期内也难以完全替代Datadog。理论上，配备足够的Agent和高素质运维人员可以实现类似效果，但实际操作难度大，除非有非常了解Datadog的人员加入。另外，Datadog在报表、APM 、日志等整合能力上表现突出，和Splunk等产品有部分重合，但侧重点不同，Splunk主要做日志分析和安全日志，Datadog则以监控为主。整体来看，Datadog在不考虑成本的情况下，是非常优秀的产品。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 中性 核心指标在理论上可以切到开源，但能否达到 Datadog 的效果还没被证明
主要看某个指标缺失或异常时带来的代价是否巨大。核心指标通常由业务或开发提出，数据采集后用于监控和分析。还有一些后发性指标，是在出现问题后补充的。如果某个指标缺失会导致重大影响，超出公司可承受范围，这类指标就是核心指标。例如成交量、交易笔数、 每秒吞吐量、关键业务响应速度等，这些通常也是KPI考核指标，直接影响用户体验和公司收益。非核心指标则是有无影响不大，公司可以承受其缺失或延迟。核心指标理论上可以切换到开源方案如Prometheus，但是否能达到Datadog的效果还无法证明 。如果关键指标无法正常采集和监控，风险较大，因此在选择监控方案时需要权衡。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 反对 Datadog 按节点数、指标数和数据量计费，这些量减少，费用自然不增
暂时没有增加。公司以前的核心指标等内容其实在减少，很多都迁移到了开源的 Grafana系统。相对来说，Datadog 上的指标数、dashboard数据、日志量等都减少了很多。Datadog 的计费方式是根据节点数、指标数和数据量来计费，只要这些减少，费用自然不会增加。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 支持 对核心用户和大企业，Datadog/Splunk 的合规、行业案例和集成经验短期难替代；AI 主要影响中低端、成本敏感的客户
Splunk和Datadog在各自领域只要不出现重大失误，依然会保持领先。AI的发展会对他们产生一定影响，尤其是对中低端、成本敏感的客户，但对于核心用户、大型企业来说，Splunk和Datadog的合规性、行业案例和系统集成经验是开源方案短 期内难以替代的。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 反对 增长放缓来自 FinOps、日志量优化、云成本控制，以及客户把低价值遥测迁到更便宜的存储/分析平台
Um, there was a higher trace volume, there was broader APM deployment. Um, I've seen expansion in security monitoring and also the observability observability requirements are growing for AI applications and agent-based architectures. Where growth, um, kind of slowed down, I think was typically due to FinOps initiatives, log volume optimization, um, cloud cost control, and where customers were shifting lower value telemetry to um, a cheaper storage, um, and analytics platform. But overall, the AI driven workloads, um, I think they are meaningful drivers to data dog consumption in 26. particularly, uh, where customers are deploying Gen AI applications into prod.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 云厂商以更低成本提供 80% 功能会抢走部分客户，基础设施监控和日志在商品化；云厂商已在挤压 Datadog 的增长和利润率
Despite the cloud providers offer native IT operations capabilities. Now, having said that, I do expect, although hyperscalars to win share at the lower end of the market and among cloud native startups, which operate primarily on one single cloud. Now if AWS can provide 80% of Terado's functionality at a much comparatively lower cost, some customers will definitely, you know, choose the native option. The pressure is likely to be more in infra monitoring and logging, which is becoming more and more commoditized. So I think the more defensible area for Datadog are higher value capabilities such as cross platform observability, APM, security analytics, AI workload monitoring, RCA, workflow automation and uh operational intelligence. Now the question is whether those high value services can grow quickly enough to offset the pricing pressure in the core monitoring business or not.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 客户因价格在评估替代品，部分工作负载已迁出 Datadog，但多为部分迁移
So, do you see that today some customers are already migrate or is migrating from Datadoc to other solutions because of the pricing reason? And how do you view the competition between Datadog and these open source competitors in the future? Um, so look, I've seen customers actively evaluate the alternatives because of the pricing and in some cases migrate part of their workloads away from Datado. The most common driver is probably not the dissatisfaction with functionality. I think it's cost. Um, as telemetry volumes grow, especially log traces and AI related workloads, the observability spending can increase too much, right? and become a material line item um for your platform teams. However, most of the migrations are partial, right? Not complete. What I often see is customer retaining data dog for high value use cases. um and like um thinking about things like APM, incident response, production observability, uh critical app monitoring, and they move the lower value log retention history or certain infra monitoring workload to cheaper alternatives such as B or open search or or Prometheus, right? The challenge of the open source competition is that the software licensing is only one component of their total cost. Customers need to consider engineering effort, operational support, platform maintenance, uh upgrade, integration, governance, security, and even 24x7 reliability. Many customers discover that while the software itself is free, the operational burden is not. Having said that, I do think the competitive threat from open source is stronger today than it was five years ago.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 开源威胁比 5 年前更强，OTel 成为标准降低锁定，AI 自动化运维会缩小开源的运维劣势
Customers need to consider engineering effort, operational support, platform maintenance, uh upgrade, integration, governance, security, and even 24x7 reliability. Many customers discover that while the software itself is free, the operational burden is not. Having said that, I do think the competitive threat from open source is stronger today than it was five years ago. Open telemetry is now, you know, a common standard which reduces the vendor lock in. Grafana has improved, open search and elastic, you know, they're maturing. and as AI helps automate the platform operations, the um historical operational disadvantage of open source solution may gradually decrease. So looking forward, I think that a dog will remain strong, you know, among large customers who prioritize operational simplicity, rapid deployment, integrated workflows, um, and enterprise support. But I also expect increasing pricing pressure.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 客户比 2–3 年前更价格敏感，AI 原生和传统客户都是
And are DataDog's customers including AI native companies and the traditional companies, are they sensitive to DataDog's price? And uh, do you know that whether Data Dog is uh taking some measures to uh let the customers to be less sensitive to the pricing or is there any solution or measures taken by Data Dog? Um, I think yes, customers are surely more price sensitive today, right? than they were two or three years ago. And that applies to both AI native customers and traditional ones, although the reasons are slightly different. Um, AI native customers, they experience extremely rapid telemetry growth because the AI workloads they generate large volume, you know, logs, traces, model metrics, inference data and GPU monitoring information. And as a result, observability cost can scale very quickly for you, which makes FinOps and cost optimization a board level concern.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 Datadog 自己投资采样、过滤、保留策略、成本分析帮客户少摄入数据 (自我蚕食单位量)
The objective there is to increase the overall value delivered per dollar spent rather than compete purely on infra monitoring costs. The other area is helping customers optimize their usage. Um, DataDog has invested in log management controls, data filtering, sampling, um, retention policies, observability cost analytics and even recommendations, which help customers reduce the unnecessary data ingestion. Now in many customer conversations today, Data Dog is not only selling more consumption, but also helping customers manage that consumption more efficiently. So, I also see DataDog becoming increasingly flexible in commercial negotiations for strategic customers.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 对战略大客户越来越灵活：批量折扣、企业协议、多产品打包、缓解不可预测用量的定价结构
Now in many customer conversations today, Data Dog is not only selling more consumption, but also helping customers manage that consumption more efficiently. So, I also see DataDog becoming increasingly flexible in commercial negotiations for strategic customers. Um, large customers, you know, who have increased spend, they receive volume discounts, enterprise agreements, multi-product packaging, and even pricing structure, which is designed to reduce the concerns around um unpredictable consumption growth. I think the challenge for Data Dog is that customers, you know, they want to benchmark them against open source alternatives, cloud native tools and AI power operations platform. So, their long-term strategy, I think, to me is moving away from being perceived as simply a monitoring tool and towards becoming a operational intelligence platform.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 云厂商原生监控在利润率上压 Datadog
and and there at Dynamics is another example and there uh Datadog keeps taking share uh and and especially given that again, Data Dog is like an ecosystem of different modules which they are the legacy players typically are not. Uh so there they are winning, so that's um uh that's the positive side. Then you have the hyperscalers and they have like uh, you know, native uh monitoring solutions and and and they are uh uh integrated of course with their own ecosystem and with their own cloud and they have they are pressuring Datadog on margins basically. Um at the same time, the hyperscalers struggle when the when the client has a multi cloud uh strategy. Uh that that's like genuinely difficult for them to solve because they they're really good at monitoring within their own cloud.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 反对 Snowflake 进入可观测性是威胁：存储商品化，客户不愿为在供应商云之间搬数据付费 (主要在基础设施监控层)
Uh and and basically here I would say basically the the underlying thing here is the storage, right? Because uh basically observability generates a lot of data uh and and like uh data dog has a sort of a bring your own cloud. uhm, but but but customers are not very keen on paying like to move a lot of data between uh various uh vendors clouds when when storage is commoditized, right? So so I think um basically data dog brings a unified analysis across storage across cloud which is very attractive. But but snowflake's move uh is uh a threat. And then sort of the fourth bucket would be uh you know, open source and and and like um niche names like Prometheus, open telemetry etc. and they are like uh yeah, I mean if you're really good engineering wise and and if you only want certain capabilities and if you're very cost conscious, they can work.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 支持 切换成本和运营风险使所见所有案例都没能完全迁走 Datadog
Uh, secondly, that there is a stickiness here in like uh, you know, uh dashboarding, monitoring, runbooks, alert logic, like you you have like uh you have like almost like a a way of working that you have set up around the data dog uh installation, right? And and if you start to to fragment that across tools, you need to rebuild that and that that is uh work and it's also like a security and observability risk. Uh, so, so uh operational risk and and and switching cost is is usually actually in every case that I've seen is that like stopping them from fully fully migrate away from data dog. However, what I'm seeing more is like a segmentation of uh or or tearing of workloads. So you can actually use, you can use uh like a uh um sort of cheaper alternatives like elastic, Grafana, etc, for uh for like the high volume, low value log tears, like, you know, um that that is a partial offload.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 反对 更普遍的是分层：高量低价值日志部分分流到 Elastic、Grafana
And and if you start to to fragment that across tools, you need to rebuild that and that that is uh work and it's also like a security and observability risk. Uh, so, so uh operational risk and and and switching cost is is usually actually in every case that I've seen is that like stopping them from fully fully migrate away from data dog. However, what I'm seeing more is like a segmentation of uh or or tearing of workloads. So you can actually use, you can use uh like a uh um sort of cheaper alternatives like elastic, Grafana, etc, for uh for like the high volume, low value log tears, like, you know, um that that is a partial offload. So that that I've seen uh being discussed and and that is technically. quickly more feasible uh compared to a full migration.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 支持 客户找 SI 帮忙谈价通常不成功，能拿到一些折扣，但 Datadog 定价权很高
quickly more feasible uh compared to a full migration. And then of course, you know, even if they don't migrate fully, uh you know, customers try to negotiate and sometimes they use us uh uh as help here as well. Um but but uh Data Dog basically knows that it's really difficult to replace them uh fully. So so uh, um it it usually doesn't work. Uh I mean you you may get some discounts, etc, but they they have pretty high pricing power. 嗯,OK,对,他说是这个节级客户的话,一直都会去比价,他们的采购部门,因为现在变得也越来越成熟了嘛,但而且这个data dog的价格确实也就是逐年增长,但是他说data dog就是你要完全从它的平台上迁移走,不是一件很容易的事情,因为首先他们的生态系统是很全面的,然后这个 全面性的话,其他的玩家很难替代,第二个就是它是有客户年性的,因为 因为就是datadog本身的各种各样的工具,企业其实是围绕它的工具去做了一系列的工作方式的调整,去配合,比如说它的仪表盘呀,它的各种警报,警告啊等等,所以你要去换其他的工具的话,你也要调整自己的这个工作方式,所以它其实是一个很费劲的过程,还有就是你调整的过程中也会涉及到一些运营的风险。 所以就是有看到客户迁移,但是没有办法完全迁移走。
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 支持 Citi：取代 Splunk、New Relic 的机会稳健，开源替代的竞争压力有限
Following DASH ‘26 and customer checks, Citi raised its PT to $270, arguing Datadog’s competitive position continues to strengthen as enterprises scale AI and agentic workloads. Customer feedback was broadly positive on AI observability demand, with Bits AI and Cloud Cost Management receiving strong interest, while displacement opportunities against Splunk and New Relic remain robust and competitive pressure from open-source alternatives appears limited. Citi acknowledges investor concerns around tokenization, export controls, and lab-customer concentration, but believes these risks are overstated relative to Datadog’s expanding AI opportunity, growing enterprise adoption, and ability to capture incremental observability spend.
来源背景：TMT Breakout (TMTB) 日报，转述 Citi (DASH 后与客户检查) 报告摘要。二手转述，看不到底稿。 - 中性 Datadog 现在靠商务条件 (价格/商务弹性) 赢 Dynatrace 等对手
Um this is probably going to be a package um based on usage maybe, uh or based on the entire environment. Um the the challenge here is, you know, the more agents you adopt the more expensive it can be. So, Data Dog is really um competing in terms of commercials. and that's where it's winning at the moment, you know, against Dynatrace or some of the other competitors. It's winning because of its commercial uh appetite. So, yes, this is a this is a good question. It will definitely expand on its opportunity by at least 20%.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 中性 他预计包装会走向 '订阅 + 高级附加组件' 的混合模式
在包装方面，我们可能会开始看到混合模式，订阅加上高级附加组件。定价可能会沿着这个方向发展。我们希望 10–20% 的增长。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 在全栈可观测性上，Elastic 很少从 Datadog/Dynatrace 赢单，主要从 Splunk 抢客户
**CM** [00:23:03]: We did not see it often with Datadog or Dynatrace. We did see it with Splunk, absolutely. Absolutely, we saw it. I think if we were in with a Datadog or a Dynatrace, I know what I would do today as of the new Gartner Magic Quadrant, which I found very interesting. We would win if it was just okay, we play up on… Custom search is critical. That was just one thing that we could obviously be very strong on and how important was that, or maybe it was just price to performance. Maybe a little bit of data retention mattered, we could win other than that. We didn't see as many of those competitive takeouts for sure with Datadog or Dynatrace. Datadog wins on ease of use. They have just a strong developer affinity. They have a really great product. I know that just from spending time at other organisations, New Relic, knowing the competition in Datadog is really pretty strong.
来源背景：Christine Malloy：Third Bridge 介绍为 Elastic 前 Director of National Strategic Partners ；页头写的是 Perforce Software 现任 Partner Ecosystem Manager；自述曾在 New Relic 工作数年。访谈日期 2026-06-10 (Analyst-led，主持人 Luke Fragnito)。视角局限：伙伴/联盟视角；离开 Elastic 的时间未说明；曾在 New Relic 工作可能影响她对 Datadog 的看法。 - 口径不齐 Elastic 赢的时候靠价格性能比/数据留存/自定义搜索
**CM** [00:23:03]: We did not see it often with Datadog or Dynatrace. We did see it with Splunk, absolutely. Absolutely, we saw it. I think if we were in with a Datadog or a Dynatrace, I know what I would do today as of the new Gartner Magic Quadrant, which I found very interesting. We would win if it was just okay, we play up on… Custom search is critical. That was just one thing that we could obviously be very strong on and how important was that, or maybe it was just price to performance. Maybe a little bit of data retention mattered, we could win other than that. We didn't see as many of those competitive takeouts for sure with Datadog or Dynatrace. Datadog wins on ease of use. They have just a strong developer affinity. They have a really great product. I know that just from spending time at other organisations, New Relic, knowing the competition in Datadog is really pretty strong.
来源背景：Christine Malloy：Third Bridge 介绍为 Elastic 前 Director of National Strategic Partners ；页头写的是 Perforce Software 现任 Partner Ecosystem Manager；自述曾在 New Relic 工作数年。访谈日期 2026-06-10 (Analyst-led，主持人 Luke Fragnito)。视角局限：伙伴/联盟视角；离开 Elastic 的时间未说明；曾在 New Relic 工作可能影响她对 Datadog 的看法。 - 反对 Elastic 在成本敏感的迁移中获胜 (主要针对 Splunk)，成本是客户离开旧平台的首要原因
**CM** [00:20:35]: I think Cisco stabilises Splunk probably in regulated accounts. Elastic would win. Elastic wins in absolutely cost-sensitive migrations. The other way we would win was it would take people… These organisations who are on Splunk have been using it for a long time, and it's part of also their culture and their people, and it's a pretty big transformation. That's part of how we, I think, would also win in saying that, "Listen, you don't want to be locked in anymore, and we can give you a roadmap that will get you out of this cost paralysis that lots of customers continue to say that they were in." Modern cloud-native workloads. It was a very big push to replace Splunk in every single account. What was interesting was on the Cisco side and with even other partners, large partners who have very good relationships and partnerships with Splunk and Cisco. They the interesting thing was that though they were getting the same feedback from customers and that, "I've got to get off Splunk. I won't continue to…" Number one was always cost. It was like, okay, well, originally, we thought, oh gosh, this is going to be tough to partner with them when they're really strong partners with Splunk, Cisco, etc. It just came down to the facts of when the customer says they want to get off, we will help you put that roadmap together.
来源背景：Christine Malloy：Third Bridge 介绍为 Elastic 前 Director of National Strategic Partners ；页头写的是 Perforce Software 现任 Partner Ecosystem Manager；自述曾在 New Relic 工作数年。访谈日期 2026-06-10 (Analyst-led，主持人 Luke Fragnito)。视角局限：伙伴/联盟视角；离开 Elastic 的时间未说明；曾在 New Relic 工作可能影响她对 Datadog 的看法。 - 中性 Datadog 是市场上较贵的工具，Zoho ManageEngine、New Relic、Dynatrace 都更便宜，所以流向 Datadog 的是收入关键型应用
In in theory that could happen, right? But if I if I look at the kind of use cases that data dog tackles, it's uh as I was saying earlier, usually the critical uh revenue facing uh uh sort of like applications which are most monitored by uh by data dog, right? Given it's a relatively expensive tool in the market. There are many cheaper alternatives. There's uh manage engine by Zoho, there's New Relic, Dino trace, all of these tools which are comparatives to data dog are actually cheaper than data dog. So the kind of use cases that we see flowing to data dog in an organization are usually the ones that are very critical and revenue impacting. All right.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 客户拿 vibe coding/换供应商当谈判筹码，一说要关掉实例就撤回
Now, due to the reasons that I said earlier about how deeply it integrates with different uh components within the uh company's coding infrastructure and also Also, uh given the nature of these revenue critical uh applications, you wouldn't really want to trust a vibe coded tool uh to uh to save money on maybe uh the the the cost of the tooling when in fact the risk from revenue leakage is actually much larger. Uh if if if the app doesn't uh really end up performing uh as per standards, right? So so far, uh I mean, we've seen isolated pockets where, you know, customers are maybe uh mentioning this as maybe a negotiation tactic, but whenever they've been told that okay, sure, like we'll turn off your data dog instance, they've actually come back and said, no, no, that's actually not something that we want and they've been called out on that being a negotiating tactic instead of something that they can actually uh realistically do, right? So, it's it's a possibility, but due to the reasons that I mentioned, again, it's not something that I'd be uh overly, overly concerned by. Mhm, okay.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 中性 客户认为 AI token 价格短期不会降反而会涨 (模型公司盈利压力)，自建反而不划算
People are already in a lot of cases blowing through what they had budgeted for AI for the whole year. So there's a costing angle when it comes to AI usage as well, which is quite expensive uh uh in in certain cases, especially when you're trying to uh build production grade applications like an observability suite like we discussed, right? And uh customers are also cognizant that uh it's not as if these AI token prices are going to decrease uh in the near future. They're actually going to increase because there's a lot of profitability pressure on all the major model providers like OpenAI, Anthropic, etc, who have which filed for uh IPOs this year and are still not profitable at the unit economics level. So there's also that recognition that even if in theory it could be cheaper today, uh it would actually A, first of all require a lot of effort to do the reasons I mentioned and also uh you know the token prices could go up tomorrow very erratically and uh that could uh that could leave people in a precarious situation. 嗯。
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 反对 印度初创大量用开源 Grafana+Prometheus+OTel，并越来越多转向 Grafana 付费版，Grafana 想做完整可观测性提供商
我来自印度，许多印度初创公司使用开源的 Grafana 加 Prometheus 堆栈以及 OpenTelemetry，且越来越多的公司正在转向 Grafana 的付费版本。他们希望成为完整的可观测性提供者。这是一种威胁，尽管我不会称之为新——他们已经存在多年。
来源背景：标题写明 'Datadog离职专家' ，作者 Lantian； 自述 '我来自印度'；内容与 Merit Datadog 产品进展调研 IV (2026-06-22 音频) 逐段对应，判断为同一专家同一场。局限同音频 IV：前员工，口头区间，可能 MNPI，偏看多。 - 口径不齐 被问是否提价，CEO 只答“用量制，数据多了自然计入”，没有回应单价
But the cost issue though in terms of your margins, are you getting challenged from people saying, hang on a second, like I want more data, more data, more insights, and you've got to manage your costs. Are you increasing your prices or... No, we're usage-based. So if customers send more data or if they use more, basically, that's built into the model. We're not seat-based. Like a lot of the traditional SaaS companies are seat-based and they're challenged in that maybe they get less seats because they have more non-human agents using their systems and maybe they need to bundle more functionality within those seats. We're usage-based.
- 支持 企业自建观测栈目前不可行，也没人在讨论；难的是运维/补丁/升级而非开发
Sid Roy: No, in my opinion, it's hard to predict the future, especially with AI advancing so quickly. But right now, building your own observability stack is not a viable option for the enterprise, nor is it something that I hear anyone in the enterprise even talking about. Because the truth is, if we wanted to, we could get on a Zoom meeting and all of us together, we could vibe code an observability solution. We could take OpenTelemetry, we can build it for our own specific use cases, but building an observability solution and then deploying it and managing it are 2 vastly different situations.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 反对 Databricks、Snowflake、ClickHouse 等数据平台会成为 Datadog 的更大威胁，因为存储和计算是增长最快的成本
So it's not a data set that's just relegated. It's moving away from a data set that was just relegated to IT users and now being considered as part of a first-class citizen in league with like business data, revenue data, supply chain data, customer data, especially as people just start to do more agentic operations. So companies like a Databricks, a Snowflake, a ClickHouse, they are going to become more of a threat to the Datadogs, the Dynatraces, the Grafanas, the Ciscos, because customers are going to be looking, how do we scale our compute? How do we scale our storage? Because those are the costs that are growing the fastest, because of the growth in data. And then on the flip side, when companies like a Databricks, an Azure Data Factory, a Snowflake, frankly, their AI ML models.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 反对 UBS：DDOG 靠整合拿份额，但定价阻力仍是主要摩擦
UBS maintains Buy after conversations with seven customers and three partners reinforced that observability remains a priority spend area as enterprises accelerate AI deployments. Checks improved versus late 2025/early 2026, with DDOG gaining share through vendor consolidation and expanding into hybrid environments, although pricing pushback remains the main friction point. UBS still sees 30%+ growth in 2026 and expects another acceleration in 2Q, but acknowledges the setup is tougher after the stock’s rally and valuation near 20x revenue and 75x FCF.
来源背景：TMT Breakout (TMTB) 日报，转述 BNP Paribas 季度经销商调查与 UBS 客户检查 (7 位客户 + 3 位伙伴) 摘要。二手转述，看不到底稿。 - 口径不齐 Bits AI 改为 AI credits 定价，1 credit=$1，每次 SRE 调查约 7–8.5 credits
+ His ranking of DDOG’s AI products: Bits AI SRE, LLM Observability, Security Analyst, GPU Monitoring; LLM Observability and Bits AI SRE have both entered his customers’ top three * **Bits AI pricing has moved from per-interaction to an AI credits model, with 1 credit = $1 and official guidance of roughly 7-8.5 credits per Bits AI SRE investigation**
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 顾问 D：年花 $20m 的客户环境扩 3 倍也不会花 $60m，供应商只覆盖约 20% 生产环境
+ All vendor solutions combined cover only about 20% of production environments; the other 80% is homegrown and open-source tooling — what DDOG and DT are really competing for is that 80%, not each other + That 80% exists precisely because customers will not pay incumbent vendor pricing across their whole environment; his example: a client spending $20m a year on DDOG is not going to spend $60m when the environment triples in size + The answer is for vendors to use AI to distinguish which data matters and tier collection accordingly — he expects an 80/20 outcome where roughly 20% of data (transaction-level, agent-behavior-level) is collected in full, and the rest
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 顾问 D 预计 80/20：约 20% 数据全量采集，其余会被采样
he expects an 80/20 outcome where roughly 20% of data (transaction-level, agent-behavior-level) is collected in full, and the rest is sampled down
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 客户 E：DDOG 定价不灵活，已开始早期评估 Grafana 和 Prometheus
* Customer E: **inflexible DDOG pricing has already pushed them into an early-stage evaluation of Grafana and Prometheus**, leaning toward a mix of managed and self-hosted; but after years on DDOG the migration would be hard, so this remains very early and Chronosphere is not under consideration
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 Infinite Cardinality Metrics：AI agent 产生更多数据时客户 '不产生额外费用'——计费单位与数据量脱钩
And we've also announced that we are bringing BYOC to metrics and traces as well. In the digital experience space, joining monitoring automatically gives a single shared view for every critical user flow. And for custom metrics data, we introduced Infinite Cardinality Metrics, which allow our users to answer arbitrarily complex questions as they generate larger amounts of data with AI agents without incurring any extra costs. Finally, we launched a number of innovations to secure the AI stack and defend against a new class of AI powered attacks. AI Guard agent discovery finds and maps every known and unknown custom agent to security teams can see what is protected and what is not.
- 反对 管理层承认高基数导致账单不可预测是长期痛点，靠技术+商业重新打包解决
And so that's one of the reasons we've invested in all those products we've mentioned earlier. And also we're seeing some of the great returns on that products already. In terms of Infinite Cardinality, that's a – I would say it's been one of the longest standing source of frustration for customers, when sometimes, they send more data or they send more fine grained tags with their data. And they get some unpredictability on the bills because of that, because it increases the cardinality of the data we're getting. And we've solved that from a technical perspective and from a commercial perspective by packaging our metrics a little bit differently. And we think it's particularly important and relevant as customers are building more applications with AI. And as they want to send basically more tags, more information and ask more complex questions and get more fine grained answers to those questions.
- 反对 AI 客户关注点 3–6 个月前转向成本优化
And so we're well positioned there, but we think this market is going to change quite a bit and the preoccupations of customers will also change over time. So for example, last year, our customers were mostly trying to validate correctness and validate that they were getting some form of outcome that it could then scale up. I would say three to six months ago, the focus has moved quite a bit towards cost. Now customers were spending a lot on AI and they were wondering how to optimize cost. And I think we'll see some variations in the concerns over time, as customers get further into the adoption and new problems emerge for them. I just want to add that when you look at what we described as some of our deals in the quarter and you look down our scripts and you'll see that a number of them have the AI products included and so that is indication that those large enterprises are using the platform and buying the AI products as well.
- 口径不齐 日志联邦查询 Databricks、ClickHouse；BYOC 让日志留在客户基础设施，并将扩展到 metrics 和 traces (可能降低单位收费)
In network monitoring, we launched Network Path and Network Configuration Management to trace changes that cause complex network issues. Within database monitoring, Bits Database Optimizer now automatically simulates and evaluates the impact of AI generated changes in order to optimize slow queries. In log management, federating logs enables users to query external data stores, including Databricks and ClickHouse. And with Bring Your Own Cloud or BYOC, customers can now use the full Datadog experience on logs that are kept within their infrastructure. And we've also announced that we are bringing BYOC to metrics and traces as well. In the digital experience space, joining monitoring automatically gives a single shared view for every critical user flow. And for custom metrics data, we introduced Infinite Cardinality Metrics, which allow our users to answer arbitrarily complex questions as they generate larger amounts of data with AI agents without incurring any extra costs.
- 支持 BYOC 赢得迄今最大单：PB 级替换老日志工具 (在线媒体公司，多年期 $30M+ TCV)
This customer chose to standardize on Datadog across its business, displacing four commercial and internal tools. Datadog also proved value beyond core observability with product analytics, CI visibility, data observability and cloud cost management. This deal includes our largest win to-date for Bring Your Own Cloud, displacing their legacy commercial logging tool at a petabyte scale. And finally, we signed a nine figure renewal with a leading AI company. This long time, very large customer uses 17 Datadog products to enable unified visibility on production workloads at a very large scale, albeit with a usage reduction starting in Q3, which we considered in our guidance and which David will speak to.
- 反对 毛利率 79.6% (Q1 80.2%、去年同期 80.9%)，连续下行；公司称预期不变，约 80% 上下
Unless otherwise noted, all metrics are non-GAAP. We have provided a reconciliation of GAAP to non-GAAP financials in our earnings release. Our Q2 gross profit was $892 million for gross margin of 79.6%. This compares to a gross margin of 80.2% last quarter and 80.9% in the year ago quarter. As we've discussed in the past, our gross margin varies from quarter-to-quarter with investments into innovations for our customers, offset by efficiency efforts. There's no change in our expectations for gross margin, which has been in the 80% plus or minus range historically. Q2 OpEx grew 26% year-over-year versus 31% last quarter and 36% in the year ago quarter. We held our DASH user conference in June, and as expected, the event cost about $15 million.
- 反对 Funda 认为 Chronosphere 在 OpenAI 内部抢份额，价格比 DDOG 低 30–40%
* 26Q2 revenue grew 35.6% YoY, in line with buy-side expectations of 35-36%, accelerating 3ppts QoQ. However, the US$46m beat was slightly below the US$50m last quarter. * 26Q3 revenue guidance was a touch light at 29.5% at the high end, where expectations were looking for a 3-handle. Applying a typical 5ppts beat would imply a marginal sequential growth deceleration in 26Q3. * **The miss appears mostly driven by usage optimization at OpenAI, DDOG’s largest customer. For 26Q2, we estimated a 20% QoQ decline in OpenAI’s spending, on which market consensus appeared too optimistic. Management confirmed it has renewed a 9-figure contract with OpenAI, but the latter’s usage will decline sequentially in 26Q3. We believe Chronosphere is gaining share within the OpenAI platform, as it prices 30-40% lower than DDOG.** * Core/non-AI usage growth continued to accelerate to high-20s from mid-20s, driven by strong cross-sell momentum and increasing usage from cloud migration and AI app development. * Full-year revenue guidance was raised by more than the 26Q2 beat flowed through, to 30-31% YoY, in line with expectations. * The margin trend remains strong — GPM dipped below 80%, but OPM beat expectations on operating leverage.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 反对 访谈：LLM 可观测性转向 Braintrust，价格压力促使客户评估替代品
Datadog usage stays sticky and AI-supported, though LLM observability is shifting toward Braintrust and pricing pressure is pushing customers to evaluate alternatives.
来源背景：FundaAI 周报：DDOG 段落为 2Q review 摘要 + Funda 当周两次专家访谈的一句话摘要 (某全球云通信平台的工程高管——疑似 Twilio，因标题标 TWLO；某全球金融信息服务商经理)。二手转述，看不到底稿，访谈全文在 Funda 平台未附。 - 反对 扩量换折扣：按量和按期限定价，客户扩大承诺有价格优势 (单价随量下降)
There's essentially capacity planning we do together. And then for the most part, you can tell with the net retention that they use more than they've committed to. And so, the motion is we then work with them to figure out what their next consumption is going to be, and they have an incentive to do it because we're volume-based pricing and term pricing. So, most of the time we have within the contract – they have a set amount, but it's in their advantage to expand that. And so what we do is we have long histories by customer of what happens. We know because every day we see the usage.
- 反对 Datadog 主动帮客户减少日志写入 (ZIRP 时期有过度使用)
And then and then we work with that. We've gotten pretty good at helping clients use it. I would say, in the bubble after COVID, there was – in the ZIRP side, there was probably some usage that was over-usage. But we step in and we help clients then use it effectively. We help them figure out what logs to put in and what logs not. And then in terms of guidance, what we're able to do and that's how we can beat and raise is we take that growth rate that we see over a long period of time and we discount it, so that we have that cushion. And then, what's been proven out in the beat and raise is that the clients spend more than the amount we're discounting.
- 支持 Snowflake 可观测性目前难与 Datadog/Dynatrace 竞争
Yeah, this is an interesting one. Um, so, I think this was at the beginning of the year when we had a lot of, um, a lot of, uh, go to market, you know, training and a lot of pre-prep in terms of, in terms of the observability platforms with Snowflake. Um, it's very difficult to compete with Datadog, Dynatrace and Datadog. You know, those two vendors are very, very, uh, very, very good at what they do. Um, especially Dynatrace, they're they're in the enterprise, Datadog more so in mid-market and small enterprises. They're very, very competitive.
来源背景：未具名专家，从'我们这边的客户''我们围绕 Snowflake 的可观测性平台开展了大量市场推广培训'推测为 Snowflake 内部 GTM 人员或紧密合作伙伴 (未明示)。视角局限：站在 Snowflake 一方；只看自己客户，未必了解 Datadog 内部数据。 - 支持 未见业务从 Datadog/Dynatrace 转向 Snowflake
It's very difficult to compete with them, very, very difficult. Um, but it it, you know, it it really depends on what value Snowflake brings. We've not seen much business move away from either Dynatrace or or Datadog, um, to Snowflake. Um, maybe next year, for this year there's really no, not much, uh, purposeful, um, you know, acquisition of new business, uh, from a from an observability platform perspective. 明白,好的。
来源背景：未具名专家，从'我们这边的客户''我们围绕 Snowflake 的可观测性平台开展了大量市场推广培训'推测为 Snowflake 内部 GTM 人员或紧密合作伙伴 (未明示)。视角局限：站在 Snowflake 一方；只看自己客户，未必了解 Datadog 内部数据。 - 中性 Snowflake 进入可观测性要等企业 AI 工作负载规模化，今年不会有实质变化
回 答 者：会，肯定会。最终，我认为 Snowflake 将在可观测性方面发挥作用，但这可能要等到更多 AI 工作负载得到采用之后。从客户角度看，AI 工作负载的增长最终将推动价值提升，但我们需要先看到企业环境中的 AI 工作负载真正增加。届时，我们才能制定市场进入策略，形成正确的市场信息和定价，从而展现其带来的价值。不过，我认为今年不会出现任何实质性变化。AI 工作负载需要进一步增加，Snowflake 才会在演进过程中获得明显增长。
来源背景：未具名专家，从'我们这边的客户''我们围绕 Snowflake 的可观测性平台开展了大量市场推广培训'推测为 Snowflake 内部 GTM 人员或紧密合作伙伴 (未明示)。视角局限：站在 Snowflake 一方；只看自己客户，未必了解 Datadog 内部数据。 - 支持 用 AI 自建观测在规模上很贵，需要自建模型，买比造划算
**EE** [00:09:52]: I think it's going to be expensive at scale. There's the traditional build vs buy argument, obviously, where building in-house is not really free. You still pay for a lot of infrastructure and you still pay for full-time employees. Even if AI doing that, someone needs to build it and maintain it, but cost could be the more significant thing here on that sense, especially at scale. If you want to do it efficiently, you need to build your own models probably and use a combination of different methods, which is small language models and large language models and industry-specific models, and building models in-house, that's a whole different thing. I think it's not just like building an application and using Claude or OpenAI because then the cost could really be very high, I think, at scale of telemetry, especially of large organisations.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 反对 大型金融科技客户以支持 OTel 为条件，Datadog 被迫投入大量资源并改变心态，还在考虑按部署方式调整定价
**EE** [00:29:21]: I don't know if you're referring to OpenTelemetry or something that's more specific to LLM observability, but Datadog has definitely seen that or there was a lot of chatter about threat from the OpenTelemetry, for example. This is probably similar. I think it took some time for sure or maybe it took them time until they realised, "Okay, this might cost us customers," or customers are really asking for it. There was a big fintech company that was kind of a condition. You have to support OpenTelemetry in this way or that way. Datadog really had to allocate a lot of resources and even change its mindset. Initially, it was like we'll support OpenTelemetry, but we're not going to talk about it. If someone wants it, then fine. By the time I left, it was also about, "Okay, how do we need to rethink our pricing for different types of deployments, if it's OpenTelemetry or something like that," because after all, it's not a one size fits all. They are looking at it. Just going back to the proprietary agents, I think, again, they're incorporating some open source, OpenTelemetry support into their proprietary agents. They're trying to give the flexibility definitely for customers to choose, but there are trade-offs. I think the defensibility, like you say, it's not necessarily there in the instrumentation and the ability to collect LLM traces, etc.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 反对 Snowflake/Databricks 自建 Token/成本监控和 Agent 评估工具，想替代 Datadog
同时，我们发现 Snowflake 和 Databricks 在本季度采取了更加积极的策略，开始逐步向平台型公司转型。一方面，它们通过 Postgre 逐渐进入 MongoDB 的市场；另一方面，也开始建立 token 消耗监控、成本监控以及 Agents 评估工具。它们希望在客户内部逐步替代 MongoDB、Datadog 等外部应用。
来源背景：Merit 研究团队 (路演主讲人，未具名)，转述渠道商和 Snowflake 专家的观察并加上自己的判断。视角局限：二手汇总，不是一手专家；主题是 Snowflake/MongoDB。 - 口径不齐 大客户短期不会替代，小公司可能先被替代
我们认为，这一进程不会特别快。MongoDB 和 Datadog 的架构相对完善，而 Postgre 类数据结构与 MongoDB 类分布式架构相比，在高并发场景下的处理能力仍存在较大差异。因此，在大型客户内部，短期内不一定会形成替代；但在部分小型公司内部，Snowflake 可能会逐步替代 MongoDB、Datadog 等产品。
来源背景：Merit 研究团队 (路演主讲人，未具名)，转述渠道商和 Snowflake 专家的观察并加上自己的判断。视角局限：二手汇总，不是一手专家；主题是 Snowflake/MongoDB。 - 反对 证伪条件：云厂商或模型厂商把观测与治理免费打包
有六种情况会证伪本文的积极判断。Agent 激活量持续增长，NRR、cRPO 和自由现金流却连续两个季度没有改善；信用点收入被席位流失、降价和推理成本抵消；客户增长必须同比例增加驻场人员；供应商 AI ARR 上升，客户每成功结果总成本和项目回收期却没有改善；云厂商或模型厂商把身份、观测与治理免费打包，第三方控制点被绕过；公司不断扩大 AI 收入口径，旧口径无法复核。
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 - 中性 大客户承诺在各产品间可互换 (fungible)，可把主机换成 AI 能力，或把日志换成 APM
If we need to repackage that in different ways, attach that to different parts, such as data volumes, the host or anything else, that's something we can easily do. Now, when you look at the reason customers consolidate on us, one of the reasons is that it's a lot easier for them to have one big pool of commitment and spending with one vendor, as opposed to managing 12 different pools with different companies and have to make sure they hit the right number of everything. So, when customers contract with us, especially large customers, what they use across different parts of Datadog is fungible. So, they can say, you know what, we're going to actually – we changed our mind, we don't need as many hosts, but we need more smarts, or we don't need – we're going to use more APM and less logs because it's more efficient in different ways. Customers don't have to care.
- 口径不齐 被问量增带来价格通缩和 OTel 商品化，CEO 说 OTel 降低接入摩擦、带来更多负载，是顺风；未回答单价
See, if you look back at the – or if you go and look at a transcript of our latest earnings call, we – I think we have, we call out like five or six customer consolidations. And all of those typically involve a number of open source products that they consolidate on us. In terms of OpenTelemetry, like – it's amazing. Like it reduces the friction to instrument. It gets us more workloads faster. It makes customers more confident in expanding and getting both feet into our platform. And so, that's great. So, we support that, we invest in it.
- 中性 CEO 用“维持毛利率和价值”回应通缩问题，没有给价格数据
And we're glad it's there and it's actually a tailwind for us. So, we feel great about that. If you look at our history, we've been pretty good at maintaining margins. We've been pretty good at maintaining the amount of value we can deliver to customers. Because, again, when you sell, there's only two reasons customers buy your product, they save money or they make money. And we're very good at making that case and demonstrating it.
- 中性 Datadog 内部把 OTel 视为机会而非威胁，自信能赢正面竞争，但承认 OTel 让迁入迁出都更容易
JC: Yeah, I was not particularly close to the OTel org or initiatives. My organization did take on some work in support of those initiatives. But from my view, OTel was definitely viewed as more of an opportunity as opposed to a threat at the time. The reason being that Datadog believes that they have and are the best-in-class product. And if they, by definition, by supporting OTel, they make it easier for people to migrate both to Datadog and off of Datadog in theory. But Datadog believes they are going to win that head-to-head fight against other providers and believe that there will be more adoption of Datadog as opposed to adoption away from Datadog.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 口径不齐 原生 OTel 让换后端比重新埋点容易，但不是改几个开关就行
JC: I think it's a mix of both. I'm quite sure that it is not literally a flip a couple flags here and change some agent configuration and boom, it works. But it certainly is easier than having to re-instrument all your systems and services in a completely different library or something like that.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 反对 选 Grafana/Dynatrace/自建而非 Datadog 的首要原因是成本；亲见上家公司把日志从 Datadog 迁到更便宜方案
JC: I have to imagine that the primary driver there is cost. Datadog is a premium product and depending on how you use it, it can be, will be quite expensive. So I imagine that is really the primary driver there. I can, from my own experience, back at a previous company, this was in probably around 2021, there was a big push to centralize all of our various observability tooling into Datadog. That was when the company had lots of money and everything was going well. Fast forward, 6 months, a year later, things were kind of less rosy and there was a push to move some of those observability workloads out of Datadog, specifically logging to cheaper alternatives. So that does happen. I have seen that happen.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 反对 Datadog 内部对是否提供降量/成本提醒能力有长期张力，因大账单引发客户不满才让步 (自我蚕食)
JC: Yeah, I do know that you brought up internal tension in terms of OTel before and I do know that this specifically is a topic where there was a lot of internal tension around if customers... if Datadog should really be notifying customers and allow them this capability. I think... over time and enough unhappy customers over very large bills, I guess the side of yes won out, but it took some time to get there.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 反对 Infinite Cardinality：指标在进入前先筛选；Datadog 主动把客户转到该 SKU，称能提高留存和毛利 (收入影响未说)
I'm using very simple language. But – and then that is good for the customer because they get more value in their metrics. That's also good for us because it doesn't help us to have metrics or even logs that flow in and are costly to us. So, there are many, many examples. I think it's a very good point of how we've been evolving the platform to meet the client at the value point. And what we've been doing is proactively converting those customers that can benefit from Infinite Cardinality to our SKU, which has essentially been retentive and also margin-enhancing for us, given the weight on our platform of all that – of large cardinality of our metrics. Yeah. The other pricing announcement, it's not really a pricing announcement, but I think we could frame it as driving value to customers, Federated Logs.
- 反对 Federated Logs：日志可存在 ClickHouse 和 Databricks，不必全部进 Datadog
It's in a group of innovations where we used to have the – all the data have to flow into Datadog to be used, but we know that's not efficient. So, Federated Logs, Data Observability Pipelines, Bring Your Own Cloud. What this means in the case of Federated Logs is that we know that not all logs have to flow into Datadog to enhance the platform, that we can link into and logs can be stored, in this case, it was ClickHouse and Databricks. So, what we're doing is, we'll say (00:20:00) we understand certain – what we want to be able to have more data available in the platform without having to port all of your data into Datadog. That, we've proven, and we've gotten feedback from clients, enhances the value of the platform, eliminating one of the objections is, I have to bring all my data into Datadog even though I'm storing it elsewhere.
- 口径不齐 被问预算是否流向 ClickHouse/Databricks，CFO 说用户集中在 Datadog 时 ARPC 更高，并主动告诉客户数据发多了
We've analyzed that. We actually thought, okay, for – we want the – a map – the logs that need to be accessed there to be in Datadog, but we want the platform to be able to use as a central point. The more we have our users centralized on Datadog, we proven that we get higher average revenue per customer. So, to close Datadog off and force them to go to another platform is not as optimal as staying in Datadog all day long and using the data efficiently. We actually – getting back to how we've evolved, we actually go to clients and we say, you set this wrong. You're flowing too much data into Datadog that you don't need for this purpose. And it creates a much better long-term value of customer, and also, in many cases, is margin enhancing for us because to store all that data in Datadog that is never used also is a burden on our system. Every two to three quarters, we all have to ask you about some new competitor in the space...
- 中性 一系列降本创新：flex logs、frozen logs、metrics without limits
...addresses one of the budget pushbacks that you would have gotten from customers. Definitely. So, I'll focus on answering Infinite Cardinality, but then I'll also mention it's one of a portfolio of technology innovations where we have met the customer, where they are, flex logs, frozen logs, metrics without limits. I have another side questions on that. Yeah.
- 反对 Grafana Cloud 相比 Datadog 通常省约 20%，自托管 Grafana Enterprise 节省可能大得多
MC: I think it typically varies, but you can see there's usually a reduction of around 20%. It's not as dramatic anymore as it was. I think overall that market is... there's a lot of competition. But I think overall, it's probably like a 20% savings if you're leveraging Grafana Cloud. Now, if you're using Grafana Enterprise in your own environment, the savings could be tremendous because you're then using your own resources, there's gear, whatever you want, and you don't need to use that. And there's a lot of advantages to that as well. So I think it really, if it's cloud versus cloud, I'd say... because you can't really deploy Datadog in your own infrastructure. I would say you'd see a 20% savings if it's Datadog, which is the cloud anyways, versus an on-premise deployment. I think it could be a lot bigger depending on the environment and how much data the metrics folks have.
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 反对 Datadog 定价仍与市场不匹配；如果团队能自己做 AI SRE agent，可以大幅降低工具支出
MC: Yeah, I think it just depends on the industry that they're in. I think that's a huge one if there's market pressures, etc. The biggest thing is being cost-conscious and also understanding that there are other priorities within organizations. I think for like Datadog, when I was at Grafana Labs, it was around cost savings. I think the pricing still doesn't align with the market and where other observability companies are or if their team can build these AI, these SRE agents, they can reduce their costs in these tools pretty significantly. So I think it's... adapting to the market and making sure that you're still competitive.
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 中性 OTel/厂商中立对他接触的企业不是主要驱动，只是 nice to have；但已埋点的客户迁出可省 APM 费用
towards that. That said, it wasn't a major driver for the companies I worked with. It was more of a nice to have. I think, depending on where folks are, it could be a much easier migration route off. It could provide them with a bunch of savings on application performance monitoring, but they need to be instrumented already. So if they're not instrumented, it's not as much of a value add, but if it is, then it definitely can provide a ton of value.
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 反对 企业按优先级分层：高优先应用留在 Datadog/Dynatrace，开发/测试环境、部分指标和日志分流到开源
MC: Yeah, it's a great question. I think some of it is very related to just how some of these companies are approaching their strategies for AI and how they're going to differentiate. So I think one is, is there existing tooling? Or can our tooling... can we separate these use cases, and where are the integration friction points? I think that's something that a lot of folks have started to look at. So a tool that is truly agnostic, right? So a Grafana, for instance, that has open APIs that can be consumed or can output data. That is a more flexible, yes, if you are moving everything to an open source or to a Grafana Enterprise type of tool or contract, you're locked in with Grafana. But that said, you're locked in on that contract, but you can still use open source, you can integrate with a lot more things, a lot more tools than you could if you were on a Datadog contract and you were only restricted to the Datadog ecosystem or Dynatrace. So I think that's one of the bigger things. And then also, when I say separation of use cases from an observability perspective, that is, can we offload our metric collection or a subset of our metric collection to a different tool, to an open source tool, or can we instrument specific applications with open telemetry versus having them instrumented with a Grafana or Dynatrace? I mean, a Datadog or Dynatrace or an AppDynamics, those are more expensive. So let's keep whatever,
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 反对 AI 编程工具让更多公司变成 'build shop'，更倾向开源工具
of ethos. But there's also just a ton of companies that are build shops and have moved more towards build shops, being build shops with the advent of the cloud codes and all of that. So I think those are also folks who are going to lean more towards the open source nature, open source tooling.
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 中性 大型金融机构年付 Datadog 2,000 万美元并质疑是否值得，找顾问评估
Bernd Harzog: Good morning, everyone, or good afternoon, depending on where you are. Briefly, I've been in this business for about 45 to 50 years. I've been on the product side of this business, primarily working for vendors. I've been a CEO of 2 venture-funded startups, sold 1 of them to Citrix. When I sold my last company to Citrix, I became a consultant. I... I consulted as a 1-person consulting firm. And I focus upon—it wasn't called observability back then, it was called APM back then—but I focus upon the APM industry or the observability industry. My clients are large customers and I help them evaluate and better use these products. I think a simple example of the kind of problem I help my large customers solve is I had a large financial institution call me up last year and say, "We're spending $20 million a year with Datadog, and we're not sure that this is the right thing to be doing. Can you help?" And the answer is, yeah, I can help with that. Or I've had somebody else call and say, "We're using New Relic today, and we're thinking we might want to change observability platforms. We want to evaluate what's out there today and then make a choice." ES: Can you run the evaluation process for us?
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 反对 给大客户降费第一步是 '找出愚蠢并消除'，如实验结束后未关闭的 1,000 个云实例仍在被监控
Bernd Harzog: Well, that's a contained enough problem so it's possible for someone to roll their own with open source. On the other end of the spectrum, I've got a client who's a major investment bank. They have 5,000 business critical applications that run in the cloud, in their own on-premise VMware environment, in their own bare metal high performance computing environment, on their mid-range systems and on their mainframes. Are they going to build their own monitoring for that incredibly heterogeneous mess? And the answer is no, they can't. It's not possible. So in between those 2 extremes lies all different kinds of shades of gray. So when customers seek to reduce their spend on observability, and this is, in fact, what I counsel them to do, and this is obvious once I say it, the very first thing I tell them to do is find stupidity and eliminate it. Examples of stupidity would be somebody decided to spin up 1,000 cloud instances up in Amazon or somewhere to conduct an experiment to see if something would work, and they implemented observability to watch the applications that are part of that experiment.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 反对 客户分散在 5 个事业部各自付费，合并后才发现观测账单过高 (CIO '震惊')
Bernd Harzog: Which is very often the case when it comes to business level applications. This kind of stuff happens because there is no centralized control of observability usage and observability spent. In the case of I have a client who spends $20 million a year with 1 particular vendor and the CIO said, "I was shocked to learn that we were spending this much money on observability." ES: Mm-hmm.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。
E05 新 AI 产品收到钱了吗
判断：还没有收到钱。产品发布很多、使用量增速很快，但收入层面管理层自己说"还看不到"，定价在改，渠道付费率低，客户 ROI 算不过来。安全是唯一有规模数字的新业务，ARR 刚过 $1 亿。
Bits AI SRE (现名 Bits Investigation)：告警触发后自动翻指标、日志、trace、变更记录，给出根因假设。2025 年 12 月 GA，2,000+ 客户用过，10 万+ 次调查。定价从"$500/月含 20 次调查"改成 AI credits (约 $1/credit，一次自动调查约 6.5 credits)。如果单位一致，单次价格从约 $25 降到约 $7，但两次口径可能不同，降幅需要核实。渠道端：北美 GSI 说他的客户里采用率不到 10%，每周处理不到 3 起事故；Funda 的大客户 D 说 30–40% 的情况 Bits 答不上来，他们的账单是每月 300 多次调查、每次约 $500 (这和另一位前销售说的"$500/月含 20 次"对不上，口径存疑)，而 30 人团队每月处理 5,000–6,000 张工单，摊下来每张约 $50，算不过来，2026 年 Bits 预算只有 $10–20 万。偏乐观的只有 Datadog 印度前员工"Jason"：Bits 占收入低到中个位数、环比 +50–60% (可能涉及内部信息，且和渠道看到的不一致)。
Agent/LLM Observability：按模型调用收费，Pro 版 $160/月含 10 万次。和 Braintrust、LangSmith、Arize 这些开发侧工具竞争，管理层自己说"要不要争、争多远"还说不清。一家年付约 $500 万的客户已把 LLM 可观测性给了便宜约 20% 的 BrainTrust；一家初创公司 CTO (前 Datadog) 在 RFP 里没选 Datadog，因为当时不支持 Gemini。
安全：ARR 超过 $1 亿，约四分之一客户在用，但在年付 $100 万以上的客户里，安全只占其 Datadog 支出不到 2%。Cloud SIEM 借日志业务进展最快；大型受监管银行不会用 Datadog 替换 Splunk 做 SIEM。
"退化成数据库"的风险：CEO 承认未来很多功能通过 MCP 交付给 agent，UI 只是其中之一；CFO 说"我们不是席位模型，按工作负载收费，谁来看无所谓"。这个说法对现在的收入是对的：只要数据在 Datadog 里，谁来读都要付钱。风险在于 Federated Logs、BYOC 这些产品已经让数据可以不在 Datadog 里。如果 agent 直接读 ClickHouse 里的日志、读 Datadog 里的实时指标，Datadog 收的是"实时那一层"的钱，单价和份额都会比今天低。
They were so impressed. They emailed us straight on that. So, it's pretty transformative for them. In terms of where it is on the results, you won't really see it yet on the results. We're actually in the process right now of repackaging these products. We used to have one SKU we sold for – specifically for investigations. And now, we're packaging it to a set of AI credits instead that can work across a lot of different surfaces, because as I said, we expanded the product to cover a lot of different surfaces. And so, you should – we probably will comment on that in the future. But, for now, there's a – the transformation's happening right now in terms of the way we're charging those. And by the way, if you zoom out, the world as a whole doesn't quite know how to charge for agents just yet.
A:The number of investigations Bits AI SRE would run is typically defined by priority. It does not make sense to employ it for every incident, as these customers still have dedicated teams. In some cases, we are running their infrastructure and technology. The number of incidents that Bits AI SRE processes is less than three per week.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。
+ Excluding agent cost, MTTR savings are ~20–30%; including cost and stability factors, the math does not work out + Even with MTTR savings of ~50%, may still be insufficient to justify Bits AI cost + Current bill: ~300+ investigations / month at ~US$500 each; the customer’s 30-person team handles 5,000–6,000 tickets / month, equating to ~US$50 per ticket on a fully-loaded basis — clearly below Bits AI cost * **Management view: DDOG core functionality will continue to be heavily used; AI agent is too early but the customer cannot afford to fall behind, so will continue trialing** — 2026 Bits AI budget capped at US$100–200k * AI observability: did not purchase DDOG’s LLM monitoring; views the product as expensive and capability-limited. The customer self-developed a Llama-3-based call center chatbot, monitored via distributed tracing on existing DDOG APM
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。
So, we're still in evolution. We're early on, definitely the way it's going to change where it's going to wind up, I don't know. But the good news is it probably doesn't matter. Meaning we're not a seed model. We are basically monetizing based on the workloads that go through and it really doesn't matter if those workloads are created or looked at by agents versus not. You're going to have to – we believe you're going to have to observe and secure this and probably all this is going to accelerate the pace of change, which means you're going to have to move from legacy. Legacy still is the biggest part of the market because it was there for 50 years. So, it's all probably going to be our friend, if history repeats itself.
Yes, there's going to be an evolution, that's certain. There's going to be a lot more automation. We see it today, like we see the -- all the signs we see point to everything moving faster, more data and more interactions, more systems, more releases, more breakage, more resolutions of those breakages, more bugs, more vulnerabilities, everything. So, we see an acceleration there. At the end of the day, humans will still have some form of UI to interact with all that. And a lot of the interaction will be automated by agent. So we're building the products to satisfy both conditions. So, we have a lot of UIs, and we are able to present humans with UIs that represent how the world works, what their options are, give them familiar ways to go through problems and to model the world. And we also are exposing a lot of our functionality to agents directly. We mentioned on the call, we have an MCP server that is currently in preview and that is really seeing explosive growth of usage from our customers. And so, it's a very likely future that part of our functionality is delivered to agents through MCP servers or the likes. Part of our functionality is directly implemented by our own agents, and part of our functionality is delivered to humans with UIs.
It is a consistently a harder job than writing the application itself. And that is why there's been this race (07:40) over time and everybody is going after it, including the leading companies in AI whether it'd be the AI builders or the hyperscalers, so that's one thing. The other thing I will say is that the kind of problems we solve are somewhat different in nature from what these fundamental models – financial models are built for, and operate at very different levels of data volume and real-time requirements. So we have five or six orders of magnitude more data than what is typically put in an LLM. We need to operate in near real-time. So the analogy I would give to understand where we stand with the problem we solve versus what these models do is compared to self-driving cars, so I mean, today, it's working pretty well. You have self-driving cars everywhere. You could, if you wanted, take a photo from your dashboard in your car and apply it to Claude and ask what you should do.
It's still very early days for Bits AI, so uh the revenue contribution is uh currently insignificant. The uh company's revenues are right now massive, right? Like they had their first uh quarter with a billion dollars of revenue. So in that uh broader scheme of things Bits AI would contribute in like the low single digits, low to medium uh single digits. So it's it's a driver of acceleration, but uh still a lot of adoption that uh can happen on Bits AI, very early days. But it's growing quickly, it's growing uh somewhere between 50, 60% uh quarter on quarter. So that's uh an encouraging sign and the growth honestly can even accelerate as they go full guns blazing uh now that the product also is getting much more uh uh uh improved feedback, right? But there's a lot of demand in the uh market for a tool like this and uh you know, a lot of our uh uh new business calls are uh revolving around the capabilities of Bits AI and I'm sure that uh you know, you'll see even more contribution from the uh product in the in the next few quarters. Mm, okay.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。
Cloud security has three components. It has how the cloud's working, it's called posture management, and set up to secure digital applications; two is cloud SIEM, how are you using logs and other information to investigate what's happening; and three is code security, how are you engineering security into the code. And so for – with – so not on-premise, not legacy, but for cloud applications, we've built this suite of the three products. I would say the one that we've probably made the most progress on is cloud SIEM. Some of the reasons it's off of our logs business and have an over $1 billion log business. We have a great, largest observability logs, and the end market has dynamics where we're been an innovator in it. So we've had two major strategies. One is attach the whole suite to cloud-native emerging companies, and two is attach cloud SIEM to more traditional companies where we have a strong observability logs.
展开本节全部 114 条原文 (支持 39 / 反对 35 / 其他 40，按时间)
- 口径不齐 Bits AI 是否单独收费还没定，可能主要靠带动平台用量变现
Yes. So -- I mean, look, it's super exciting. The feedback is very good on it. I mean, we've been collecting all the -- so I read one quote, we have dozens that look just like that that was sent to us by customers. And so that's very, very exciting. The -- we also started I think some customers buy -- come to it just to show value and to make sure we're on to the right product mix. And so we feel good that this is something that is high quality and we can monetize. In terms of the impact for next year, on the packaging side, I'm not completely sure yet whether the biggest impact will be seen from what we charge for itself or for the rest of the platform, that it gets what it is on the differentiation of Bits AI. I think that's more of a broader question of packaging and monetization of AI. And remember that we have a product that is usage based. So anything that drives usage of and adoption from customers is good for us and is very, very monetizable. But what we can tell you is this is differentiating. This is good. It works significantly better than anything else we've seen order in the market, and we are doubling down on it. We have many, many teams now working on deepening CSRs to making sure it goes further into the resolution doesn't just point to the issue, but it is the code have all these kind of things, working hard on that. We're also working on breadth, making sure that we train it on many more types of data, I mean, in terms of sources sometimes even systems that are also systems that are not tap so we can cut across to other systems our customers are using. So we are very, very aggressively developing Bits AI. It's resonating by well in the market.
- 支持 安全 ARR 增速升到 mid-50s (上季 mid-40s)
And this includes, in particular, a very fast ramp for Product Analytics which has already seen adoption by more than 1,000 customers. We also want to call out our security suite of products where we are executing and accelerating growth. Security ARR growth was in the mid-50s as a percentage year-over-year in Q3, up from the mid-40s we mentioned last quarter. We're starting to see success in including Cloud SIEM in larger deals, and we'll get back to that in a bit in our customer examples. And we're seeing positive trends beyond Cloud SIEM, including fast uptake of good security and an increasing number of wins in cloud security.
- 中性 Bits AI (推测) 在欧洲还没开始推广；他猜定价会按积分/消耗计，并和平台捆绑
A:不，没看到具体的。他们可能在试验。根据类似平台的经验，它很可能是基于积分的，也就是按消耗计费，并与整体平台捆绑。如果你有较大的平台占比，他们可以给折扣，并会试图引导你使用整体平台。我还没有看到任何官方数字。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 反对 Bits AI 在他的客户里采用率不到 10%，处于试点和生产混合阶段
A:At this point, a little under 10% of my clients-mostly in selected verticals-are using Bits AI in a mix of pilot and production. In those examples, we have deployed Bits AI for Security, and there is another component, which is AI Token Processing.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 中性 已付费客户每月在这些 AI 功能上的花费不超过 5–7.5 万美元
A:I believe they are using Security Analyst and the AI Token Pricing model. In my view, their monthly cost for these features is not more than 50,000 to 75,000 dollars. They are just getting started.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 反对 Bits AI SRE 每周处理的事故不到 3 起，低于官网每月 20 次调查的最低档
A:The number of investigations Bits AI SRE would run is typically defined by priority. It does not make sense to employ it for every incident, as these customers still have dedicated teams. In some cases, we are running their infrastructure and technology. The number of incidents that Bits AI SRE processes is less than three per week.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 反对 Bits AI SRE 还不能实质性改善 MTTR，只能把 MTTU 缩短至少 25%
MTTR measures the average time required to restore service after an incident. In large enterprises, it is computed over a fortnight or a month, and they identify common root causes. Bits AI SRE today is far from materially contributing to MTTR; it may do so in a few quarters. What is emerging with Bits AI SRE and similar tools is MTTU-mean time to understand. These tools triage information and hand it off to a human for further processing. Because of technological nuances, the benefit is in triaging, which accelerates the ability to understand incidents. Today, in terms of efficiency improvement, it is at least 25%.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 中性 他推测到年底 Bits AI SRE 系列可能占收入的 5–7%
In the second half of this year, as adoption of Bits AI or the wider Digital AI product portfolio grows, AI products will become more mainstream in their use, even if not dominant. Even if the Bits AI SRE family starts contributing five to 7% of revenue by the end of the year, that would be notable for a product launched in June of last year.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 反对 多数客户仍在 pilot/PoC，付费的很少
A:Yes, it is a mix of both. The pilots-beta pilots-are effectively businesses. The majority are in pilot/PoC, and very few are actually paying, and they do have a commercial agreement with Datadog.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 中性 Enterprise 档客户有免费 token，有 3 个客户超额，但还没单独收费
+ Enterprise tier客户有免费的token，目前有三个客户用超了免费的量但还未正式单独收费
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 反对 Bits AI SRE 的初步反馈平淡，低于去年 demo 时的预期
**客户初步反馈比较平淡，认为低于去年demo时的预期**
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 反对 没有看到针对 AI 模型的额外 license，AI 功能捆绑在分级许可里
他们在带来的一些新的 AI 项目上也有一定杠杆作用，特别是在可观测性方面。即便如此，我们没有看到针对 AI 模型定义的额外许可；他们已经将这些功能捆绑在分级许可中。我们没有看到立即产生巨大影响。我们听到一些风声，客户也在沿这些方向质疑我们，但 Datadog 已经有一些应对方案。它可能不会成为一个大因素，但值得关注。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 口径不齐 Watchdog 等 AI 功能大多不单独收费，收费的是 LLM 可观测性
监控将是总体支出和供应商所提供价值的关键方面。随着所有变化，这将驱动 Datadog 的新收入。他们对 Watchdog 等 AI 功能感兴趣；他们大多不为这些功能单独收费。他们在收费的是 LLM 可观测性和类似能力。根据观点，这个比例可能是 5–6%；我只是举例说明。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 中性 BCI (Bits AI) 按 token 配额收费，他的客户中约 3 家买了额外 token
A:是的。我的意思是，在代理管理或基础设施监控方面没有单独收费。收费是基于幕后的令牌。客户会获得一定配额的令牌可以使用，尽管我不记得确切数字。超过该配额的部分需要额外付费。在我们的客户中，大约有三位购买了额外的令牌容量。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 反对 Datadog 宣称约 200 个开箱用例，实际可行的只有 20–30%
A:是的。每个供应商关于代理都有很多兴奋点。但其价值并不像最初预期的那样强。Datadog 表示大约有 200 个开箱即用的用例。 我们看到在这 200 个中，可能有 20–30% 是可行的。对于其他用例，客户必须更有效地管理和存储他们的数据。这就是为何它显得让人失望：它并非对所有事情都有效。在某种程度上，这并非我们特有；SAP、Oracle 等类似供应商也面临相同约束。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 支持 Bits AI SRE 12 月 GA，过去一个月 2000+ 试用和付费客户跑过调查
So first, let's look at AI for Datadog. These are AI products and capabilities that make the Datadog platform better and more useful for customers. We launched Bits AI SRE Agent for general availability in December to accelerate root cause analysis and incident response. Over 2,000 trial and paying customers have run investigations in the past month, which indicates significant interest and shows great outcomes with Bits AI SRE. And we're well on our way with Bits AI Dev agent, which detects code level issues, generates fixes in production context and can even help release and monitor a fix and Bits AI Security Agent, which autonomously triages SIEM signals, conducts investigations and delivers recommendations. The Datadog MCP server is being used by thousands of customers in preview.
- 口径不齐 CEO 承认：未来部分功能将通过 MCP 交付给 agent，UI 只是其中之一 (界面价值可能被摊薄)
Yes, there's going to be an evolution, that's certain. There's going to be a lot more automation. We see it today, like we see the -- all the signs we see point to everything moving faster, more data and more interactions, more systems, more releases, more breakage, more resolutions of those breakages, more bugs, more vulnerabilities, everything. So, we see an acceleration there. At the end of the day, humans will still have some form of UI to interact with all that. And a lot of the interaction will be automated by agent. So we're building the products to satisfy both conditions. So, we have a lot of UIs, and we are able to present humans with UIs that represent how the world works, what their options are, give them familiar ways to go through problems and to model the world. And we also are exposing a lot of our functionality to agents directly. We mentioned on the call, we have an MCP server that is currently in preview and that is really seeing explosive growth of usage from our customers. And so, it's a very likely future that part of our functionality is delivered to agents through MCP servers or the likes. Part of our functionality is directly implemented by our own agents, and part of our functionality is delivered to humans with UIs.
- 中性 CEO 认为 LLM 擅长事后分析，Datadog 的护城河在流式实时分析，"还没做到"
And you also need to be able to run specialized models that can act on that data as opposed to just taking everything and summarizing everything after the fact 10, 15 minutes later. And that's what we're uniquely positioned to do. We are building that. We're not quite there yet, but we think that a few years from now, that's what the world is going to run, and that's what makes us significantly different in terms of how we can apply anomaly detection, intelligence and preemptive resolution into our systems.
- 支持 过去一个月超过 2000 个客户在测试或部分付费使用 Bits AI SRE，一些七位数续约里包含了 Bits AI
* 超过2000个客户在过去一个月在测试或已部分付费Bits AI SRE, 部分七位数续约大单里面已包括Bits AI 项目 * 根据我们此前调研，Bits AI Dev agent和Bits AI Security Analyst都可能于1Q26末GA; 预计在Investor Day将会披露更多细节
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 中性 AI 群体客户中 LLM Obs + Bits AI 不到其 DDOG 支出 10%
+ AI cohort represents ~30% of the expert’s DDOG book by budget + Such customers value Bits AI’s native operation at the telemetry layer, and view it as well-positioned to integrate more deeply into AI-era workflows as AI gains importance + AI cohort customers typically use 2–3 LLMs in parallel; feedback on Bits AI is **“broadly in line with expectations”** — it has not fully automated problem resolution, but no better product is currently available + AI cohort customers spend ~US$200k / customer on DDOG on average, of which LLM Observability + Bits AI is <10% (LLM Observability 6–6.5%, Bits AI 3–3.5%) + **<10% of customers paying, ~40% in POC; year-end target is 16–22% paying customers** * Non-AI cohort customer feedback on Bits AI is muted
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 中性 Bits AI 付费客户 <10%，约 40% 在 POC；非 AI 客户付费 <2%
+ AI cohort represents ~30% of the expert’s DDOG book by budget + Such customers value Bits AI’s native operation at the telemetry layer, and view it as well-positioned to integrate more deeply into AI-era workflows as AI gains importance + AI cohort customers typically use 2–3 LLMs in parallel; feedback on Bits AI is **“broadly in line with expectations”** — it has not fully automated problem resolution, but no better product is currently available + AI cohort customers spend ~US$200k / customer on DDOG on average, of which LLM Observability + Bits AI is <10% (LLM Observability 6–6.5%, Bits AI 3–3.5%) + **<10% of customers paying, ~40% in POC; year-end target is 16–22% paying customers** * Non-AI cohort customer feedback on Bits AI is muted
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 Bits AI 主要价值是更快 triage，根因仍需人确认，目前更像 nice to have
+ Expects 3,000–4,000 customers using Bits AI, with each contributing US$20–30k incremental ARR per year * Customer feedback: **Bits AI’s main value is faster triage, but root-cause confirmation still requires humans; today it remains more “nice to have”**
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 中性 70–80% 客户试用过 Bits AI，约 20% 转付费；定价 $500/月 20 次完成的调查
+ 70–80% of customers have trialed Bits AI, of which ~20% have converted to paid + Pricing: US$500 / month for 20 completed investigations; most customers underutilize early on, making the per-use
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 大客户 D：30–40% 情况 Bits AI 答不上来或只给泛泛信息
30–40% of cases Bits AI cannot answer or only returns generic information — content that humans can already retrieve from DDOG’s existing data without an agent
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 大客户 D 的 ROI 算不过来：每张工单全成本约 $50，明显低于 Bits AI 成本
+ Excluding agent cost, MTTR savings are ~20–30%; including cost and stability factors, the math does not work out + Even with MTTR savings of ~50%, may still be insufficient to justify Bits AI cost + Current bill: ~300+ investigations / month at ~US$500 each; the customer’s 30-person team handles 5,000–6,000 tickets / month, equating to ~US$50 per ticket on a fully-loaded basis — clearly below Bits AI cost * **Management view: DDOG core functionality will continue to be heavily used; AI agent is too early but the customer cannot afford to fall behind, so will continue trialing** — 2026 Bits AI budget capped at US$100–200k * AI observability: did not purchase DDOG’s LLM monitoring; views the product as expensive and capability-limited. The customer self-developed a Llama-3-based call center chatbot, monitored via distributed tracing on existing DDOG APM
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 反对 大客户 D 未买 DDOG LLM 监控，认为贵且能力有限，自建 chatbot 用现有 APM 追踪
+ Excluding agent cost, MTTR savings are ~20–30%; including cost and stability factors, the math does not work out + Even with MTTR savings of ~50%, may still be insufficient to justify Bits AI cost + Current bill: ~300+ investigations / month at ~US$500 each; the customer’s 30-person team handles 5,000–6,000 tickets / month, equating to ~US$50 per ticket on a fully-loaded basis — clearly below Bits AI cost * **Management view: DDOG core functionality will continue to be heavily used; AI agent is too early but the customer cannot afford to fall behind, so will continue trialing** — 2026 Bits AI budget capped at US$100–200k * AI observability: did not purchase DDOG’s LLM monitoring; views the product as expensive and capability-limited. The customer self-developed a Llama-3-based call center chatbot, monitored via distributed tracing on existing DDOG APM
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 支持 产品层级：26 个产品中 5 个 ARR >$100M，3 个在 $50–100M
Our total ARR now exceeds $4 billion, and our quarterly revenue exceeded $1 billion for the first time in Q1. This is a big achievement for all of us at Datadog, and is a product of years of investment in building, and innovating for our customers, but we're still just getting started. Of our 26 products, five are over $100 million in ARR, and another three are between $50 million and $100 million ARR. We're working hard to build and deliver further growth in those products, and this leaves 18 other products which are earlier in their life cycles. We believe each has a potential to grow to more than $100 million over time.
- 支持 LLM Observability 替换客户原有的独立工具 (招聘平台案例)
Next, we signed a seven-figure annualized expansion for an eight-figure annualized deal with a leading online recruiting platform. This customer is centralizing on Datadog to reduce complexity, drive developer velocity, and improve efficiency. With this expansion, they will replace a standalone tool with Datadog LLM Observability to correlate LLM signals with APM and user experience data. This customer will grow to 16 Datadog products including Datadog MCP server. Next, we signed a seven-figure annualized expansion for an eight-figure annualized deal with a Fortune 500 bank.
- 中性 Bits AI Security Agent 把数小时的调查缩到 30 秒 (效率数字，无收入)
With MCP server, developers access live production data to debug their applications directly in their AI coding agent or IDE. We delivered Bits AI security agent, which autonomously triages Datadog Cloud SIEM signals, conducts in-depth investigations of potential threats, and delivers actionable recommendations. We've seen Bits AI Security Agents reduce investigations that could take hours to as little as 30 seconds. We also shipped Bits Assistant now in preview, which helps customers search, and act across Datadog using natural language prompts. Moving on to Datadog for AI.
- 口径不齐 Bits AI SRE 目前按调查次数收费，长期定价模式未定；管理层称因为本来就是用量制所以“不重要”
And what do you think that would look like once you start being able to do that? I think is it charging per like issue that you've proactively resolved or how would you think about that? I mean, right now, we charge per investigation for this. Yeah. Long-term, we don't actually know what the model is going to be... Yeah. ...in part because the broader market is still figuring out how to package intelligence, and so it's not clear yet what customers relate to the most there? And so, I guess, we'll see. The short of it though is, for us, it doesn't really matter because we have a usage-based business model in general. Yeah. And so, it doesn't matter whether we have a new dimension in usage that relates to this particular type of investigations or whether that gets attached to other parts of usage we have whether it's on the data volume we process, the number of events, the footprint our customers have in their cloud like there is a number of different ways to look at it.
- 支持 客户要求 Bits 跨非 Datadog 系统 (其他日志、安全、开源工具) 调查，Datadog 在做跨系统连接
Now, give me a button to fix it or maybe even better fix it for me and then tell me about it. So we're working on that. Second way (00:22:04) we hear from customers is okay, this is great, this works really well, but I want it to work across my other systems, so it should work to solve issues within Datadog but also I'm using some other logging system, I'm using some other security tools, I'm using all these different things, I have a lot of open source, can it work across all of that? So we're also pushing so that Datadog actually connects to all those and Bits can investigate across all of those different systems. Another area we're investing in is in getting more proactive and predictive.
- 支持 安全：代码安全和 Cloud SIEM (日志后端 + Bits AI 安全助手) 需求好，未给数字
So, it's exciting. I think there's two areas, in particular, where we find great satisfaction over the past couple of quarters. One is code security. I mean, that whole field is upended by coding models, and so there's so much more to be done there, and we find a great amount of demand for it. So that's exciting. The second area, we find a lot of success in our cloud SIEM product.
- 支持 Bits AI SRE 拉动强，客户要求更快走向自动修复
Pomel: *“We see a ton of pull for it. One thing we keep hearing from customers is, ‘I want it to go faster into auto-resolution.’ Initially we were worried that if it does too much, it creates a trust issue. But with AI taking hold in pretty much every single part of our customers’ businesses, I think that train’s leaving the station. Folks are getting comfortable with automation. They’re pushing for more and more end-to-end automation. ‘Okay, it’s fine, you told me what was the issue, why it was there, who introduced it, how to fix it — now give me a button to fix it. Or maybe even better, fix it for me and then tell me about it.’”*
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 口径不齐 客户要求 Bits 跨其他日志系统、安全工具、开源一起调查
*“A second area of pull — ‘This is great, this works really well, but I want it to work across my other systems.’ So it should work to solve issues within Datadog, but also my other logging system, security tools, open source. We’re pushing so that Datadog connects to all those and Bits can investigate across all of those different systems.”*
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 支持 Cloud SIEM 与 code security 是安全上的两个增长点
Pomel: *“There are two areas where we find great traction. One is code security — that whole field is upended by coding models. There’s so much more to be done there, and we find a great amount of demand.”*
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 支持 回应 agent 替代论：可观测数据比 LLM 输入大 5-6 个数量级、需近实时，LLM 做会差且贵 (自动驾驶类比，$10,000/天)
And that is why there's been this race (07:40) over time and everybody is going after it, including the leading companies in AI whether it'd be the AI builders or the hyperscalers, so that's one thing. The other thing I will say is that the kind of problems we solve are somewhat different in nature from what these fundamental models – financial models are built for, and operate at very different levels of data volume and real-time requirements. So we have five or six orders of magnitude more data than what is typically put in an LLM. We need to operate in near real-time. So the analogy I would give to understand where we stand with the problem we solve versus what these models do is compared to self-driving cars, so I mean, today, it's working pretty well. You have self-driving cars everywhere.
- 反对 CEO 承认：通用 LLM 借 MCP 工具取数也可能做出类似的排障推理，Datadog 的差异要靠数据平面内的专用模型
And so, we do that as well. So, that's what we do today. If you squint, you could say, hey, some of that you could also do from a (39:07) and you can then – or MCP tools to acquire data and maybe they can come to the similar reasoning and come to a similar conclusion. The next level of what we're building though is about having that intelligence directly inside the data plane, with all of our real time data and to drive the automation itself. So, models are specialized that can run on very large volumes of data at a very attractive cost and with very low latency. So, we can fully automate the operations there.
- 口径不齐 加速“数学上”主要来自现有产品；前 20 家 AI 公司中 14 家用的是和其他客户一样的传统产品，新 AI 产品尚未成为驱动
There's several things that you've been putting out that you, I think, have been very hopeful that could become some of those next S-curves, whether or not it's security, whether or not it's AI monitoring, the GPUs, monitoring that you guys have been putting out. When you look at the acceleration that's going on right now, how much can you kind of think those new things for this versus kind of your existing footprint? What's interesting is most of the acceleration like mathematically, it comes from the existing stuff. And when you think of like we have the – I think we said like 14 of the top 20 AI companies using us, including the very largest of those, like, they use the same traditional products as the rest of our customer base, so by and large. So that gives you a sense of the shape of the market. When you look at the observability category, so there's this Gartner number that we put every year.
- 口径不齐 LLM 观测在开发侧 (LangChain、Braintrust 等) 是否要争、争多远，管理层自己说不清
So, I remember people being very opinionated about the (25:37) they use and things like that. And whereas in observability, you have only one production application, so like everything has to plug into one environment, to one thing in the end, and that's platformy in nature. So, what's not completely clear to us today is how much of that is the purely developer, purely local, non-platformy, lots of diversity of usage in any company, some homegrown, some open source, et cetera, and what parties and how far basically back there we have to go to own fully the production side of things. So, when you kind of look at that opportunity going forward, I think perhaps one of the challenges is it's very dynamic, like how we want to do all of this stuff is somewhat of a work in progress. How do you think about your own investment in kind of being part of the frontier there? So, you're in the right place at the right time when things start to solidify and become a little bit more common across organizations. Yeah. So I mean, we spend a ton of time, like, watching our customers and what is it they do and how they work, and so we have the luxury of catering to every single layer of this time into customers. So, most of our business is with large enterprises.
- 反对 企业 AI 落地中观测也会受益，但观测预算没有安全多
预计Search受益较多，很多企业AI落地的切入点是RAG。包括在中国与腾讯、阿里云的增长也主要来自向量库。安全也会受益于，因为黑客已经用AI进行攻击，安全分析必须用AI，否则会有落差。特别是考虑到Anthropic之前发布的Claude Security，安全厂商市值都掉了。Elastic的安全解决方案与Anthropic、Datadog等都有合作，在ES的解决方案都是用这几家的模型，用户可以自主选择这些模型，然后在ES的场景中做数据分析。用的效果最好的还是Anthropic，token消耗最多，做安全分析的时候效果非常突出。分析告警相关数据，再由大模型生成分析报告，告知入侵情况、阶段、态势分析、修复方案和合规方案。这些都可以由模型生成，所以现在企业在做安全分析的时候是一定会把AI加上去的。观测性也可以受益，但用户在观测上的预算没有安全那块多。
来源背景：未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。 - 反对 DDOG 安全业务没有 CISO 关系，作者给的份额只有 4%
**关于 Cloud Security 给的低份额**:DDOG 在安全上的短板是没有 CISO 关系——Bits AI Security Agent 是战略投入,但 CRWD/PANW 直接面对 CISO 是 DDOG 永远绕不过的销售短板。4% 份额是合理的"够大但难突破"。
来源背景：Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。 - 支持 LLM Observability 增长显著、开始有收入，但仍早期，未量化
Think of it as you have databases, you have code, lines of code. So that is where you have LLMs in a production model application and you're using signals to understand, is that affecting the performance of the application? And I think we've said that, that is germane to having LLMs in production, and we have – we've had significant growth there, still early days, so we're getting revenues from that. Then we have GPU, which is essentially like an infrastructure product, but instead of CPU, GPU, where you're seeing how the application might – or the model may interact with GPUs in delivering. And that's like similar to our other products where you have to see how the servers or the GPUs are doing.
- 支持 Bits SRE 2,000 客户使用，10 万+ 次调查
And so for different end markets, so SRE would be the systems reliability engineers, et cetera. So that product is out there. There are 2,000 customers using that. I think we have 100,000 or more investigations. And then we also have products in – for development and for security that we're rolling out in this Bits suite. We also have, as I mentioned, the ability to connect to the code generation through our MCP server when – that will enable that information to get into Datadog and monitor that.
- 支持 新增 token 成本/agent 使用监控 (由云成本管理延伸)
And then we also have products in – for development and for security that we're rolling out in this Bits suite. We also have, as I mentioned, the ability to connect to the code generation through our MCP server when – that will enable that information to get into Datadog and monitor that. And we also have a bunch of other products, (00:31:16), et cetera, that is being able to monitor on sort of the cost and management side how much you're spending on tokens, who's using it. We have cloud cost monitoring and now we're extending that into agentic monitoring. So that's a – that was a long talk, apologize for that. There's a lot of names but those are some of the product lines that are being put out to market...
- 支持 安全 >$100M，Cloud SIEM 借 >$1B 日志业务进展最快
It has how the cloud's working, it's called posture management, and set up to secure digital applications; two is cloud SIEM, how are you using logs and other information to investigate what's happening; and three is code security, how are you engineering security into the code. And so for – with – so not on-premise, not legacy, but for cloud applications, we've built this suite of the three products. I would say the one that we've probably made the most progress on is cloud SIEM. Some of the reasons it's off of our logs business and have an over $1 billion log business. We have a great, largest observability logs, and the end market has dynamics where we're been an innovator in it. So we've had two major strategies.
- 反对 安全分析领域 Palo Alto 在争夺本应流向 Splunk 或 Datadog 安全业务的预算
回 答 者：DataDog 的核心是可观测性和 AI 负载监控，而 Palo Alto 的定位更偏向于一个平台。在大多数大型客户中，两者是共存的。我认为，双方在安全分析领域的重叠正在增加。DataDog 提供安全监控和威胁检测，而 Palo Alto 则拥有 Cortex 和 XCM 产品。因此，Palo Alto 正在试图争取本应流向 Splunk 或 DataDog 安全业务的预算。
来源背景：未具名专家，Accenture (埃森哲) 前任人员 (身份由同场 MeritForum 06-04 标题'埃森哲离职专家'推得；本篇无自我介绍)。以'my portfolio'口吻谈客户，推测为咨询/集成侧的安全业务负责人。视角局限：咨询商视角，只见自己客户组合；主题是 Palo Alto/CyberArk 安全，DDOG 为顺带；Chronosphere 数字为举例性量级。 - 中性 AI 可观测性是新兴领域，玩家包括 Chronosphere、Datadog、Dynatrace、LangFuse
再次是 AI 可观测性。这是一个迅速崛起的新兴领域，主要参与者包括 Chronosphere、DataDog、Dynatrace 和 LangFuse。这与传统的网络安全有所不同。
来源背景：未具名专家，Accenture (埃森哲) 前任人员 (身份由同场 MeritForum 06-04 标题'埃森哲离职专家'推得；本篇无自我介绍)。以'my portfolio'口吻谈客户，推测为咨询/集成侧的安全业务负责人。视角局限：咨询商视角，只见自己客户组合；主题是 Palo Alto/CyberArk 安全，DDOG 为顺带；Chronosphere 数字为举例性量级。 - 反对 产品都在 API/CLI 化、面向 agent，企业不需要 UI 上有很多功能
但在近期 AI 重心的背景下，其实并不是明显转向企业，因为所有产品现在有一个趋势，是面向 AI 开发的产品，既不是面向企业，也不是面向开发者生态，而是面向 AI 和 agent。面向 agent 有一个非常重要的趋势，就是所有产品功能都必须 API 化、CLI化，文档要写清楚每个参数怎么用，让 agent 能学习到，并放在互联网上，进入大模型的 pretrain 材料里，让大模型天生知道如何调用你的平台、组装 API，完成企业级需求。你要有足够丰富的 API，覆盖足够多场景。现在企业客户其实不需要软件定制化，不需要企业为他定制产品，所有定制化都由 AI 完成，不需要 UI 上有很多功能，agent 接入后可以直接定制化想要的内容。现在比较火的是 MCP apps，可以作为 plugin 或 extension 嵌入 cloud，用户以对话方式实现功能，比如画图表、数据分析，直接在 agent 上呈现，完成定制化功能，不需要找企业开发，直接和 agent 说就能实现。如果找企业开发，可能要等好几个迭代才能上线。所以现在的重点一定是面向 agent 做开发。
来源背景：未具名；以 Elastic (ES) 内部视角回答 (清楚 ES 组织架构调整、API 文档策略、Jina 整合)，推测为 Elastic 现任或前任员工，带明显中国区视角 (国内定价、阿里/腾讯 ES 服务)。局限：竞争对手视角；对 Datadog 的判断是间接的。 - 反对 大型受监管机构不会用 Datadog 或其他观测平台替换 Splunk 做安全 SIEM，只有中小机构可能整合
Expert: I don't think so. Here, perhaps maybe smaller or medium-sized organisation might look into consolidation. If you look at Datadog and Dynatrace, I was talking about these two, they came in the market as an observability platform, logging and monitoring going a little bit beyond. When Splunk came in, they were actually purely a security play. This is a security SIEM solution. Let's say, if you want to replace Splunk with my security SIEM, I would not go replacing with the Datadog or with any other observability. They play a total different role. That's why you had to see whether in your organisation, how actually you're using these solutions within your organisation. I wouldn't say it's displacing. Here, larger organisation, regulation and all that, it's required that you have that multiple touch point. You have that complete defence in depth then observability, two different platforms. There's overlap, but it's just additional functionality.
来源背景：匿名专家 ([Expert])。Third Bridge 主持人介绍为 'former Head of CCoE [Cloud Centre of Enablement] and Platform Engineering at PNC Financial Services'；但专家自述 '我是' PNC Cloud CoE 和平台工程负责人、在该行两年 。负责容器、中间件、数据平台及平台安全 (本地和云)；上一家机构曾把 Splunk 整合迁移到 Google Chronicle 。视角局限：受严格监管的大型银行买方，偏 '最佳单品 + 多厂商冗余' 策略，不代表一般企业；角色偏平台/安全，不直接管观测账单细节；对 Datadog 的评价较笼统、无数字。 - 反对 公司在试把 Datadog 的指标喂给大模型/agent 做根因分析，Datadog 只是数据源之一
按照传统运维模式，根因定位其实比较慢，需要一层层排查，这和运维的专业能力有关，最终定位问题可能需要较长时间。如果遇到重大事故，客户的容忍时间可能只有几分钟，长时间未解决会对公司形象等方面造成很大影响。此时AI Agent的作用是什么？公司现在也在尝试，将CMDB和各种指标，比如Datadog里的指标，作为数据源输入到大模型中，一旦发生问题就抛给AI Agent。但AI Agent为什么现在不敢随便用？因为生产环境下，AI Agent权限过高可能带来风险，出问题还是运维人员担责，不是AI担责。所以目前生产环境下还在观察，不敢大胆尝试，但在少量非核心系统上可以用AI Agent，能在几分钟内定位并解决问题，从而降低公司的损失。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 中性 AI agent 在生产环境权限过高、责任归属不清，目前只用在非核心系统
按照传统运维模式，根因定位其实比较慢，需要一层层排查，这和运维的专业能力有关，最终定位问题可能需要较长时间。如果遇到重大事故，客户的容忍时间可能只有几分钟，长时间未解决会对公司形象等方面造成很大影响。此时AI Agent的作用是什么？公司现在也在尝试，将CMDB和各种指标，比如Datadog里的指标，作为数据源输入到大模型中，一旦发生问题就抛给AI Agent。但AI Agent为什么现在不敢随便用？因为生产环境下，AI Agent权限过高可能带来风险，出问题还是运维人员担责，不是AI担责。所以目前生产环境下还在观察，不敢大胆尝试，但在少量非核心系统上可以用AI Agent，能在几分钟内定位并解决问题，从而降低公司的损失。
来源背景：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 - 支持 Wedbush：Bits AI、Agent Console、Journey Monitoring、AI Guard 客户采用强
Following DASH 2026, Wedbush came away more constructive on DDOG, highlighting 100+ product launches and a growing focus on autonomous operations across development, security, and observability. The firm believes Datadog is evolving beyond monitoring into a full-stack platform that helps enterprises manage increasingly complex AI-driven environments, with new products spanning AI observability, security, workflow automation, remediation, and developer productivity. Wedbush noted strong customer adoption of AI-focused offerings such as Bits AI, Agent Console, Journey Monitoring, and AI Guard, arguing these products expand Datadog’s role across the software lifecycle while creating additional share-gain opportunities.
来源背景：TMT Breakout (TMTB) 日报，转述 Wedbush (DASH 2026 后) 与 BofA (基础设施软件 'Fab Five') 报告摘要。二手转述，看不到底稿。 - 支持 Bits AI、AI Guard、数据库/网络监控、查询优化兴趣特别强
Following DASH 2026 and discussions with customers, BNP Paribas came away more constructive on DDOG’s growth durability, noting customer enthusiasm was stronger than expected across observability, security, and AI-related offerings. The firm highlighted Bits AI, AI Guard, Database Monitoring, Network Monitoring, and Query Optimization as products generating particularly strong interest, with several customers viewing AI as a catalyst for increased observability spend rather than a substitute for it. Customer conversations also suggested Datadog is benefiting from growing demand for AI workload monitoring, application tracing, security safeguards, and full-stack visibility across increasingly complex cloud environments.
来源背景：TMT Breakout (TMTB) 日报，转述 BNP Paribas DASH 2026 后客户调研摘要。二手转述，看不到底稿。 - 反对 AI 成本监控是新类别，但客户也会更严苛审视可观测性本身的成本
And they may want to know not only whether an agent is working, but also whether it is delivering sufficient business value compared to tokens and infra consumer. And that creates both an opportunity and a challenge for data doc. The opportunity is that the AI cost monitoring becomes new observability category. The challenge is that customers will scrutinize the observability platforms, you know, on cost more aggressively. Now if a company is spending heavily on AI tokens, and then pay a significant premium to monitor them, then finance teams will start questioning the total economics. So looking forward, I think the winner will be the platform that can help customer reduce the overall AI operating cost rather than simply measure them. If data dog, BIA or similar tools can identify that waste, optimize token usage, improve the model selection, and reduce the infra spending, I think customers will view the observability spend as part of the saving mechanism rather than additional cost layer.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 更大风险是 AI 推理被商品化，届时差异化转向谁拥有运营数据，Bits AI 本身不是护城河
I think the bigger risk is not that Anthropic can replace Datadog entirely. I think... The bigger risk is that the AI reasoning becomes commoditized. So if Anthropic, OpenAI, Google and others, they start offering highly capable agents, then the differentiator pivots from the AI model itself to who will own your operational data, workflows, integrations and customer relationship. And if that happens, that Datadog's competitive advantage will come down, you know, come from its install base, deep telemetry coverage, workflow integration and operational context, rather than from the bits AI itself. So customers may ultimately care less about which AI model is generating the recommendation and more about which platform has the most complete view of their environment. So, my personal perspective is that Anthropic could become a competitor in AI driven troubleshooting, security analysis and automated remediation.
来源背景：未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 - 反对 纪要中的举例：在 AI 上花 50 万美元再为监控付显著溢价，财务会施压 (音频原话没有具体金额)
机会在于 AI 成本监控成为一个新的可观测性类别。挑战在于客户会更严格地审查可观测性平台本身的成本。如果一家公司在 AI 上花了五十万美元，然后又为监控它支付显著溢价，财务团队会对整体经济性进行施压。
来源背景：标题写明 '北美软件渠道商离职专家' ，作者 Lantian；内容与 Merit Datadog 产品进展调研 I (2026-06-16 音频) 逐段对应，判断为同一专家同一场的英文转写再翻译版。局限同音频 I：只看到渠道客户，无样本量，回答偏框架化。 - 反对 Datadog 只覆盖 Agent 治理的一部分，没有端到端工作流和风险面板；ServiceNow AI Control Tower 在争这一层
What their AI control tower is doing is actually managing all of those processes along the way. For a Datadog, to be able to do that, they would have to have complete access to every single system. In my opinion, they do pieces of it, but they don't do it end-to-end, and they obviously don't provide a dashboard to be able to have someone understand how everything is performing along AI agents, if there's any risk, if there's any problem with anything along the way, somebody doing something they shouldn't. It also tracks it from a life cycle management perspective. From intake into deployment, that tracks AI along the way.
来源背景：ServiceNow 前 Senior Executive, Alliances & Channels (标题明示)，有约 25 年软件行业经验；访谈实际日期 2026-04-17 (Client-led)。视角局限：ServiceNow 一方，自然强调 ServiceNow 的价值；不熟悉 Datadog 产品细节。 - 中性 Bits AI 有客户兴趣但还看不到真实数字
Uh like log traces, etc. And then it sort of uh it proposes fixes to to the code based on that. So so uh it closes the gap between the signal and then how you need to change the code. So so that is like, that is like a a very relevant use case. Um and and we see an interest here, right? But but it's still too early to see like real the numbers. I think maybe by Sorry, there's some background noise there. 对,就是有一点背景的杂音。
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 反对 Bits AI 主打的自主修复技术上可行，但合规/护栏/人在回路角度短期不会落地
Then of course, uh, the friction is between, you know, the keynote speeches and production. So for example, they they they they pitch this as sort of autonomous remediation, which is basically like, you know, you would go fully autonomous from the telemetry data to the code fixes. And that I feel is, you know, a bit too uh uh uh too uh uh, technically it works, but but from a compliance and and guard rails and and human in the loop perspective, it's not going to work, uh or it's not going to fly uh anytime soon. Uh, but Bits AI is absolutely um, absolutely something we feel will be a growth driver going forward. That's what I can say.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 中性 Agent Console 现在不是大增长驱动，但让 Datadog 卡位 agent 层的中立监控
And there, uh monitoring that output as well is getting increasingly important. So so basically that is the story of of the agent console that that it will allow a centralized monitoring of uh you know, the actions that AI agents are taking. So again, it's not a massive growth driver just right now, but it positions data dog uh very sort of cleanly to be a monitoring layer across the agentic layer if you will. and that I think is great. And and also since it would have been different if the if the intelligence layer was because then then that platform could build the observability functionality themselves.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 支持 token 花费过高不会减缓 AI 采用，反而推动 token 监控和可观测性
And second question is what kind of impact do you see this phenomenon on the AI infra ecosystem or the software companies? Yeah, and you're absolutely right. So so we've seen that and and basically I don't think it will slow AI adoption in in like uh in in the broad sense. Uh but I do think this is a case for uh you know, token monitoring and and and this is partly coming from the model platforms themselves, but partly also this is a driver for uh you know, observability uh in a general sense because basically what I foresee is that you will have a more disciplined process in like steering your uh your applications and your developers to the cheapest models, you will have a uh increased cost control and you will have uh basically a more uh disciplined approach to token usage, right? And and and uh observability is is a way to gain visibility into that. So if anything, that plays into the the Data Dog story.
来源背景：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 - 反对 Bits AI 兴趣大、POC 多，但实际采用还不多；成本 (硬件涨价挤 IT 预算) 和数据合规评估是阻碍
It's a lot about helping customers formulate a framework, um whether it is on root causes or analyzing root causes, whether it is about detecting issues and fixing bugs, you know, within within the environment. So there is definitely an automation um I would say value add. Um we have a huge amount of interest. Not a lot of up takes as of yet. Um I think we have a number of challenges at the moment. Cost is one of them from from a customer IT budget perspective, because of this year we've been challenged really challenged by increases of hardware generally which is affecting the entire IT budgets. Um, and customers are really evaluating AIs across um all infrastructure, um as it pertains to data essentially. Um so, yes, we are taking a lot of, we are seeing a lot of increases in AI adoption, but customers are also really going through and analyzing to make sure data is really protected and they are compliance.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 反对 Datadog AI 产品尚早、对企业不够成熟，渠道 GTM 还没准备好；明年才可能起飞
Uh particularly when we look at when we look at Data Dog, I think this will be uh more of a takeoff for Data Dog in terms of AI. I think customers um, and in a way, I think Data Dog AI is really um it's at the early stages, I would say. I don't think it's completely mature um for enterprises to adopt. Uh there's a lot of interest, but I think there's still things to improve, go to market strategies in the channel, for example, that's not quite ready yet. So there's a lot to to work in the background, I think. 嗯,OK,听上来他觉得今年就是传统的软件肯定是受到了压力的,然后呢,他说硬件还有整个AI的采纳这一块,其实都是受到了这个成本的制约,但是 呃就是客户他也没有去停止这个AI的部署,因为他们也不想落后嘛,而且他们也不确定明年这个价格是否会回落。
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 中性 agent 场景定价未明，可能不按 agent 计价，而是按用量或整体环境打包；agent 带来至少 20% 增量机会
Um we haven't really seen um for clarity in terms of the pricing model. That will change. Um is it going to be per agent? Probably not. Um this is probably going to be a package um based on usage maybe, uh or based on the entire environment. Um the the challenge here is, you know, the more agents you adopt the more expensive it can be. So, Data Dog is really um competing in terms of commercials. and that's where it's winning at the moment, you know, against Dynatrace or some of the other competitors. It's winning because of its commercial uh appetite. So, yes, this is a this is a good question. It will definitely expand on its opportunity by at least 20%. So the growth is at least 20% from the agents. We know that.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 中性 未来定价可能是订阅 + agent 高级附加组件的混合包
So I think Datadog and all the other competitors will completely absolutely benefit because of AI. Uh, absolutely. Now, in terms of the package, we will probably start to see a hybrid packaging model whereby, you have a, maybe a hybrid model whereby you have some sort of subscription and then you have some premium add-on, add-ons on the agents in terms of pricing. That's what's going to drive the cost. 20, we're hoping 20 to 30% growth.
来源背景：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 - 反对 Bits AI 兴趣大但采用率不高，成本是障碍之一
A:兴趣非常大——有些兴趣已经转化为使用，但目前有很多概念验证。它主要与自动化和帮助客户建立框架有关。无论是用于识别根本原因、分析记录，还是检测问题并修复环境中的缺陷，确实存在自动化的角度。我们有大量兴趣，但采用率还不高。我们面临一些挑战； 成本是其中之一。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 反对 Datadog AI 还处于早期，没到企业级采用的成熟度
所以我认为明年会看到更多的 AI 采用。就 Datadog 而言，我认为这将更像是起飞。Datadog AI 还处于早期阶段。它尚未完全成熟到企业级采用的程度。兴趣很大，但还有需要改进的地方——例如在渠道的上市策略等——所以它还不完全准备好。后台还有大量工作要做。
来源背景：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 - 支持 Bits AI (AI SRE，自动修复/根因) GA 仅 6–7 个月，已带来部分收入加速
So uh one of them is that uh the adoption of AI case, AI use cases continue to accelerate. And uh some of the Datadog products which have been out in general availability for some time now have accelerated in terms of adoption. So for example, the Bits AI product, uh which is the artificial intelligence site reliability engineer, which does automatic remediation, automatic root cause analysis, that has only been in general availability for let's say the last six, seven months, it's a relatively recent launch. So that product has picked up a lot of uh uh velocity, which is driving some uh revenue acceleration, right? Another uh driver of the growth uh is that Datadog has had security products for about the last uh couple of years or so. So, uh increasingly now we're starting to see big security deals in addition to observability.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 安全产品开始出现大单，GTM 飞轮启动；安全增速在上升
So that product has picked up a lot of uh uh velocity, which is driving some uh revenue acceleration, right? Another uh driver of the growth uh is that Datadog has had security products for about the last uh couple of years or so. So, uh increasingly now we're starting to see big security deals in addition to observability. They always had observability products. Now they're adding cloud security products, code security, API security, etc uh to to their roster, right? And increasingly now they're finally seeing the GTM flywheel finally uh kick off for security as well. So these two, these two are the biggest uh uh biggest drivers I would say of the mild revenue acceleration, although uh the the year-on-year growth in Q1 and Q2 remains roughly constant, the mild acceleration is due to uh these these two reasons. 嗯哼, okay。
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 Bits AI 收入占比低到中个位数，环比增长 50–60%
The uh company's revenues are right now massive, right? Like they had their first uh quarter with a billion dollars of revenue. So in that uh broader scheme of things Bits AI would contribute in like the low single digits, low to medium uh single digits. So it's it's a driver of acceleration, but uh still a lot of adoption that uh can happen on Bits AI, very early days. But it's growing quickly, it's growing uh somewhere between 50, 60% uh quarter on quarter. So that's uh an encouraging sign and the growth honestly can even accelerate as they go full guns blazing uh now that the product also is getting much more uh uh uh improved feedback, right? But there's a lot of demand in the uh market for a tool like this and uh you know, a lot of our uh uh new business calls are uh revolving around the capabilities of Bits AI and I'm sure that uh you know, you'll see even more contribution from the uh product in the in the next few quarters. Mm, okay. 他说那个,整个反馈还是比较积极的,因为,呃,确实是这样的,就是Bits AI的话,它其实也是基于底层的模型嘛,所以模型的质量其实会影响到Bits AI的表现,然后他说之前的话,其实因为模型质量的问题,所以客户可能对这个Bits AI在做根因分析啊,在做问题的发现和纠正等等的准确度方面其实是有一点点小小的complain的,但是呢,随着底层模型质量的不断提升,他说包括我们看到Anthropic最近发的这个Fable的模型,就是模型质量显著提升之后,Bits AI它的发挥也是质量有显著的提升,所以相对来说客户现在的反馈是很积极,呃,然后他说从这个revenue的贡献来看的话,虽然现在有一些Datadog的大客户在部署Bits AI,但是它的这个revenue的贡献还是比较小的,因为现在还是非常早期的阶段,呃,然后他说因为Datadog现在的营收也还不错嘛,所以在整个公司的营收方面,Bits AI可能贡献的也就在5%左右,但是他觉得这个,呃,这个Bits AI产品的潜力是非常大的,因为如果看一下季度环比的增速的话,是有50%到60%的这个增速。 他觉得未来是一定会有一个这个部署的加速的, 所以在未来的话会看到它对整个公司的 revenue 有一个更显著的 增长。 嗯,明白。
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 中性 Bits AI 底层用 OpenAI、Anthropic 等外部模型混合，内部在讨论自研；时间序列有自研模型 Toto
还是说我们有自研的一些东西, 对。 So the underlying model for Bits AI, is it uh Anthropic's model or Datadog is using some of the internal models? So, there's been a lot of conversation on considering developing a fundamental model that is proprietary, but currently they're using a mix of uh different models from providers like OpenAI, Anthropic, etc. Uh it may change to a proprietary model at some point of time. Like for example, for the uh for the time series modeling, they have a proprietary model called Toto, uh which is doing some of the forecasting for the metrics that you might be monitoring for uh with with Datadog. But uh currently the models in the backend for Bits AI are uh from the large model providers. It's not a proprietary model.
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 安全收入占比小但增速在上升，未见 AI 安全公告带来实际收入冲击
And also security is actually a relatively nascent uh part of the data dog offering. So, the core revenue still remains observability. Security is a relatively uh uh smaller part of the uh revenue, but there we see the growth rate actually increasing versus decreasing even though the uh revenue basis also increasing right? So, at least in terms of numbers, we're not seeing anything there and I think it's more of a uh sort of like, you know, move for the public markets where uh SaaS stocks start falling when they announce that the product has uh come in versus there being any actual revenue impact on uh data dog, right? So, that would be my opinion. 哦，OK。 他说是那个Anthropic是去年成为DataDog的客户的，然后几个月前的话，他们把这个合同做了一个延期。 呃，然后也看到就是Anthropic对DataDog这块需求的一个增长，不管是产品使用的数量上，还是它的这个产品 用的这个广度和深度都是有一个增长的。 然后它这个营收贡献的话,他说跟OpenAI差不多,现在就是5%左右,所以它这个数字并不是很大,但是他们觉得Anthropic在DataDog这一块的花费会继续的增加。
来源背景：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 - 支持 BofA：Adaptive ML 增加强化学习能力，支持 Bits SRE、Bits Code、Bits Security Analyst
BofA views Datadog’s acquisition of Adaptive ML as a strategic addition that accelerates its AI roadmap by adding reinforcement learning capabilities to support proprietary AI agents and observability products. The firm believes the deal enhances Datadog’s ability to build differentiated AI offerings such as Bits SRE, Bits Code, and Bits Security Analyst, making it incrementally more confident the company can meet or exceed near-term targets while strengthening its long-term growth trajectory. No financial terms were disclosed.
来源背景：TMT Breakout (TMTB) 日报，转述 BofA (DDOG) 与 BTIG/BNP (PANW，涉及 Chronosphere) 报告摘要。二手转述，看不到底稿。 - 支持 用量制相对席位制的优势：agent 减少席位不影响 Datadog
So if customers send more data or if they use more, basically, that's built into the model. We're not seat-based. Like a lot of the traditional SaaS companies are seat-based and they're challenged in that maybe they get less seats because they have more non-human agents using their systems and maybe they need to bundle more functionality within those seats. We're usage-based. We're still bound to the same dynamics, right? At the end of the day... Right. ...we have to show value, and there's only two reasons people buy software. They buy software to make more money or to save money.
- 支持 Adaptive ML：停售原产品，技术用于自研专用模型，比前沿模型快 100 倍、便宜 10 倍 (管理层说法)
And the – Adaptive is – what the product does is it helps you train your own models based on evaluations that you build with maybe some frontier models or other models you can already use in your product, and it basically allows you, through reinforcement learning, to own the models and reach a different level of performance and cost. That's amazing. And that's something that – we're actually not continuing that product, like we're not selling... Right. ...that product anymore, but that's technology that we're using internally to build our own models, whether that's today models that replicate what you might do with some of the frontier models. So, basically thinking models but specialize on our use case that can be 100 times faster and 10 times cheaper, so that's definitely something we want. But in the future, we also have our eyes on building models that can be end-to-end world models for our problem space. So instead of a post-hoc investigation, say, for example, I've had an outage and I'm asking Claude or ChatGPT to explain what happened and they're going to call some tools and reason and give you something, instead, think of a model that takes your telemetry data as an input.
- 支持 愿景：端到端“世界模型”，直接以遥测为输入预测故障并给出处置，而非 LLM 事后调工具调查
So, basically thinking models but specialize on our use case that can be 100 times faster and 10 times cheaper, so that's definitely something we want. But in the future, we also have our eyes on building models that can be end-to-end world models for our problem space. So instead of a post-hoc investigation, say, for example, I've had an outage and I'm asking Claude or ChatGPT to explain what happened and they're going to call some tools and reason and give you something, instead, think of a model that takes your telemetry data as an input. So, say, you send your metrics, your logs, your network telemetry, your code, your traces and everything into the model, and then that comes in on one end, and on the other end, what you get is, oh, this is what's going to happen in the future and that's the problem and this is what you should do about it. Right.
- 中性 自愈软件比自动驾驶难；自主修复需远超 90% 的准确率，会花时间
Do you see us get to a place where everything is 100% or do you still see human intervention coming in? There's a couple of angles there. The first one is self-healing software is actually a lot harder than self-driving cars. And the reason for that is the roads pretty much all look the same. There's not a high level of variability, they don't change very fast.
- 支持 Datadog 客户有专门的安全日志 (SIEM) 预算，这是相对 Dynatrace 的优势
how we intend on hitting our 35% target. Datadog, it's a little bit more, I'd say. One aspect with Datadog versus Dynatrace is that there is, when we think about logging... again, a key theme here is logging, semi-structured, unstructured data... is that with Datadog, we also have the benefit, unlike Dynatrace, of having customers spend allocated for security logging. So what we consider like SIEM, for security incident event monitoring. So we're very confident on the Datadog number. In addition to the meat and potato type stuff, like infrastructure, cloud monitoring, APM, logging, we're also with Datadog definitely more consolidation opportunities around things where customers are looking to consolidate legacy implementations of things like X Matters or PagerDuty onto Datadog On-Call. And with it, it's not always as apparent with Dynatrace, but with Datadog, they have a very discrete SKU related to LLM observability, and we are starting to see uptick in customers explicitly acquiring capacity for LLM observability. In essence, it's the same thing from a Dynatrace standpoint, where the customers are requiring this capacity to basically enable LLM tracing and LLM logging within their telemetry data sets.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 Datadog 有独立的 LLM Observability SKU，开始看到客户明确购买其容量 (未量化)
how we intend on hitting our 35% target. Datadog, it's a little bit more, I'd say. One aspect with Datadog versus Dynatrace is that there is, when we think about logging... again, a key theme here is logging, semi-structured, unstructured data... is that with Datadog, we also have the benefit, unlike Dynatrace, of having customers spend allocated for security logging. So what we consider like SIEM, for security incident event monitoring. So we're very confident on the Datadog number. In addition to the meat and potato type stuff, like infrastructure, cloud monitoring, APM, logging, we're also with Datadog definitely more consolidation opportunities around things where customers are looking to consolidate legacy implementations of things like X Matters or PagerDuty onto Datadog On-Call. And with it, it's not always as apparent with Dynatrace, but with Datadog, they have a very discrete SKU related to LLM observability, and we are starting to see uptick in customers explicitly acquiring capacity for LLM observability. In essence, it's the same thing from a Dynatrace standpoint, where the customers are requiring this capacity to basically enable LLM tracing and LLM logging within their telemetry data sets.
来源背景：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 - 支持 客户常把 Monte Carlo 的指标抽出来放到 Datadog 统一查看；若 Datadog 能力相当，会标准化到 Datadog
**TL** [00:50:02]: I think it's a real risk. There's a lot of customers in the space. Again, this doesn't matter if they're using Monte Carlo or anybody else, they actually just pull all of the metrics out of Monte Carlo and surface them in Datadog because they really want a standardised way of actually running all of their observability, be it whether it's application or data. It can often sit with the same teams anyway. If Datadog is able to offer a solution that is as good, then I think a lot of companies would go the route of standardising on something like Datadog instead of having two platforms, especially if you're going to the effort of maintaining a connection between Monte Carlo and Datadog when you could have it all in Datadog. Certainly something that we're starting to see across the board.
来源背景：Terry Lucy (TL)。Third Bridge 页眉 'Executive, Go-to-market at CodeRabbit Inc'，介绍为 'former Strategic Account Executive at Monte Carlo Data' (明示)；自述在 Monte Carlo 多年，是伦敦办公室区域第二号员工，负责企业/战略客户 。视角局限：数据观测 (非 APM) 销售视角，对 Datadog 的看法来自竞争一侧；主题是 Monte Carlo，涉及 Datadog 的内容只有几段。 - 中性 AI 产品目前只占收入 3–7%，Bits AI 仍是个位数；预计 4Q 到 6–10%
* **AI products contribute only 3-7% of revenue today, and on an implementation basis Bits AI is still single digit**; he expects 6-10% by 4Q, with a material impact starting in 1Q27
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 LLM 可观测性/GPU 监控/Bits AI 兴趣高，但做了很多 POC，企业级落地还没出现
+ He expects LLM-related activity alone to add at least ~20% to his DDOG business, with the opportunity concentrated in GPU monitoring, metrics, logs and AI workload performance + **Plenty of POCs have been run, but genuine enterprise-level adoption has not yet materialized**; AI products typically take 3 months and sometimes 6 to move from evaluation to deployment, with compliance the bottleneck * For the channel, Bits AI’s primary value is stickiness — the degree of integration makes DDOG harder to unpick and supports the consolidation narrative
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 支持 LLM Observability 是独立 SKU，已看到客户明确购买
* **LLM Observability is a discrete SKU at DDOG, and he is already seeing customers explicitly buy that capacity**, which differs from DT, where the equivalent capability lacks such a clean selling unit * The most common practical AI monitoring requirement among his customers is **performance monitoring of MCP servers** (agent-to-agent communication), followed by monitoring of frontier model API calls and of coding/co-worker assistant interactions
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 反对 客户 E 的 LLM 可观测性已选 BrainTrust：便宜约 20%、保留期更长
+ **If detection and remediation capability keep improving, he expects SRE headcount could come down 30-40% in phases within about a year**; today the agent can only detect and has no remediation capability * Purchase path: even if they buy, they would start with a very small user group, and reaching a broader population would take 1-1.5 years — he specifically noted headcount itself will shrink by then, so it may no longer be the current 500-person base * **DDOG has already lost the LLM observability workload here** — the customer chose BrainTrust because it is roughly 20% cheaper and includes 5GB of data, 50,000 scores and 30-day retention, against DDOG log retention of only 15 days with extensions costing extra (metrics are 90 days)
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 中性 Bits AI 改为 AI credits 定价模式 (正在推出)，因覆盖面从告警调查扩到 chat、监控、代码、测试等
There's all sorts of different use cases that we built into it that broadened the surface of contact. And we see a lot of adoption across all of those different areas. We also are changing the way we package it. So we have a new model with AI credits that we're rolling out, just because the surface of contact is so much wider now than the specific feature. So there's quite a bit that is going on there. The explosion of activity that I mentioned earlier about other parts of our – or other AI surfaces is happening also in Bits AI.
- 中性 LLM/AI 监控：一年前量很小，近两季快速加速；但品类 '非常非常早'，产品、用量和打包都可能大变
Yeah. Yeah. And look, the volume, like the – there used to be very little volume a year ago, it started growing quite a bit into the second half of last year. And now it's been very rapidly accelerating over the past couple of quarters. So we've seen like an explosion basically of the volume we're getting there. And we get more usage from different kinds of companies, so we definitely see that. We see it also across traditional companies and some more recent AI natives. So we see a little bit of both. I would say for that category, it's still super, super early, like we expect the products to change quite a bit. We expect the usage and maybe also the packaging to change over time quite a bit. Got it. Thank you.
- 支持 GPU monitoring 在 neo-lab 和 AI-first 客户中用量不少；agent 监控量爆发 (无数字)
There's opportunities at the very top end when you measure what the agents are doing and whether you're getting the right outcomes or whether you're getting the right alignment, and there's opportunities at every layer in between, just looking at the LLM itself, just looking at the tool calls and the applications that are being called by the agents, like everything is an opportunity in there. We see growing adoption from the products we already have there. We mentioned our GPU monitoring product, is actually getting quite a bit of usage in a number of neo-labs and very AI-first types of customers. We're also seeing an explosion of volume in our agent monitoring product. And so we're well positioned there, but we think this market is going to change quite a bit and the preoccupations of customers will also change over time. So for example, last year, our customers were mostly trying to validate correctness and validate that they were getting some form of outcome that it could then scale up.
- 中性 GPU 监控、agent 可观测性、Bits AI、MCP、AI 安全有早期牵引，但变现仍在早期
Barclays says CEO/CFO meetings reinforced that DDOG’s acceleration is broader than AI-native customers, with non-AI growth improving from roughly 18% a year ago to the high-20% range as AI drives cloud migration, infrastructure modernization and new application development. Barclays notes DDOG intentionally de-risked 3Q guidance by assuming its largest AI customer contributes only at contractual minimums, while broader usage and pipeline trends remain strong. Barclays also says GPU monitoring, agent observability, Bits AI, MCP and AI security are gaining early traction, but remain in the early innings of monetization and represent a multi-year growth opportunity.
来源背景：TMT Breakout (TMTB) 日报，转述 Barclays 与 DDOG CEO/CFO 财报后 fireside 会面的报告摘要。二手转述，看不到底稿。 - 中性 Datadog for AI 产品按数据量或调查次数收费，GA 时公布价格
So, if they're training models and they're not putting in production and it's in-house or they're using it for their marketing collateral, internal, it doesn't tend to be our market, but more and more of it is being our market. So, we're monetizing it through pricing. We publish as we put into GA and it generally is on – like everything else, it generally is on the amount of data consumed or the amount of investigations and things like that. Still early days but really good growth signs. That's Datadog for AI.
- 口径不齐 Bits 定价从按调查扩到按 token；被问毛利平衡时 CFO 未正面回答
There's two things going on. One is, you're right, we have sort of – and this is very typical of Datadog. We play around with pricing, see where we think we can add value. And so that is being launched. And early signs are a lot of good reception and use. We also are expanding what we're doing with Bits. So, we started out with these investigations, right, and now we're also doing it in on to the left with development and security.
- 支持 收费按工作负载，不在乎是 agent 还是人在看，界面价值下降不影响计费
We're early on, definitely the way it's going to change where it's going to wind up, I don't know. But the good news is it probably doesn't matter. Meaning we're not a seed model. We are basically monetizing based on the workloads that go through and it really doesn't matter if those workloads are created or looked at by agents versus not. You're going to have to – we believe you're going to have to observe and secure this and probably all this is going to accelerate the pace of change, which means you're going to have to move from legacy. Legacy still is the biggest part of the market because it was there for 50 years.
- 支持 Mizuho：MCP server 用量强，Bits AI 扩展到 agentic 用例
Mizuho reiterates Outperform and a $300 PT after meetings with Datadog management, saying AI demand is broadening across the business while observability, security and newer products continue to scale. Mizuho highlights strong MCP server usage, expansion of Bits AI into agentic use cases and healthy new-customer contribution, while management sees AI as a secular tailwind rather than a narrow AI-native customer story. GTM remains strong, though security is still more of a work in progress, particularly around SIEM and Code Security.
来源背景：TMT Breakout (TMTB) 日报，转述 Mizuho 与 DDOG 管理层 NDR 后报告摘要。二手转述，看不到底稿。 - 反对 Mizuho：安全仍在打磨，尤其 SIEM 和 Code Security
Mizuho reiterates Outperform and a $300 PT after meetings with Datadog management, saying AI demand is broadening across the business while observability, security and newer products continue to scale. Mizuho highlights strong MCP server usage, expansion of Bits AI into agentic use cases and healthy new-customer contribution, while management sees AI as a secular tailwind rather than a narrow AI-native customer story. GTM remains strong, though security is still more of a work in progress, particularly around SIEM and Code Security.
来源背景：TMT Breakout (TMTB) 日报，转述 Mizuho 与 DDOG 管理层 NDR 后报告摘要。二手转述，看不到底稿。 - 反对 在 LLM trace 采集/埋点层面 Datadog 没有护城河，壁垒只在平台层的关联和内嵌 AI 调查
**EE** [00:29:21]: I don't know if you're referring to OpenTelemetry or something that's more specific to LLM observability, but Datadog has definitely seen that or there was a lot of chatter about threat from the OpenTelemetry, for example. This is probably similar. I think it took some time for sure or maybe it took them time until they realised, "Okay, this might cost us customers," or customers are really asking for it. There was a big fintech company that was kind of a condition. You have to support OpenTelemetry in this way or that way. Datadog really had to allocate a lot of resources and even change its mindset. Initially, it was like we'll support OpenTelemetry, but we're not going to talk about it. If someone wants it, then fine. By the time I left, it was also about, "Okay, how do we need to rethink our pricing for different types of deployments, if it's OpenTelemetry or something like that," because after all, it's not a one size fits all. They are looking at it. Just going back to the proprietary agents, I think, again, they're incorporating some open source, OpenTelemetry support into their proprietary agents. They're trying to give the flexibility definitely for customers to choose, but there are trade-offs. I think the defensibility, like you say, it's not necessarily there in the instrumentation and the ability to collect LLM traces, etc.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 中性 LLM 观测创业公司很多，对 Datadog 不致命，但客户可能先从它们起步
**EE** [00:26:19]: I've led the LLM observability product when I was at Datadog. I've seen some of these start-ups back then already, and I'm sure there are more today. If I'm looking at LLM observability specifically, yes, I think there's many start-ups. It could be errors, it could be reaching to sensitive data, it could be high costs. It could be many things. I do think there is a lot of companies already doing that. I don't think it's existential to Datadog. I think, if anything, companies might start with that. Again, all these start-ups usually, they target developers, from the beginning of the development cycle. It's easy. Many of them are open source and then try to upsell maybe to a paid solution. Developers find it probably easy to get started. Some of these start-ups also provide the framework to building the LLM applications or the agents. It's even easier to plug themselves into the monitoring of those applications, at least in the early stages.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 口径不齐 Agent 时代观测需求确定，但单一告警功能可能被更大的控制平面整合
**第三类是身份、安全和可观测性。** Agent 能够调用更多工具以后，权限范围和错误半径同步扩大。Okta 等身份平台、Palo Alto Networks 与 CrowdStrike 等安全平台、Datadog 等可观测平台，都有机会承接动态授权、审计、异常检测和运行监控。这个方向的需求较确定，利润归属仍取决于谁能跨系统看到完整任务，单一告警功能可能被更大的控制平面整合。
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；不是专家访谈，是作者自己的分析。 - 口径不齐 Bits 没有单独披露 ARR，分不清 AI 客户使用和 Agent 带来的增量
可观测性天然满足“持续发生、可计量、任务绕不开”三项条件。Datadog 的收入和现金流也比许多 AI 产品发布更扎实。当前缺口是 Bits 尚未单独披露 ARR，投资者仍要区分“AI 客户使用 Datadog”与“Datadog 因 Agent 获得多少增量收入”。
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 - 反对 Bits AI 收入“还看不到”，正从单一调查 SKU 改为 AI credits
They emailed us straight on that. So, it's pretty transformative for them. In terms of where it is on the results, you won't really see it yet on the results. We're actually in the process right now of repackaging these products. We used to have one SKU we sold for – specifically for investigations. And now, we're packaging it to a set of AI credits instead that can work across a lot of different surfaces, because as I said, we expanded the product to cover a lot of different surfaces. And so, you should – we probably will comment on that in the future. But, for now, there's a – the transformation's happening right now in terms of the way we're charging those.
- 口径不齐 安全 agent 与 Cloud SIEM 解耦，AI 能力可用于不进 Datadog 的数据，按 AI credits 在数据消费之上另收
Nothing about what comes out, nothing about what happens inside the platform, because the volumes of data were enough to model basically the usage and the value we provide to our customers. In the future, so we are shipping those AI agents. We have some AI agents that even work on data we don't ingest. So, we're – we announced the decoupling of our security agent from our Cloud SIEM, for example. So, you can actually use our AI smarts on all the data source today. So, there's – and we also can generate code, we can do things that don't necessarily manipulate a lot of data. So, for now, we're packaging that as AI credits that our customers can buy on top of the rest of the data consumption. So far, it seems to be very well accepted. But again, we'll see where this goes. If we need to repackage that in different ways, attach that to different parts, such as data volumes, the host or anything else, that's something we can easily do.
- 中性 业界还不知道怎么给 agent 定价，Datadog 称可随时改 SKU
And so, you should – we probably will comment on that in the future. But, for now, there's a – the transformation's happening right now in terms of the way we're charging those. And by the way, if you zoom out, the world as a whole doesn't quite know how to charge for agents just yet. At the low level, you have the model companies are charging by token. And then, above that, it's a little bit unclear. As far as we're concerned, we don't really care. We are completely usage-based.
- 支持 安全工具被模型颠覆，CEO 认为这抹平了安全行业在位者的优势，对 Datadog 是机会
And so, they are – the DevSecOps is going to – is not going to be an emerging trend anymore, but it's going to be the way most companies are running things. So that's number one. Number two is, and that one is even newer, like we've seen over the past three months, that security tooling has been completely upended. We've seen that the models are so good and they move so fast that pretty much anything that was built more than a few years ago in security is obsolete. And the very notion that you're going to have 12, 15, 25 different security products installed that are all contributing to a central repository that then humans review and prioritize, that's just not going to work anymore.
- 中性 Agent 观测本质上与传统 APM 差异不大 (评估、成本、定位出错子步骤)
JC: Sure. So, there's the evaluation aspect, there's the cost aspect, hallucinations are part of that evaluation portion. And then, being able to dig into the traces and figure out what tools are being called and, like, let's say, if you have an agentic flow, that means that there's generally more than 1 step. So if something is going wrong, so to speak, as part of that flow, you need to be able to identify which substep is going haywire and then dive deeper into it. So it really is not so dissimilar from classical APM solutions.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 反对 自主修复 (Bits AI 修复) 难在客户不愿放开生产系统权限
JC: I think the difficult part is the actual remediation part. Datadog and other observability tools and platforms have historically been more about surfacing things that are going wrong or need some level of investigation for then humans to step in and make some decisions about that. Even going back a year and a half, prior to this AI boom, there was already a push to make Datadog more of an action-oriented platform. But there is understandably reticence and hesitation from the customer perspective to allow unfettered access to production systems to take these actions, if that makes sense.
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 反对 他现在的公司 LLM 观测没选 Datadog (RFP 时 Gemini 支持不足)，计划一个月后重新评估
JC: Yeah, for me, I am definitely biased to the 1 tool as opposed to a constellation of specific tools in general, unless there is really a feature gap or again, the cost is not something that we want to bear. In my current organization, we actually are not using Datadog for LLM observability because at the time when we did the RFP, I think Gemini support
来源背景：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 - 支持 Toto/Bits 愿景：客户按问题类型选择自动修复或人工确认
That will be – when we get there – we're not there yet – will be a combination of the evolution of the platform. Our Bits product line is what is going to use these models in data to be predictive. And then clients, in our vision, will be able to make the choice of how much to auto-remediate to say, for this type of issue, the platform can auto-remediate. For this type, there'll be a suggestion, and then someone will have to press yes. And that has tremendous ramifications, both in terms of the speed and also the efficiency in human capital in this endeavor.
- 支持 Resolve AI 等 AI SRE 公司听起来好但不提供真正的观测数据底座 (无指标处理)
MC: I mean, I think Resolve AI is one that on the surface sounds good. There are deep pockets backing it up, but they are not actually providing the actual observability. There is no metric handling, etc. Conversely, Datadog or Grafana, who have that back end, that type of infrastructure... from a cloud-based or a self-hosted perspective like with Grafana, I think those companies are the ones that are less risky if you want to take a stab at some of their AI for observability versus observability for AI, or just the kind of traditional observability type of play. That is the difference. There are some businesses who are just strictly the AI for observability component and not providing the whole kit and caboodle, which I think is required if you are an enterprise organization.
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 支持 AI 创业公司更像 'snake oil' 风险而非颠覆者
MC: I don't believe that. I guess it's, yeah, it's too soon to tell. I don't think that there's... I haven't seen from, you know, I keep a keen eye on the market and what tools are coming out. I haven't really seen anything that's super compelling from a startup perspective that would, aside from potential like acquisition candidates, nothing that would go in to completely disrupt the industry. That said, I do see some of these... I think there's a bigger risk that there's snake oil out there than that there's something that's going to be a major disruptor. From the AI startup side, I think it's a perspective, yeah.
来源背景：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 - 支持 安全 ARR 超 $100M，>$1M ARR 客户中安全支出占比 <2%
Truist says meetings with DDOG IR reinforced that guidance remains conservative around its largest AI-native customer, with 4Q assumptions based on contractual minimum usage even though 3Q is tracking above that floor. Truist says IR described growth as broad-based across products, cohorts and verticals, with non-AI-native customers also accelerating as “getting-ready-for-AI” investments move into production. Truist also sees security as an underappreciated growth leg after crossing $100M ARR, with SIEM leading adoption and security still <2% of spend among >$1M ARR customers, leaving substantial expansion runway.
来源背景：TMT Breakout (TMTB) 日报，转述 Truist 与 DDOG IR 会面后的报告摘要。二手转述，看不到底稿。 - 口径不齐 Datadog 当前受益仅来自 LLM 监控；agent 管控层 '收入上还没开始'，可能仍是研发项目
Bernd Harzog: No, no, they're not. No, what they're benefiting from is something as simple as LLM monitoring. Their customers are building AI functionality into the applications that they use to run their business, and they need that AI functionality to be monitored to make sure it's working correctly. So, they're benefiting from the very first part of this, as is Dynatrace. And they've only begun to benefit from the rest. They've only begun to benefit from vibe coding. They've only begun to benefit from agents in production. They've not started. I mean, if you want, I'm sure they started from an R&D perspective, but they've not started it at all from a revenue perspective to benefit from the question of how do we notice bad AI behavior and control it. That's a brand new problem that dramatically expands the TAM and the value of observability solutions. And I would venture to guess that that's an R&D project at both companies.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 口径不齐 agent 在生产中的监控/管控尚未贡献收入
Bernd Harzog: No, no, they're not. No, what they're benefiting from is something as simple as LLM monitoring. Their customers are building AI functionality into the applications that they use to run their business, and they need that AI functionality to be monitored to make sure it's working correctly. So, they're benefiting from the very first part of this, as is Dynatrace. And they've only begun to benefit from the rest. They've only begun to benefit from vibe coding. They've only begun to benefit from agents in production. They've not started. I mean, if you want, I'm sure they started from an R&D perspective, but they've not started it at all from a revenue perspective to benefit from the question of how do we notice bad AI behavior and control it. That's a brand new problem that dramatically expands the TAM and the value of observability solutions. And I would venture to guess that that's an R&D project at both companies.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 支持 在 LLM 监控上 Datadog 进展领先所有厂商；Dynatrace 在自动化修复型 AI 上领先
Bernd Harzog: I would say Dynatrace is ahead of everyone from that perspective. Okay. All right. If you define AI as being able to monitor LLM, right? I would say Datadog has made more progress in monitoring LLM for their customers than any other vendor. Okay. Now, the next thing we'll have to watch closely is agents in production, right? You know, Datadog is very, very good at noticing what are the new things that their customers are doing and putting in place a focused effort to capture those new things.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 支持 一位客户用 DDOG SRE 调查 agent 几周，平均约 3 分钟到根因，无需人工和 dashboard
* One customer ran Datadog’s SRE investigation agent over a few weeks, averaging just over three minutes to root cause with no human and no dashboards opened, which is concrete evidence for paid AI-SRE adoption. * Li says customers keep asking Datadog to raise rate limits on its own platform, implying agents now query Datadog APIs at machine scale; that could open a usage-based monetization path or create cost pressure (our read). * Li frames incident war rooms of tens to hundreds of people as mostly teams proving their innocence; investigation agents remove that labor, which supports a labor-replacement pricing story rather than a seat-based one (our read). * Datadog builds its own models, and Li says a lot of AI’s lift to observability demand still comes from core metrics, traces and logs, so agents should boost the base business, not just new SKUs.
来源背景：TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。 - 反对 Li 承认存在 '谁来消费遥测' 的争夺：轻量 MCP、DDOG 自家 agent 还是第三方跨厂商 agent
On how much context to surface via MCP, Li says Datadog has had to curate formatting and query mechanics rather than ship a raw API wrapper, and notes an emerging “interesting battle” over whether customers consume telemetry via lighter MCP access, Datadog’s own curated AI agent, or third-party agents synthesizing data from multiple vendors. She argues dashboards aren’t disappearing — some customers still find clicking through known dashboards faster than natural-language queries — but consumption surfaces are proliferating (Slack, CLI, IDE), and AI agents increasingly generate dashboards just-in-time for a specific investigation rather than reading pre-built ones.
来源背景：TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。 - 中性 AI 的 ROI/tokenomics 尚未解决，价值与成本不像云那样直接相关
Li identifies deep agent-behavior understanding as only “the first step,” with trust and security as the next layer — she calls AI security unresolved and “multi-edged,” since AI can be the target, the attacker, or the attack vector, and argues security and observability increasingly overlap because both require understanding every agent action. She also flags tokenomics/ROI as an open problem: unlike cloud infrastructure, where footprint growth tracked business growth, AI’s value-to-cost relationship is not directly correlated, and the industry hasn’t yet linked token-level cost efficiency to business outcomes. Datadog, she adds, already performs both static and dynamic model routing internally, mixing large models, small models, and basic ML models depending on reliability needs.
来源背景：TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。
E06 预期和仓位
判断：预期很高，而且市场已经在"指引 + 惯常 beat"的基础上定价。
指引方法：CFO 反复解释：用近几个月的趋势和长期 NRR，再打一个折扣，客户实际花得比打折后多，所以能"beat and raise"。历史上每季大约超指引 4–5 个点。2026 年的轨迹是：2 月首次指引全年 +18–20% (剔除最大客户 ≥20%) → 5 月上调到 +25–27% → 8 月上调到 +30%。2025 年初也是指引 19%，实际 26%+。
现在的指引隐含什么：Q3 指引 $1.135–1.145B，同比 +28–29%，比 Q2 实际只多 $1,400–2,400 万；全年指引 $4.45–4.47B 减去上半年和 Q3 中值，隐含 Q4 约 $11.9 亿，同比约 +25%。即使照惯例超 4–5 个点，Q4 也是约 30%，比 Q2 的 36% 明显减速。8 月 Q2 业绩后股价跌 13%，就是因为 Q3 指引低于买方约 31% 的期待。
卖方和买方：卖方目标价大多在 $260–340 (BofA $260、Citi $270、Wedbush $275、MS $300、Truist $300、Wolfe $340)；Bernstein 7 月以估值降级到 Market Perform。估值上 UBS 承认约 20 倍收入、75 倍 FCF，MS 承认约 16 倍 2027 年销售"不便宜"。仓位：Funda 在 Q2 前提示"股价自上次财报涨 83%，仓位拥挤"；BNP 的经销商调查里 DDOG 的前瞻支出分数从 5.4 降到 2.0。
I would say, in the bubble after COVID, there was – in the ZIRP side, there was probably some usage that was over-usage. But we step in and we help clients then use it effectively. We help them figure out what logs to put in and what logs not. And then in terms of guidance, what we're able to do and that's how we can beat and raise is we take that growth rate that we see over a long period of time and we discount it, so that we have that cushion. And then, what's been proven out in the beat and raise is that the clients spend more than the amount we're discounting. So, let's say net retention is blank. We discount net retention and then we have a long time series.
As a reminder, we base our guidance on trends observed in recent months and imply conservatism on these growth trends. Regarding our largest customer, we have seen a usage reduction, which is incorporated in our Q3 and full year 2026 guidance. As Olivier noted, this customer has recently renewed with us. For the third quarter, we expect our revenue to be in the range of $1.135 billion to $1.145 billion, which represents a 28% to 29% year-over-year growth. Non-GAAP operating income is expected to be in the range of $260 million to $270 million, which implies an operating margin of 23% to 24%. And non-GAAP net income per share is expected to be in the $0.63 to $0.65 per share range based on approximately 378 million weighted average diluted shares outstanding. For the full fiscal year 2026, we expect revenue to be in the range of $4.45 billion to $4.47 billion, which represents a 30% year-over-year growth.
Truist says meetings with DDOG IR reinforced that guidance remains conservative around its largest AI-native customer, with 4Q assumptions based on contractual minimum usage even though 3Q is tracking above that floor. Truist says IR described growth as broad-based across products, cohorts and verticals, with non-AI-native customers also accelerating as “getting-ready-for-AI” investments move into production. Truist also sees security as an underappreciated growth leg after crossing $100M ARR, with SIEM leading adoption and security still <2% of spend among >$1M ARR customers, leaving substantial expansion runway.
来源背景：TMT Breakout (TMTB) 日报，转述 Truist 与 DDOG IR 会面后的报告摘要。二手转述，看不到底稿。
* **We expect 3Q26 guidance around +30%, above consensus of +25% and matching buyside bogey; actual growth should hold near 35% or better** * **We expect FY26 guidance to be raised by roughly $100m or 300bps to about $4.42B/+29%**, an increase larger than the combined ~$80m from the 2Q beat and the above-consensus 3Q guide * On valuation: the shares are up 83% since the last print (versus IGV +6%), and positioning is crowded, so while results are likely to stay strong, the post-print share reaction will be hard to replicate at last quarter’s level
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。
Bernstein raised its long-term estimates for Datadog, citing increasing confidence in the company’s AI platform opportunity, but downgraded the stock to Market Perform on valuation and elevated near-term expectations. The firm expects a strong 2Q, but believes growth could disappoint beginning in Q3/Q4 as ex-AI demand (~85% of revenue) flattens and AI lab growth starts to normalize. Bernstein expects Q4 revenue growth to slow roughly 500bps to ~29% y/y, well below investor expectations for 30%+ growth into next year and peak growth in the high-30% to 40%+ range.
来源背景：TMT Breakout (TMTB) 日报，转述 Bernstein 降级报告摘要。二手转述，看不到底稿。
展开本节全部 89 条原文 (支持 40 / 反对 11 / 其他 38，按时间)
- 中性 年初指引 19%，实际跟踪到 26% 以上；指引方法为"按近月趋势再打折"
First, our guidance philosophy overall remains unchanged. As a reminder, we based our guidance on trends observed in recent months and imply conservatism on these growth trends. For the fourth quarter, we expect revenue to be in the range of $912 million to $916 million, which represents a 24% year-over-year growth. Non-GAAP operating income is expected to be in the range of $216 million to $220 million, which implies an operating margin of 24%. Non-GAAP net income per share is expected to be in the range of $0.54 to $0.56 per share based on approximately 367 million weighted average diluted shares outstanding.
- 中性 分析师指出年初预期 19% 增长、现在跟踪 26%+；管理层称 AI 采纳快于预期
I mean, look, the -- so first, I apologize for reading on the results. We might do it again, but we'll see. The -- I think the biggest surprise for us has been that the -- so AI in general has -- AI adoption has grown faster than we thought it would at the beginning of the year. So we've seen that across our AI cohort. We've seen also that we got some of our new products and new -- like the changes we're making on the go-to-market side to click perhaps earlier than we would have thought otherwise. So all in all, we saw the leading part of the business with AI growth faster, not the lag, but the slower growing, more tensional property business also accelerate and that gives us where we are today.
- 中性 高盛：DDOG 等公司业绩好但股价下跌，反映市场担心 AI 冲击
* 过去两周，软件行业的市场情绪变得更加悲观：尽管一些公司公布了良好的财报（例如 DDOG、DT、HUBS、TEAM），但其股价却下跌；人工智能是否是一种颠覆性力量，它会改变（甚至蚕食）定价模式，降低新进入者的门槛，并最终压缩领先SaaS公司可获得的利润空间？ * 基础设施层已投入约1万亿美元的资本支出，但迄今为止，企业软件应用领域创造的价值却有限（例如 Palantir），因此我们认为，SaaS领域的领导者有机会利用大型语言模型（LLM）工具来弥合这一差距，从而提高企业生产力。 * 为了直接回应这场争论：我们不认为软件行业“已死”或处于长期衰退之中——在某些情况下，我们认为人工智能有可能成为该行业领先供应商的倍增器。 * 我们认为，该行业正在经历一场转型，其中蕴藏着机遇，尤其是在2026年，随着周期性利好因素（例如客户续约）缓解增长压力，新的人工智能贡献可以为关键指标（例如净收入留存率）提供稳定性，并为可持续的多年增长铺平道路。
来源背景：Acecamp 作者整理的卖方报告和买方讨论 (不是专家访谈)；涉及 DDOG 的是转述的公开数字和高盛的观点。 - 中性 该渠道 2026 年的 Datadog 业务目标是 +30%
A:我们有一个非常粗略的预测。他们以我们去年所做为基础，而且几乎不会把它降低——如果有的话，通常是提高。我们为 2026 年定了 30% 的增长目标，我希望我们能实现它。
来源背景：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 - 支持 1H26 预计 22–24.5%，比 2025 年加速
A:For H1 2026, we currently believe we will be able to hit anywhere from 22 to 24.5%. That is an uplift over most of 2025.
来源背景：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 - 中性 作者预计 FY26 实际增速 24–25%，首次指引 18–19% (符合买方预期，低于卖方一致预期 21%)
* 在上个季度的调研中我们提到，尽管DDOG已经在3Q将Log management迁移到自研的基于Clickhouse的OLog上，随着mid-training的兴起，DDOG对于LLM researcher重要性大幅提升，使得OpenAI与DDOG完成续约，并且2026年用量仍可能实现增长 * **然而我们最新的调研显示，Chronosphere在metrics上的功能持续提升，因此最近在OpenAI的内部用量也在快速上升，从而导致OpenAI对DDOG的依赖有所下降。**我们预计metrics在4Q25占OpenAI DDOG ARR的40%，该部分可能在2026年中完全迁移到Chronosphere中。余下的trace部分仍将继续在DDOG上使用。**总体而言，我们预计OpenAI DDOG ARR到2026年底将下滑20%以上。** * 另一方面，DDOG其他AI native客户的用量在快速提升，将弥补部分OpenAI用量的损失。其中Anthropic在小范围部署database monitoring及infra monitoring * **综合考虑，我们初步认为DDOG 2026收入增速有望达到24-25%，因此首次给出的指引可能落在18-19%yoy，符合买方市场预期**，低于卖方一致预期的21%。在目前市场对软件的悲观看法下，不够出彩的指引仍难以改善市场情绪
来源背景：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 - 反对 他对 2026 年的预期从 28–30% 下调到约 25%
A:如果我没记错，我们说过预计增长在 20 多个百分点的高位。具体来说，我们预计接近 28% 到 30%。我们现在的预测略低，约为 25%。我们在 2022–23 年的结果为 25%，所以我们只是预测会有几个百分点的加速，而不是以同样速度达到 20 多个百分点的高位。一开始你提到我们比三个月前更谨慎有几个原因。第一是竞争。
来源背景：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 - 中性 FY26 指引 $4.06-4.10B (+18-20%)，按剔除最大客户的业务至少增长 20% 建模
Non-GAAP operating income is expected to be in the range of $195 million to $205 million, which implies an operating margin of 21%. Non-GAAP net income per share is expected to be in the $0.49 to $0.51 per share range based on approximately 367 million weighted average diluted shares outstanding. And for the full fiscal year 2026, we expect revenues to be in the range of $4.06 billion to $4.10 billion which represents 18% to 20% year-over-year growth. This includes modeling within our guidance that our business, excluding our largest customer, grows at least 20% during the year. Non-GAAP operating income is expected to be in the range of $840 million to $880 million, which implies an operating margin of 21%. And non-GAAP net income per share is expected to be in the range of $2.08 to $2.16 per share based on approximately 372 million weighted average diluted shares.
- 中性 CFO 解释：指引隐含核心业务增速高于最大客户；对最大客户"不控制、做了非常保守的假设"
Yes. There are 3 questions in there. The first is overall on guidance, except what we're going to speak about next, we took the same approach as we looked at the organic growth rates and the attach rates and then the logo accumulation rates and discounted that. So for the overall business, which is quite diversified, we talked about diversification by industry, by geography, by SMB, mid-market and enterprise, we took the same approach. We noted that with the guidance being 18% to 20% and the non-AI or heavily diversified business being 20% plus, that would imply that the growth rate of that core business assumed in the guidance is higher than the growth rate of the large customer. It doesn't mean the large customer is growing any which way. It's just that in our consumption model, we essentially don't control that. And so, we made a very conservative assumption there. And the last point, I think you mentioned is the highly diversified. We said 650 names in the AI cohort is quite diversified, essentially would be very similar to our overall business, which we have a range of customers, but not the concentration level. And what we're seeing there is significant growth. But like our overall distributed customer base, a growth and then potentially some working on how the product is being used, but nothing out of the ordinary relative to the overall customer base in the very diversified AI set of customers outside the largest customer. Hopefully that's helpful.
- 中性 Q1 2026 指引 $951-961M (+25-26%)，营业利润率 21%
Our guidance philosophy overall remains unchanged. As a reminder, we based our guidance on trends observed in recent months and apply conservatism on these growth trends. For the first quarter, we expect revenues to be in the range of $951 million to $961 million, which represents a 25% to 26% year-over- year growth. Non-GAAP operating income is expected to be in the range of $195 million to $205 million, which implies an operating margin of 21%. Non-GAAP net income per share is expected to be in the $0.49 to $0.51 per share range based on approximately 367 million weighted average diluted shares outstanding.
- 中性 1Q26 指引 25–26%；作者认为按 5pt beat 算会到 30% 以上
* 4Q adj OPM扩张0.7pt qoq, 超市场预期, 反应公司在去年开始扩张的大企业销售团队产效持续提升 * 1Q26收入指引25-26% yoy，超市场预期1pt，**若假设5pts beat则意味着1Q收入将加速增长30%以上** * **最受关注的FY26全年收入指引初始值定在18.5-19.5%yoy，比市场担忧的18-19%或更低要好；考虑到core仍在持续加速，该指引较为保守地假设了OpenAI的优化进展**
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 支持 FY26 指引 18.5–19.5%，好于市场担忧的 18–19%，作者认为假设保守
* 4Q adj OPM扩张0.7pt qoq, 超市场预期, 反应公司在去年开始扩张的大企业销售团队产效持续提升 * 1Q26收入指引25-26% yoy，超市场预期1pt，**若假设5pts beat则意味着1Q收入将加速增长30%以上** * **最受关注的FY26全年收入指引初始值定在18.5-19.5%yoy，比市场担忧的18-19%或更低要好；考虑到core仍在持续加速，该指引较为保守地假设了OpenAI的优化进展**
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 支持 作者把 FY26 收入预期上调到 +26%
我们略为上调对FY26全年收入预期至26%yoy
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 中性 渠道专家 A 自己 book 1Q26 仅约 21%，低于内部 22–24.5% 目标
* Internal 1H26 target of **+22–24.5% y/y; 1Q26 came in at only ~21% y/y, below plan**
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 支持 仅靠现有 pipeline 2H26 约 22–23%，加新签可到 25%
+ Based on existing pipeline alone, 2H26 tracks ~22–23% y/y; with new bookings, growth could reach 25% + 2H acceleration is driven by LLM observability, Bits AI and DevOps tools
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 中性 Funda 财报前预期：1Q 实际 +31–32%、2Q 指引 +24–25% (卖方约 20%)
* 1Q26: core stable at the lower end, AI-native (ex-OpenAI) continuing to contribute incrementally; **we estimate actual revenue growth of +31–32% y/y, extending the trend of beating the high end of guidance by ~5 ppts over the past several quarters** * **2Q26 guidance is expected at +24–25% y/y, above sell-side expectations of ~20%**
来源背景：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 - 中性 对最大客户沿用上季做法，施加更高的保守度
First, our guidance philosophy overall remains unchanged. As a reminder, we base our guidance on trends observed in recent months and apply conservativism on these growth trends. In addition, as with last quarter, we are applying a higher degree of conservatism to our largest customer. So, for the second quarter, we expect revenues to be in the range of $1.07 billion to $1.08 billion, which represents a 29% to 31% year-over-year growth. This guidance implies sequential revenue growth of $64 million to $74 million, or 6% to 7%, due to the strong growth of revenue in Q1 and into April.
- 中性 CFO 澄清：对最大客户的折扣力度高于其他客户，方法论与以往相同
It's both. It's the same guidance we used, and we're being very explicit. For all the business except for this largest customers, we've always taken the drivers and discounted them. And for this particular customer, we took a higher degree of conservatism than the other part of the customer base and discounted it more. And we were, I think, in the remarks and you interpret it correct, very explicit. And you're correct.
- 中性 指引方法：基于近几个月趋势再打折扣
Our cash flow from operations was $335 million in the quarter, after taking into consideration capital expenditures and capitalized software, free cash flow was $289 million and free cash flow margin was 29%. And now for our outlook for the second quarter and for the fiscal year 2026. First, our guidance philosophy overall remains unchanged. As a reminder, we base our guidance on trends observed in recent months and apply conservativism on these growth trends. In addition, as with last quarter, we are applying a higher degree of conservatism to our largest customer. So, for the second quarter, we expect revenues to be in the range of $1.07 billion to $1.08 billion, which represents a 29% to 31% year-over-year growth.
- 支持 Q2 指引 $1.07–1.08B (+29–31%)，隐含环比 +$64–74M
As a reminder, we base our guidance on trends observed in recent months and apply conservativism on these growth trends. In addition, as with last quarter, we are applying a higher degree of conservatism to our largest customer. So, for the second quarter, we expect revenues to be in the range of $1.07 billion to $1.08 billion, which represents a 29% to 31% year-over-year growth. This guidance implies sequential revenue growth of $64 million to $74 million, or 6% to 7%, due to the strong growth of revenue in Q1 and into April. Non-GAAP operating income is expected to be in the range of $225 million to $235 million, which implies an operating margin of 21% to 22%. As a reminder, in Q2 we will be holding our DASH user conference, which we estimate to cost about $15 million and which we have reflected in our operating income guidance.
- 支持 FY26 指引上调到 $4.30–4.34B (+25–27%)
As a reminder, in Q2 we will be holding our DASH user conference, which we estimate to cost about $15 million and which we have reflected in our operating income guidance. Non-GAAP net income per share is expected to be $0.57 to $0.59 per share, based on approximately 369 million weighted average diluted shares outstanding. And for fiscal 2026, we expect revenues to be in the range of $4.3 billion to $4.34 billion, which represents 25% to 27% year-over-year growth. Non-GAAP operating income is expected to be in the range of $940 million to $980 million, which implies an operating margin of 22% to 23%. And non-GAAP net income per share is expected to be in the range of $2.36 to $2.44 per share, based on approximately 372 million weighted average diluted shares outstanding.
- 支持 billings +37%、RPO $3.48B +51%、cRPO mid-40s，多年期合同占比上升
Next, regarding our retention metrics, our trailing 12-month net revenue retention percentage was in the low 120% percent, up from about 120 last quarter, and our trailing 12-month gross retention percentage remains in the mid-to-high 90s. Now moving on to our financial results. Billings were $1.03 billion, up 37% year-over-year, and remaining performance obligations, or RPO, was $3.48 billion, up 51% year-over-year, with current RPO growing in the mid-40s percent year-over-year. RPO duration increased year-over-year as the mix of multi-year deals increased in Q1. As a reminder, we continue to believe revenue is a better indicator of our business trends than billings and RPO, given their variability. Now, let's review some of the key income statement results.
- 支持 1Q26 收入 +32%，比买方预期高 1pp，较 4Q25 加速 3pp
* 1Q26 revenue grew 32% yoy, beat buyside expectation by 1ppt and in line with our higher-than-consensus estimate, representing a 3ppt acceleration from 4Q25.
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 2Q 指引 +30%，比市场预期高 6–7pp；按 4–5pp 惯常 beat 实际可达 35%
+ The “core” non-AI customer cohort **grew at a mid-20% year-over-year rate, up from 23% in 4Q25** and 19% in 3Q25. + This implies **AI-native customer growth accelerated to 17% QoQ**, after growth was impacted by OpenAI’s migration out in 4Q, demonstrating Anthropic and other AI native customers’ consumption picking up. * 4Q adj OPM beat expectation by 1ppt, demonstrating strong operating leverage despite GPM being lower, likely due to the discount offered to large customers * 2Q26 revenue guidance is set at 30% year-over-year growth, surpassing market expectations by 6-7 percentage points. **Based on a normal beat cadence of 4-5 percentage points, actual growth could reach 35% year-over-year, the highest** **since 2022.** * Full-year revenue guidance is now 26-27% year-over-year growth. Factoring in **typical beat cadence, FY26 revenue is projected to accelerate to low-to-mid 30% growth.**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 FY26 指引 26–27%，加上惯常 beat 预计全年低到中 30s
+ The “core” non-AI customer cohort **grew at a mid-20% year-over-year rate, up from 23% in 4Q25** and 19% in 3Q25. + This implies **AI-native customer growth accelerated to 17% QoQ**, after growth was impacted by OpenAI’s migration out in 4Q, demonstrating Anthropic and other AI native customers’ consumption picking up. * 4Q adj OPM beat expectation by 1ppt, demonstrating strong operating leverage despite GPM being lower, likely due to the discount offered to large customers * 2Q26 revenue guidance is set at 30% year-over-year growth, surpassing market expectations by 6-7 percentage points. **Based on a normal beat cadence of 4-5 percentage points, actual growth could reach 35% year-over-year, the highest** **since 2022.** * Full-year revenue guidance is now 26-27% year-over-year growth. Factoring in **typical beat cadence, FY26 revenue is projected to accelerate to low-to-mid 30% growth.**
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 支持 DDOG 1Q 后 +30%，FY26 指引 26% 远超街 20%，缓解 OpenAI 断崖担忧
* **DDOG +30%** after their stellar beat and raise with revs reaccelearting to 32% from 29% in prior q. FY26 guide now calls for 26% ahead of street which was closer to 20%. The midpoint of FY26 was raised more than combined Q1 beat and upside to Q2 guide, implying a strong 2H and mitigating worries of growth cliff related to OpenAI. Overall, call was very bullish with mgmt saying that before they were more focused on inference. Now they are increasingly seeing training become a market as they are landing hyperscalers who are using DDOG despite having a whole suite of homegrown capabilities. As *Josh S* put it in TMTB Slack: *feels like this call was pretty much as bullish as u could hope for. only thing they could've done more was maybe just spend 5 minutes saying "anthropic" 50x*
来源背景：TMT Breakout (TMTB) 日报，作者市场评论 + TMTB Slack 成员 Josh S 的评论，转述业绩会要点。二手转述，看不到底稿。 - 中性 TMTB 区分：DDOG 和网络安全受 AI 需求推动，WDAY/NOW/FIG 等应用软件不同
Good afternoon. **QQQs -12bps** but one of those days where the story underneath the surface tells a lot more interesting story. High beta momentum down 8% with SPX/NDX down less than 50bps. Semis -3% vs. Software +3% and the gap widened through the day. The former weighted down by ARM/COHR misses and the latter helped by DDOG/FTNT’s better prints, giving software bulls hope that there is alpha to be had in the space and that a good print will be rewarded with a good move. However, DDOG and FTNT different than let’s say most names of app sw. One could make a strong argument that DDOG & FTNT (or cyber in general) is helped by AI demand, while the case is hard to make for something like WDAY or NOW or FIG. It’s still unclear how that latter group would respond to strong numbers. Some unwindy action in internet too with META outperforming favorite longs GOOGL/AMZN on the Megacap side while shorted names like **RBLX +2%** and **DUOL +8%** although **DASH +2%** sold off throughout the day and **CART -5%** followed through on yesterday’s miss so price action outside of **AMZN/GOOGL** continues to feel anemic here.
来源背景：TMT Breakout (TMTB) 日报，作者市场评论 + TMTB Slack 成员 Josh S 的评论，转述业绩会要点。二手转述，看不到底稿。 - 支持 TMTB：若继续 5–6% beat，年底可达 high-30s
The implied 2H numbers big story here. If we assume more 5-6% beats you could exit the year in the high 30s.
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 1Q26 公布数字与街上预期 (Street 口径来源未注明)，转述 CEO 评论。二手转述，看不到底稿。 - 支持 1Q 收入 $1.006B (+32%) vs 街 $960M；billings $1.03B (+37%) vs 街 $932M
* Revenue $1.006B, +32% y/y (last q +29%) vs Street $960M, with growth accelerating driven by continued enterprise adoption, AI-native customer strength, and core observability demand. Calculated billings were $1.03B, +37% y/y vs Street $932M, a notable reacceleration and well ahead of expectations. * Operating margin came in at 22.2% vs Street ~21.3%.FCF was $289M vs Street $274M. * 2Q revenue guide of $1.07-1.08B (+29-31% y/y) came in ahead of Street at $994M (+20%), while FY26 revenue guidance was raised to $4.30-4.34B (+25-27%) vs Street $4.12B. The implied guide raise was materially above the size of the 1Q beat, pointing to a stronger 2H setup. * FY26 operating margin guidance was also lifted to 21.9-22.6% vs Street ~21.4% * Customer metrics remained healthy, with $100K+ ARR customers growing to 4,550 from 4,310 last quarter, while net retention rate improved into the low-120% range after bottoming near ~120% last quarter. Total customers reached 33.2K. * Mgmt commentary and investor focus centered around AI-native customer contribution, momentum in newer products like LLM Observability, Cloud SIEM and CNAPP, as well as whether broader enterprise AI deployments are beginning to drive incremental consumption. Investors also focused on whether FY guide conservatism implies further upside potential if AI-related workloads continue ramping through 2H.
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 1Q26 公布数字与街上预期 (Street 口径来源未注明)，转述 CEO 评论。二手转述，看不到底稿。 - 支持 2Q 指引 $1.07–1.08B vs 街 $994M；FY26 $4.30–4.34B vs 街 $4.12B
* Revenue $1.006B, +32% y/y (last q +29%) vs Street $960M, with growth accelerating driven by continued enterprise adoption, AI-native customer strength, and core observability demand. Calculated billings were $1.03B, +37% y/y vs Street $932M, a notable reacceleration and well ahead of expectations. * Operating margin came in at 22.2% vs Street ~21.3%.FCF was $289M vs Street $274M. * 2Q revenue guide of $1.07-1.08B (+29-31% y/y) came in ahead of Street at $994M (+20%), while FY26 revenue guidance was raised to $4.30-4.34B (+25-27%) vs Street $4.12B. The implied guide raise was materially above the size of the 1Q beat, pointing to a stronger 2H setup. * FY26 operating margin guidance was also lifted to 21.9-22.6% vs Street ~21.4% * Customer metrics remained healthy, with $100K+ ARR customers growing to 4,550 from 4,310 last quarter, while net retention rate improved into the low-120% range after bottoming near ~120% last quarter. Total customers reached 33.2K. * Mgmt commentary and investor focus centered around AI-native customer contribution, momentum in newer products like LLM Observability, Cloud SIEM and CNAPP, as well as whether broader enterprise AI deployments are beginning to drive incremental consumption. Investors also focused on whether FY guide conservatism implies further upside potential if AI-related workloads continue ramping through 2H.
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 1Q26 公布数字与街上预期 (Street 口径来源未注明)，转述 CEO 评论。二手转述，看不到底稿。 - 支持 Funda 重申：2Q 指引 +30% 按 5pp beat 约 35%，FY26 上调实际低中 30s，OpenAI 断崖叙事退场
**Review|DDOG 1Q26: Big Guidance Upside Surprise, No Growth Cliff Concern Anymore.** We were bullish into the event but earnings were even stronger than our higher-than-consensus estimates. 1Q26 met expectations but the real story was the guide — 2Q26 +30% YoY (implying ~35% with the typical 5pt beat) and FY26 raised to 26-27% (low-to-mid 30s on actuals), effectively retiring the 2H OpenAI cliff narrative. Two new hyperscaler AI training wins reinforce DDOG’s tech moat over in-house and OTel alternatives.
来源背景：FundaAI 周报，DDOG 部分是 5/5 preview 和 5/7 review 的摘要重述；仓位部分引用 Coatue 演示材料。二手转述，看不到底稿。 - 中性 Coatue 半导体+基础设施多头敞口从年初 35% 升到 58%，以软件/互联网为代价
A week dominated by tons of earnings reports and a relentless rally in AI semis, especially memory and CPU stocks. As our long-term readers would recall, we have consistently highlighted the “mis-positioning” by many US funds, which significantly underweight semi/hardware while overweighting software/internet, and, more importantly, they are starting to adjust. This was confirmed by Coatue’s latest presentation, in which they disclosed that Semis + Infra’s long exposure % jumped to 58% as of late, up from 35% at the beginning of the year. This is at the expense of Software, Internet + Other exposure decreased materially.
来源背景：FundaAI 周报，DDOG 部分是 5/5 preview 和 5/7 review 的摘要重述；仓位部分引用 Coatue 演示材料。二手转述，看不到底稿。 - 支持 Whale Rock 与 Maplelane 在 Sohn AI panel 上称 DDOG 是 '一切的明确赢家'
Software will have winners and losers. Datadog is viewed as the clear winner from everything. Over the last 45 days, more separation has emerged. New ways of writing code are fundamentally different. Software is sticky for a period, but it is no longer at the top of the CIO list — tokens and AI are.
来源背景：TMT Breakout (TMTB) 日报，转述 Sohn 投资大会上 Whale Rock (Alex Sacerdote) 和 Maplelane (Leon Shaulov) 的 AI panel 观点。买方观点，TMTB 摘要。二手转述，看不到底稿。 - 反对 CEO 自己预警：AI 公司收入爆发的钱来自别处，“合理化”可能比预想来得早
That's for later. That later might come sooner. If we see the explosion of the AI company's revenue continue at the rate it is right now, that money is coming from somewhere and there's probably going to be a push for rationalization sooner rather than later. But the mode right now is still very much. Let me get it to work.
- 支持 BofA 目标价 $225→$260
BofA raised PTs on $DDOG ($260 from $225) and $FROG ($85 from $72) following channel checks with Cognizant that pointed to improving demand trends across observability, security, and software supply-chain tooling. The firm said increasing AI-driven application complexity and enterprise modernization are driving stronger need for best-of-breed monitoring/security platforms, with Datadog benefiting from rising observability requirements and JFrog seeing healthy traction around software supply-chain/security workflows. BofA also believes both companies could continue delivering upside to Street estimates given improving execution and resilient infrastructure software spending trends.
来源背景：TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (渠道检查对象为 Cognizant)。二手转述，看不到底稿。 - 中性 此前买方仓位多偏好 DDOG 而非 SNOW
Just a good print all around with not much to nitpick at. Main fear had been 3 fold: deceleration, Cortex pressuring margins, and some concern around where SNOW fits in the AI picture. Earnings flipped the script on all of those in favor of the bulls. With the guide at 30%, if we assume average 3-4ppts beat, we get close to mid 30s growth next quarter on a 5ppts tougher comp. Then comps get easier, so hearing some bulls talking about mid/high 30s growth as we get into 2H’FY27. Positioning here had been on the lighter side with many favoring DDOG, and even MDB as of late, so the big move capturing some of that lighter sentiment this morning. We think path of least resistance likely higher here as SNOW still 15% below 2025 lows.
来源背景：TMT Breakout (TMTB) 日报，作者对 SNOW 财报的评论中顺带提及 DDOG 仓位。二手转述，看不到底稿。 - 支持 TMTB 认为 DDOG 与 SNOW 的 AI 消费叙事优于 MDB
Overall, we think this was a solid print, the stock’s likely path is up, and the set up going forward is decent to good, but think narrative around DDOG and SNOW better in the AI consumption space.
来源背景：TMT Breakout (TMTB) 日报，MDB 财报点评中顺带提及 DDOG。二手转述，看不到底稿。 - 中性 Q1 2026 收入 10.06 亿美元，+32%
**当前基准**:Q1 2026 营收 $1.006B(+32%),年化 ~$4B 9。
来源背景：Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。 - 支持 作者预测 DDOG 2030 年 175 亿美元，5 年 4.4 倍，年化 34%
**5 年增长**:**4.4x**,年化 **34%**。
来源背景：Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。 - 反对 CFO 承认消费型云软件会有投资期和优化期，增长不会直线
In terms of is it going to be a straight line, meaning on every month, is it going to have the exact same thing? We don't know. It's likely, if proven, you're going to have periods of investment, you may have periods of optimization. It's cloud software, it's consumption cloud software. So we can't predict exactly what the line's going to be in every moment, but we have – confident that it's very durable and it's very long because we're attached to huge market drivers. Now we also – if history repeats itself, we found that new technologies of which AI is happening is a huge accelerant of this conversion. And so, again, it's still too early, but if history repeats itself, that is likely to inflect the line upward faster to get to the same place 25, 50 years out.
- 中性 Wells Fargo：SNOW 对 DDOG 的估值差在缩小
Wells Fargo remains constructive on SNOW following Summit and Investor Day, arguing the company is emerging as one of the clearest enterprise AI beneficiaries. The firm highlighted strong early traction for CoCo and CoWork, with customer checks suggesting CoCo is accelerating migrations, improving developer productivity, and increasing workload consumption. Wells also pointed to continued strength among large customers, noting management believes spending by top accounts can continue expanding significantly over time. While adoption of Cortex AI Suite remains early, the firm sees growing evidence that AI tools are broadening Snowflake’s use cases and supporting sustained 25%+ revenue growth, improved profitability, and a narrowing valuation gap versus DDOG.
来源背景：TMT Breakout (TMTB) 日报，日程提示 + 转述 Wells Fargo 对 SNOW 的观点。二手转述，看不到底稿。 - 支持 Wedbush 目标价上调到 $260；DASH 发布 100+ 新品
Following DASH 2026, Wedbush came away more constructive on DDOG, highlighting 100+ product launches and a growing focus on autonomous operations across development, security, and observability. The firm believes Datadog is evolving beyond monitoring into a full-stack platform that helps enterprises manage increasingly complex AI-driven environments, with new products spanning AI observability, security, workflow automation, remediation, and developer productivity. Wedbush noted strong customer adoption of AI-focused offerings such as Bits AI, Agent Console, Journey Monitoring, and AI Guard, arguing these products expand Datadog’s role across the software lifecycle while creating additional share-gain opportunities.
来源背景：TMT Breakout (TMTB) 日报，转述 Wedbush (DASH 2026 后) 与 BofA (基础设施软件 'Fab Five') 报告摘要。二手转述，看不到底稿。 - 支持 BofA 把 DDOG 列为 2H26 首选基础设施软件 'Fab Five' 之一
BofA highlighted DDOG, FROG, MDB, SNOW, and TWLO as its preferred infrastructure software names for 2H26, arguing all five are benefiting from the combination of strong execution, AI tailwinds, differentiated products, and improving go-to-market effectiveness. The firm noted each company delivered better-than-expected results in 1Q, with AI increasingly becoming a growth accelerator rather than a future opportunity. BofA is particularly constructive on DDOG’s AI business, accelerating adoption at SNOW, AI-driven demand at FROG and MDB, and improving product growth at TWLO. The key takeaway is that AI spending is increasingly flowing into infrastructure software platforms, supporting continued beat-and-raise potential and further relative outperformance across the group
来源背景：TMT Breakout (TMTB) 日报，转述 Wedbush (DASH 2026 后) 与 BofA (基础设施软件 'Fab Five') 报告摘要。二手转述，看不到底稿。 - 支持 Truist 升级到 Buy、目标价 $300，认为街上低估消费增长的持久性
Truist upgraded DDOG to Buy and raised its PT to $300, arguing investor concerns around AI optimization are being overwhelmed by the urgency of enterprise AI adoption. Recent customer and partner conversations suggest Datadog remains in the early innings of AI-driven growth, with non-AI-native enterprises increasingly deploying agents that require more telemetry, monitoring, and observability. The firm also highlighted improving confidence in the durability of DDOG’s relationships with frontier AI labs, reducing a key overhang on the stock. While valuation has expanded, Truist believes Street still underestimates the durability of consumption growth and the size of the opportunity as AI adoption broadens across the enterprise.
来源背景：TMT Breakout (TMTB) 日报，转述 Truist 升级报告摘要 (基于客户与伙伴交流)。二手转述，看不到底稿。 - 支持 Citi 目标价上调到 $270
Following DASH ‘26 and customer checks, Citi raised its PT to $270, arguing Datadog’s competitive position continues to strengthen as enterprises scale AI and agentic workloads. Customer feedback was broadly positive on AI observability demand, with Bits AI and Cloud Cost Management receiving strong interest, while displacement opportunities against Splunk and New Relic remain robust and competitive pressure from open-source alternatives appears limited. Citi acknowledges investor concerns around tokenization, export controls, and lab-customer concentration, but believes these risks are overstated relative to Datadog’s expanding AI opportunity, growing enterprise adoption, and ability to capture incremental observability spend.
来源背景：TMT Breakout (TMTB) 日报，转述 Citi (DASH 后与客户检查) 报告摘要。二手转述，看不到底稿。 - 反对 Bernstein 预计 4Q 增速放缓约 500bp 到约 29%，低于买方 30%+
Bernstein raised its long-term estimates for Datadog, citing increasing confidence in the company’s AI platform opportunity, but downgraded the stock to Market Perform on valuation and elevated near-term expectations. The firm expects a strong 2Q, but believes growth could disappoint beginning in Q3/Q4 as ex-AI demand (~85% of revenue) flattens and AI lab growth starts to normalize. Bernstein expects Q4 revenue growth to slow roughly 500bps to ~29% y/y, well below investor expectations for 30%+ growth into next year and peak growth in the high-30% to 40%+ range.
来源背景：TMT Breakout (TMTB) 日报，转述 Bernstein 降级报告摘要。二手转述，看不到底稿。 - 中性 Bernstein 因估值和近端预期过高降级到 Market Perform
Bernstein raised its long-term estimates for Datadog, citing increasing confidence in the company’s AI platform opportunity, but downgraded the stock to Market Perform on valuation and elevated near-term expectations. The firm expects a strong 2Q, but believes growth could disappoint beginning in Q3/Q4 as ex-AI demand (~85% of revenue) flattens and AI lab growth starts to normalize. Bernstein expects Q4 revenue growth to slow roughly 500bps to ~29% y/y, well below investor expectations for 30%+ growth into next year and peak growth in the high-30% to 40%+ range.
来源背景：TMT Breakout (TMTB) 日报，转述 Bernstein 降级报告摘要。二手转述，看不到底稿。 - 中性 收入规模在 $4-5B，增长约 30%
So I think $3 billion revenue was the wrong estimate, right? Yes. Technically, we're between $4 billion and $5 billion now, growing about 30%. Well, pretty incredible. This company got started over 16 years ago, and we'll talk a little bit about what Olivier's learned about himself, but CEO, Co-Founder for 16 years, growing at 30%.
- 反对 客户 AI 成本飙升且算不清产出，CEO 称有大量优化空间 (未来可能波及观测支出)
Can you keep up that 30% growth? Yeah, I think the – so on our end, it's a big opportunity because everybody right now across the whole customer base, from the startups to the AI natives and with all of the big enterprises like the big banks and train companies and whatnot in the middle, everybody is busy adopting AI. And everybody is seeing their costs skyrocket and everybody is also finding that they are not in a position where they can really understand what's actually needed. Are they spending the right amount to get the right outcome? What are the outcomes, by the way? Like it's super hard. When you zoom in into specific parts of the business, you can see the outcomes, you can see value, you can see things. When you try to add everything up across the whole business, it's a lot more difficult, and I expect that to remain difficult for maybe a year or two, basically. But for us, that creates a huge opportunity because we can be there to help understand who's doing what with AI, what outcomes they're reaching, where they – how much they're spending to do that and how they can get better at it.
- 反对 BNP：DDOG 前瞻支出分数从 5.4 降到 2.0
BNP’s quarterly reseller survey found cybersecurity remained the top IT spending priority over the past three months, while Datadog sales tracked a few points above plan on product functionality, competitive wins and strength in infrastructure monitoring, APM and AI-related use cases. However, overall IT spending intentions weakened sharply from April, and resellers expect cybersecurity momentum to trail other categories over the next three months, suggesting demand may have been concentrated in 2Q. DDOG’s forward spending score also fell to 2.0 from 5.4, creating a more mixed setup despite solid recent execution.
来源背景：TMT Breakout (TMTB) 日报，转述 BNP Paribas 季度经销商调查与 UBS 客户检查 (7 位客户 + 3 位伙伴) 摘要。二手转述，看不到底稿。 - 中性 UBS 维持 2026 年 30%+ 增速，承认估值约 20x 收入、75x FCF
UBS maintains Buy after conversations with seven customers and three partners reinforced that observability remains a priority spend area as enterprises accelerate AI deployments. Checks improved versus late 2025/early 2026, with DDOG gaining share through vendor consolidation and expanding into hybrid environments, although pricing pushback remains the main friction point. UBS still sees 30%+ growth in 2026 and expects another acceleration in 2Q, but acknowledges the setup is tougher after the stock’s rally and valuation near 20x revenue and 75x FCF.
来源背景：TMT Breakout (TMTB) 日报，转述 BNP Paribas 季度经销商调查与 UBS 客户检查 (7 位客户 + 3 位伙伴) 摘要。二手转述，看不到底稿。 - 支持 MS 目标价 $225→$300，预期 2Q 中 30s (街 30.5%)、3Q 指引约 32%、FY26 至少 30%
Morgan Stanley reiterates Overweight and raises its price target to $300 from $225, saying stronger enterprise checks and robust public-cloud demand support mid-30% 2Q revenue growth versus Street near 30.5%. Morgan Stanley says AI-native demand is becoming more measurable, with one major LLM customer, two research labs and broader enterprise workloads driving consumption across observability, GPU monitoring, security and data tools. Morgan Stanley expects a roughly 32% 3Q guide and at least 30% FY26 growth, while Morgan Stanley raises FY26/FY27/FY28 revenue estimates by about 1%/4%/5%; Morgan Stanley acknowledges the stock is no longer cheap at roughly 16x CY27 sales but believes continued acceleration supports the premium.
来源背景：TMT Breakout (TMTB) 日报，转述 BTIG (伙伴检查) 与 Morgan Stanley (企业检查) 报告摘要。二手转述，看不到底稿。 - 中性 MS 承认约 16x CY27 销售不便宜
Morgan Stanley reiterates Overweight and raises its price target to $300 from $225, saying stronger enterprise checks and robust public-cloud demand support mid-30% 2Q revenue growth versus Street near 30.5%. Morgan Stanley says AI-native demand is becoming more measurable, with one major LLM customer, two research labs and broader enterprise workloads driving consumption across observability, GPU monitoring, security and data tools. Morgan Stanley expects a roughly 32% 3Q guide and at least 30% FY26 growth, while Morgan Stanley raises FY26/FY27/FY28 revenue estimates by about 1%/4%/5%; Morgan Stanley acknowledges the stock is no longer cheap at roughly 16x CY27 sales but believes continued acceleration supports the premium.
来源背景：TMT Breakout (TMTB) 日报，转述 BTIG (伙伴检查) 与 Morgan Stanley (企业检查) 报告摘要。二手转述，看不到底稿。 - 支持 BofA：AWS 收入分析显示预期保守，有望从 32% 再加速
BofA keeps DDOG as a top 2Q26 pick, saying its AWS revenue analysis suggests current estimates are conservative and leaves room for another acceleration from the 32% y/y growth reported in 1Q. BofA says additional frontier-lab wins, including two hyperscaler AI-research deals and five $10M+ AI customers, diversify the customer base and make renewal volatility less concerning. BofA also argues the recent OpenAI incident reinforces the need for agent observability as enterprises require greater visibility into autonomous AI behavior, expanding DDOG’s long-term opportunity.
来源背景：TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (AWS 收入分析)。二手转述，看不到底稿。 - 支持 专家 A 把 2H26 目标从 26–28.5% 大幅上调到 32–37%
* **The 2H26 target has been raised sharply from 26-28.5% to 32-37%**, making DDOG one of the few vendors for which he has taken targets up this year * Three reasons for the 2Q acceleration:
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 中性 股价自上次财报涨 83% (IGV +6%)，仓位拥挤，财报后涨幅难复制
* **We expect 3Q26 guidance around +30%, above consensus of +25% and matching buyside bogey; actual growth should hold near 35% or better** * **We expect FY26 guidance to be raised by roughly $100m or 300bps to about $4.42B/+29%**, an increase larger than the combined ~$80m from the 2Q beat and the above-consensus 3Q guide * On valuation: the shares are up 83% since the last print (versus IGV +6%), and positioning is crowded, so while results are likely to stay strong, the post-print share reaction will be hard to replicate at last quarter’s level
来源背景：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 - 中性 Wells Fargo 调查比 1Q 略软，部分因受访者含更多小客户
Wells Fargo says its 2Q survey, representing roughly $95M of annual Datadog spend, showed most large customers expect to increase spending in 2026 and continue shifting share toward DDOG from other observability vendors. Wells Fargo notes the survey was modestly softer than 1Q, partly because the latest respondent mix included smaller customers with narrower platform usage, but still views demand as constructive into 2H26. Wells Fargo says the results reinforce Datadog’s durable share-gain story and continued enterprise expansion.
来源背景：TMT Breakout (TMTB) 日报，转述 FundaAI 预览 (同 8/3 Funda 原文) 和 Wells Fargo 2Q 客户调查 (约 $95M 年 DDOG 支出) 摘要。二手转述，看不到底稿。 - 中性 指引方法不变：近几个月趋势打折扣；最大客户减量已计入
After taking into consideration capital expenditures and capitalized software, free cash flow was $279 million for a free cash flow margin of 25%. And now for our outlook for the third quarter and the fiscal year 2026. Our guidance philosophy overall remains unchanged. As a reminder, we base our guidance on trends observed in recent months and imply conservatism on these growth trends. Regarding our largest customer, we have seen a usage reduction, which is incorporated in our Q3 and full year 2026 guidance. As Olivier noted, this customer has recently renewed with us.
- 口径不齐 Q3 指引 $1.135–1.145B (+28–29%)，较 Q2 实际 $1.12B 仅环比 +$15–25M (按公司数字推算)
Regarding our largest customer, we have seen a usage reduction, which is incorporated in our Q3 and full year 2026 guidance. As Olivier noted, this customer has recently renewed with us. For the third quarter, we expect our revenue to be in the range of $1.135 billion to $1.145 billion, which represents a 28% to 29% year-over-year growth. Non-GAAP operating income is expected to be in the range of $260 million to $270 million, which implies an operating margin of 23% to 24%. And non-GAAP net income per share is expected to be in the $0.63 to $0.65 per share range based on approximately 378 million weighted average diluted shares outstanding.
- 支持 FY26 指引上调到 $4.45–4.47B (+30%)
Non-GAAP operating income is expected to be in the range of $260 million to $270 million, which implies an operating margin of 23% to 24%. And non-GAAP net income per share is expected to be in the $0.63 to $0.65 per share range based on approximately 378 million weighted average diluted shares outstanding. For the full fiscal year 2026, we expect revenue to be in the range of $4.45 billion to $4.47 billion, which represents a 30% year-over-year growth. Non-GAAP operating income is expected to be in the range of $1.01 billion to $1.03 billion, which implies an operating margin of 23% and non-GAAP net income per share is expected to be in the range of $2.50 to $2.54 per share based on approximately 376 million average diluted shares outstanding. And for some additional notes on guidance, we expect net interest and other income for the fiscal year 2026 to be approximately $180 million.
- 支持 这次对最大客户 '完全去风险'，不希望它成为加速中的业务的包袱
Yeah. (00:31:39) I say is. In this case, we did choose to fully de-risk our largest customer. And the reason for that is we don't want that to be an overhang on what is otherwise a business that is accelerating and performing extremely well. So we have the same overall conservatism as we always do when we look at our numbers, but in this case we also weighted (00:32:04) this one a little bit differently. Okay, perfect.
- 支持 Q2 环比 +11% 是 2022Q2 以来最高，环比增量 $115M 大幅创纪录
Thanks, Olivier. Our Q2 revenue was $1.12 billion, up 36% year-over-year. Within that, our 11% quarter-over-quarter revenue growth is the highest since Q2, 2022. And our quarter-over-quarter revenue added of $115 million is a record by a significant margin. We continue to see robust usage growth from existing customers, as well as a strong ramp in our new customers. Revenue growth accelerated with our broad base of customers, excluding AI customers to the high-20s year-over-year, up from the mid-20s percent last quarter and 18% in the year ago quarter.
- 反对 RPO $3.47B 环比基本持平 (Q1 $3.48B)，同比 +43% (Q1 +51%)；cRPO 约 +40% (Q1 mid-40s)，前瞻合同指标放缓
Now, moving on to our financial results. Billings were $1.18 billion, up 38% year-over-year. Remaining performance obligations, or RPO, was $3.47 billion, up 43% year-over-year. Current RPO grew about 40% year-over-year and RPO duration increased year-over-year. As we previously mentioned, we continue to believe revenue is a better indication of our business trends than billing and RPO. Now let's review some of the key income statement results.
- 中性 2Q 收入 +35.6% 符合买方 35–36%，beat $46m 小于上季 $50m
* 26Q2 revenue grew 35.6% YoY, in line with buy-side expectations of 35-36%, accelerating 3ppts QoQ. However, the US$46m beat was slightly below the US$50m last quarter. * 26Q3 revenue guidance was a touch light at 29.5% at the high end, where expectations were looking for a 3-handle. Applying a typical 5ppts beat would imply a marginal sequential growth deceleration in 26Q3. * **The miss appears mostly driven by usage optimization at OpenAI, DDOG’s largest customer. For 26Q2, we estimated a 20% QoQ decline in OpenAI’s spending, on which market consensus appeared too optimistic. Management confirmed it has renewed a 9-figure contract with OpenAI, but the latter’s usage will decline sequentially in 26Q3. We believe Chronosphere is gaining share within the OpenAI platform, as it prices 30-40% lower than DDOG.** * Core/non-AI usage growth continued to accelerate to high-20s from mid-20s, driven by strong cross-sell momentum and increasing usage from cloud migration and AI app development. * Full-year revenue guidance was raised by more than the 26Q2 beat flowed through, to 30-31% YoY, in line with expectations. * The margin trend remains strong — GPM dipped below 80%, but OPM beat expectations on operating leverage.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 反对 3Q 指引上限 29.5%，买方要 3 字头；按 5pp beat 意味环比小幅减速
* 26Q2 revenue grew 35.6% YoY, in line with buy-side expectations of 35-36%, accelerating 3ppts QoQ. However, the US$46m beat was slightly below the US$50m last quarter. * 26Q3 revenue guidance was a touch light at 29.5% at the high end, where expectations were looking for a 3-handle. Applying a typical 5ppts beat would imply a marginal sequential growth deceleration in 26Q3. * **The miss appears mostly driven by usage optimization at OpenAI, DDOG’s largest customer. For 26Q2, we estimated a 20% QoQ decline in OpenAI’s spending, on which market consensus appeared too optimistic. Management confirmed it has renewed a 9-figure contract with OpenAI, but the latter’s usage will decline sequentially in 26Q3. We believe Chronosphere is gaining share within the OpenAI platform, as it prices 30-40% lower than DDOG.** * Core/non-AI usage growth continued to accelerate to high-20s from mid-20s, driven by strong cross-sell momentum and increasing usage from cloud migration and AI app development. * Full-year revenue guidance was raised by more than the 26Q2 beat flowed through, to 30-31% YoY, in line with expectations. * The margin trend remains strong — GPM dipped below 80%, but OPM beat expectations on operating leverage.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 中性 FY26 指引上调幅度大于 2Q beat 的传导，到 30–31%，符合预期
* 26Q2 revenue grew 35.6% YoY, in line with buy-side expectations of 35-36%, accelerating 3ppts QoQ. However, the US$46m beat was slightly below the US$50m last quarter. * 26Q3 revenue guidance was a touch light at 29.5% at the high end, where expectations were looking for a 3-handle. Applying a typical 5ppts beat would imply a marginal sequential growth deceleration in 26Q3. * **The miss appears mostly driven by usage optimization at OpenAI, DDOG’s largest customer. For 26Q2, we estimated a 20% QoQ decline in OpenAI’s spending, on which market consensus appeared too optimistic. Management confirmed it has renewed a 9-figure contract with OpenAI, but the latter’s usage will decline sequentially in 26Q3. We believe Chronosphere is gaining share within the OpenAI platform, as it prices 30-40% lower than DDOG.** * Core/non-AI usage growth continued to accelerate to high-20s from mid-20s, driven by strong cross-sell momentum and increasing usage from cloud migration and AI app development. * Full-year revenue guidance was raised by more than the 26Q2 beat flowed through, to 30-31% YoY, in line with expectations. * The margin trend remains strong — GPM dipped below 80%, but OPM beat expectations on operating leverage.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 反对 3Q 指引高于街但低于买方约 31% bogey，暗示 4Q 更明显减速
Revenue growth accelerated to 36% and margins beat, while the FY26 raise exceeded the quarterly beat pass-through. The pushback is the F3Q guide, which beat Street but landed below the ~31% buyside bogey and implies a sharper deceleration into F4Q. Given the crowded long positioning, some reset was understandable but lots questioning the 20% move down. Call
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 2Q26 公布数字、街上与买方预期，并评论股价反应。二手转述，看不到底稿。 - 中性 仓位拥挤，股价回调可以理解，但很多人质疑跌 20% 过度
Revenue growth accelerated to 36% and margins beat, while the FY26 raise exceeded the quarterly beat pass-through. The pushback is the F3Q guide, which beat Street but landed below the ~31% buyside bogey and implies a sharper deceleration into F4Q. Given the crowded long positioning, some reset was understandable but lots questioning the 20% move down. Call
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 2Q26 公布数字、街上与买方预期，并评论股价反应。二手转述，看不到底稿。 - 支持 2Q 收入 $1.121B (+36%)，符合买方 35–36%
* F2Q revenue $1.121B, +36% y/y (last q +32%) vs. Street $1.08B and buyside 35–36%
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 2Q26 公布数字、街上与买方预期，并评论股价反应。二手转述，看不到底稿。 - 中性 3Q 指引 $1.135–1.145B (+29%)，街 $1.106B，买方约 31%
* F3Q revenue guide $1.135–1.145B, +29% y/y at midpoint vs. Street $1.106B and buyside ~31% * FY26 revenue guide $4.45–4.47B, +30% y/y vs. Street ~$4.35B and prior $4.30–4.34B
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 2Q26 公布数字、街上与买方预期，并评论股价反应。二手转述，看不到底稿。 - 中性 DDOG 没做错什么但仍下跌，原因是预期满、仓位拥挤
Software was the axis. PLTR grew 93% YoY with US Commercial up 149% to $764mn, which is a direct rebuttal to the pre-earnings sell-side channel checks arguing US Commercial was rolling over; management then raised FY26 by $496mn, and only a quarter of that is the beat flowing through. NET accelerated to 36% and, more importantly, raised the full year by more than the beat — about $26mn of genuine second-half revision, with DBNRR back to 120%. DDOG did nothing wrong: 35.6% growth, core non-AI accelerating to high-20s, a renewed nine-figure OpenAI contract. It fell anyway, because expectations were full and positioning was crowded into the print, exactly the risk-reward our preview flagged. Fundamentals and setup are two different trades. The other pair worth sitting with is AppLovin and Unity: APP’s gaming revenue grew just 2% QoQ ex-commerce, the slowest in years, while Unity Ads put up 23% QoQ on the D28 rollout. We do not think that is a coincidence.
来源背景：FundaAI 周报：DDOG 段落为 2Q review 摘要 + Funda 当周两次专家访谈的一句话摘要 (某全球云通信平台的工程高管——疑似 Twilio，因标题标 TWLO；某全球金融信息服务商经理)。二手转述，看不到底稿，访谈全文在 Funda 平台未附。 - 中性 3Q 指引偏轻只因 OpenAI 优化
**DDOG 26Q2 — unchanged fundamentals met with too-high expectations.** Our preview said the print would be solid but that full 3Q expectations and crowded positioning left poor risk-reward into the event; the quarter grew 35.6% with core non-AI accelerating to high-20s, and the guide came light only because of OpenAI optimization.
来源背景：FundaAI 周报：DDOG 段落为 2Q review 摘要 + Funda 当周两次专家访谈的一句话摘要 (某全球云通信平台的工程高管——疑似 Twilio，因标题标 TWLO；某全球金融信息服务商经理)。二手转述，看不到底稿，访谈全文在 Funda 平台未附。 - 中性 DDOG 财报下跌后反弹 +11%，有逢低买入的迹象
Good afternoon. **QQQs -30bps** as market volumes continue to slow heading into the middle of August post-earnings, when Bloomberg terminals start feeling more like expensive screensavers. Price action post-earnings continues to be a lot better in internet & software. Investors there are still showing an appetite for buying names up post-earnings, especially where numbers accelerated and the narrative improved, even if just marginally. **PLTR +2%; TEAM +2%; TWLO +3.5%; U +25bps; UBER +4%; SAP +1%; MSFT +1%; ABNB +4%; TTWO +2%** are a few of the names that printed solid earnings and have follow-through. This shift in flows out of semis and into hyperscalers started a couple weeks and has now expanded. We even started to see some willingness to buy the dip on misses, with **DDOG +11%** and **FIG +9%** today while something like **AKAM +6.5%** didn’t stay down long.
来源背景：TMT Breakout (TMTB) 日报，市场评论。二手转述，看不到底稿。 - 支持 3Q 指引刻意假设最大 AI 客户只按合同最低额贡献
Barclays says CEO/CFO meetings reinforced that DDOG’s acceleration is broader than AI-native customers, with non-AI growth improving from roughly 18% a year ago to the high-20% range as AI drives cloud migration, infrastructure modernization and new application development. Barclays notes DDOG intentionally de-risked 3Q guidance by assuming its largest AI customer contributes only at contractual minimums, while broader usage and pipeline trends remain strong. Barclays also says GPU monitoring, agent observability, Bits AI, MCP and AI security are gaining early traction, but remain in the early innings of monetization and represent a multi-year growth opportunity.
来源背景：TMT Breakout (TMTB) 日报，转述 Barclays 与 DDOG CEO/CFO 财报后 fireside 会面的报告摘要。二手转述，看不到底稿。 - 中性 合同：take-or-pay、1-3 年、有最低承诺，客户不能少于承诺
And then, how do you guide (00:09:44) manage the Street in that way? No question. So, we have take or pay contracts and they tend to be a year or more. They tend to be out to three years and then we have a base of commitment so they can't spend less than that. And then for the most part, this is really about – it's the same thing with AWS and hyperscalers. There's essentially capacity planning we do together.
- 支持 指引方法：把长期增长率/NRR 打折留缓冲，客户实际花得比打折后多，所以能 beat and raise
But we step in and we help clients then use it effectively. We help them figure out what logs to put in and what logs not. And then in terms of guidance, what we're able to do and that's how we can beat and raise is we take that growth rate that we see over a long period of time and we discount it, so that we have that cushion. And then, what's been proven out in the beat and raise is that the clients spend more than the amount we're discounting. So, let's say net retention is blank.
- 支持 Datadog 文化保守，指引不会画大饼
**EE** [00:42:33]: I think it's a bit tricky. Datadog, again, like I said, they're pretty conservative in nature and in culture. They would never try and depict the future where it's going to be much higher than what they expect. They're usually on the conservative side of things. I think usage, it depends because, again, there are a lot of commitments and drawdown type of contracts. I think that they know their customer base. They know the trends, they are on top of that. I think, again, things like usually logs could be one of those things where people can easily send less or more and sometimes traces as well. Again, it's not like I have any proof or a data point, but they see and they have tools to predict and see how these things are trending or know ahead of time. Obviously, they need to report only once per quarter, not necessarily in the middle of the quarter or every month on these things. I think more or less, they're probably on the right direction. Again, if you're saying suddenly, there's going to be a financial crisis and companies will have to reduce budgets and pay less for cloud providers and OpenAI and those kind of things, then, yes, obviously, it will impact. They will also look into observability and start thing, "Okay, where can we save more or maybe pay less for a given amount of time," and we've seen that in previous waves as well.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 支持 若无系统性危机，未来 12 个月增速维持约 30%
**EE** [00:45:13]: I don't know about the number itself. I think, again, it's pretty high, I think, for Datadog at their stage. It's a positive signal, in my opinion. It was already around 25-29% in previous quarters. It's great to see that it's above 30%. I'm assuming that it's driven by companies building more applications, deploying AI also to production. This whole trend that we're seeing is, obviously… Technology is the key thing behind it, and that's what Datadog is monitoring and providing observability for. For the next 12 months, again, if there is no major financial crisis or something in the economy that's bigger or more systemic, then I definitely see them continuing around that 30%. What's going to drive? Again, I think a lot of expansions still. Again, it's a flywheel. We haven't talked too much about the Datadog go-to-market. The land and expand motion that they put together from the early days just proved itself and now they're seeing the success of it, and it became a flywheel. Just think about how many customers with only two products or four products they have today and where they can get to.
来源背景：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 - 支持 BNP：DDOG 7 月招聘加速，销售岗位空缺增加
BNP Paribas says DDOG and CDNS hiring accelerated in July, while sales job openings also increased versus six months ago, a potentially positive leading indicator for future top-line growth. Rep quota attainment improved across much of software, particularly PANW, CRWD, ZS, WK and NAVN, although DDOG remains among the names with the lowest percentage of reps hitting quota. CDNS stands out more positively, with both accelerating hiring and relatively strong quota attainment, while falling headcount at SNPS, ADSK and PTC could support margins.
来源背景：TMT Breakout (TMTB) 日报，转述 BNP Paribas 招聘/销售生产力数据摘要。二手转述，看不到底稿。 - 反对 DDOG 是完成配额销售比例最低的公司之一
BNP Paribas says DDOG and CDNS hiring accelerated in July, while sales job openings also increased versus six months ago, a potentially positive leading indicator for future top-line growth. Rep quota attainment improved across much of software, particularly PANW, CRWD, ZS, WK and NAVN, although DDOG remains among the names with the lowest percentage of reps hitting quota. CDNS stands out more positively, with both accelerating hiring and relatively strong quota attainment, while falling headcount at SNPS, ADSK and PTC could support margins.
来源背景：TMT Breakout (TMTB) 日报，转述 BNP Paribas 招聘/销售生产力数据摘要。二手转述，看不到底稿。 - 中性 Datadog TTM EV/Sales 约 20.5 倍 (8/28 收盘)
截至 2026 年 8 月 28 日收盘，九家公司 TTM EV/Sales 的分布已经非常极端：Cloudflare 约 42.2 倍、CrowdStrike 约 40.7 倍、Palo Alto Networks 约 28.5 倍、Snowflake 约 22.3 倍、Datadog 约 20.5 倍；另一端，Okta 约 8.7 倍、Zscaler 约 8.9 倍、Elastic 约 5.4 倍、HubSpot 约 3.5 倍。Salesforce 作为对照约
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 - 反对 高兑现、高定价，'Agent 很重要'已不足以支持继续上调
**第一类是高兑现、高定价。** CrowdStrike、Palo Alto Networks、Datadog 已经展示增长、ARR 和现金流，市场也给出了近乎无瑕疵的增长要求。它们更适合跟踪财报兑现与估值消化，“Agent 安全很重要”已经不足以支持继续上调空间。
来源背景：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 - 中性 CEO：观测是 AI 时代“最后剩下的软件类别”，市场最终是 Datadog 加一两家
I think it's also pretty clear that the question is not just – is not who is going – whether we're going to be there, the question is, who's going to be there with us in the end? Like it's a market, great market, great opportunity. It's going to be us, maybe one or two other companies. So, with that in mind, I think it would be foolish not to invest in Datadog, that's my take. So...
- 支持 与 2021-23 优化期不同：当时增长 70%、ZIRP、云原生占比高；现在更分散，并有合同延期、量价折扣等缓冲
We also have built out a number of functions to provide information transparency to our customers, work with them, account managers, SKU and non-SKU types of arrangements where we help our clients use it, use the product in the right way. So, I think we've gotten a lot better in helping our clients. We also have things with contract extension, volume pricing, a number of things to ameliorate the risk. So, I think you're right on, that the end market has learned. We are much more diversified, and we've learned, I think, how to be better in helping our clients to have long-term growth with us.
- 支持 Wedbush 接手覆盖给 Outperform，目标价 $275
Wedbush says DDOG is accelerating growth while maintaining strong margins and remains the best-positioned pure-play beneficiary of rising cloud- and AI-native workloads. Wedbush sees observability demand strengthening as AI infrastructure and applications increase system complexity and monitoring requirements. Wedbush assumes Outperform and raises PT to $275 from $260.
来源背景：TMT Breakout (TMTB) 日报，转述 Wedbush 报告摘要。二手转述，看不到底稿。 - 中性 Citi：DT 相对 DDOG 仍有明显估值折价
Citi says GTM execution is improving into an FY27 renewal base up ~50% y/y, with Logs, strategic accounts and AI monetization still offering incremental upside largely absent from guidance. Citi sees Arize accelerating DT’s AI observability roadmap by ~18 months and says FY27 NNARR could exceed $370M, ~34% y/y, while organic growth holds near ~14%. Citi reiterates Buy, arguing DT still trades at a sizable discount to DDOG despite improving ARR momentum and potential Investor Day catalysts.
来源背景：TMT Breakout (TMTB) 日报，转述 Bernstein (AWS 网站流量数据) 与 Citi (DT) 报告摘要。二手转述，看不到底稿。 - 支持 IR：4Q 指引按最大 AI 原生客户合同最低用量，3Q 跟踪高于该底线
Truist says meetings with DDOG IR reinforced that guidance remains conservative around its largest AI-native customer, with 4Q assumptions based on contractual minimum usage even though 3Q is tracking above that floor. Truist says IR described growth as broad-based across products, cohorts and verticals, with non-AI-native customers also accelerating as “getting-ready-for-AI” investments move into production. Truist also sees security as an underappreciated growth leg after crossing $100M ARR, with SIEM leading adoption and security still <2% of spend among >$1M ARR customers, leaving substantial expansion runway.
来源背景：TMT Breakout (TMTB) 日报，转述 Truist 与 DDOG IR 会面后的报告摘要。二手转述，看不到底稿。 - 支持 预计上市观测厂商收入和利润继续增长，因为 TAM 和问题集都在扩大
Bernd Harzog: Well, so first of all, as a just nature of my business, right, I don't have visibility into any form of unannounced business performance, and I generally don't work with the sales teams at the vendors where I work with vendors, okay? I work with the product team, okay? And this is by design. I've designed my business not to have access to what you would consider as unannounced or, you know, material non-public information, right? I just want to make sure that I never cross that bridge, right? So with that caveat in place, what I expect is I expect revenue and profits from specifically data, observability, and tracing to publicly traded vendors in this space to continue to grow. And the reason is because the market that they are addressing is growing quite rapidly. The problems that the customers want them to solve are growing quite rapidly. And so you have an expanding TAM and an expanding problem set, okay? Conditions are ripe for you, therefore, to grow more rapidly.
来源背景：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 - 支持 Wolfe 上行情景 FY25–28 收入 CAGR 约 32%，共识约 25.5%；目标价 $340
Wolfe says ecosystem checks show agents materially increasing observability intensity as every tool call, retry and handoff needs to be traced, with traces costing roughly 2x metrics and already driving higher consumption. AI coding tools are also accelerating software-development velocity by ~2–3x, creating more code/services/pipelines for DDOG to monitor, while Wolfe’s initial conversations suggest an increasing likelihood of DDOG becoming directly involved with Meta’s Muse. Wolfe’s upside case now implies ~32% FY25–28 revenue CAGR vs. ~25.5% consensus and sees DDOG as one of the best-positioned ways to play inference/agentic demand.
来源背景：TMT Breakout (TMTB) 日报，转述 Wolfe Research 生态检查报告摘要。二手转述，看不到底稿。
E07 股东经济学
判断：差，这是 DDOG 和 mapping 里其他赢家共同的问题，而且 DDOG 没有在改善。
过去 12 个月 (2025 Q3–2026 Q2)：收入 $39.7 亿，FCF $10.7 亿 (27%)，SBC $8.2 亿 (21%)，扣掉 SBC 的 FCF 只有 $2.5 亿 (6%)。没有回购。稀释股数从约 359M 涨到 Q3 指引的 378M，两年约 +5%。毛利率框架"80% 上下"不变，但最近三个季度都在 80% 以下或附近。FY26 营业利润率指引一开始只有 21% (低于此前的 mid-20s)，CFO 的说法是先按保守收入做投入计划，超出部分"一部分"流到利润率；R&D 一直约收入的 30%，构成在从人转向 token 和 GPU。
公司治理：同股不同权，6 月股东大会上提案人说内部人只持有约 7% 的股份，却控制约 70% 的投票权，取消绝对多数投票要求的提案没有通过。
所以从股东的角度，收入加速并没有变成每股现金流的加速：SBC 相当于 FCF 的约 77%。
Yes, the framework is we try to plan with more conservative revenues, understanding that if the revenues exceed above the targets that we give, it's difficult in the short term to invest incrementally. So, what we're trying to do is invest first in the revenue growth and then layer in additional investment as we see -- if we see excess of target. So generally, it reflects, one, the continued investment, which we think is paying off, both in terms of the platform and R&D as well as in -- and including AI as a go-to-market. And then as we've seen over the years in our beat and raise, we've tended to have some of that flow through into the margin line and then re-up again for the next phase of growth.
As Olivier noted, this customer has recently renewed with us. For the third quarter, we expect our revenue to be in the range of $1.135 billion to $1.145 billion, which represents a 28% to 29% year-over-year growth. Non-GAAP operating income is expected to be in the range of $260 million to $270 million, which implies an operating margin of 23% to 24%. And non-GAAP net income per share is expected to be in the $0.63 to $0.65 per share range based on approximately 378 million weighted average diluted shares outstanding. For the full fiscal year 2026, we expect revenue to be in the range of $4.45 billion to $4.47 billion, which represents a 30% year-over-year growth. Non-GAAP operating income is expected to be in the range of $1.01 billion to $1.03 billion, which implies an operating margin of 23% and non-GAAP net income per share is expected to be in the range of $2.50 to $2.54 per share based on approximately 376 million average diluted shares outstanding. And for some additional notes on guidance, we expect net interest and other income for the fiscal year 2026 to be approximately $180 million.
And we're not at the point where we saturated the market. We still have many areas, I can spend a lot of time on that. So, that we're going to continue to invest in. When it comes to R&D, I think we're going to continue to invest at a high rate, but it's very likely and it is happening that the percentage of that, the allocation of that is going to move more towards tokens than it is towards humans. We're playing around with it. But you asked about margins. You can see our margins haven't changed, right? But we are also – we're not tokenmaxxing. We're basically using tokens in order to create good software and products to release to clients. But you can see it's more of a distribution than a margin erosion, and that's what we see right now.
So those are willing to pay a premium for shares of companies that have excellent corporate governance. The supermajority voting requirements, like those of Datadog's have been found to be one of six entrenching mechanisms that are negatively related to company performance according to what matters in corporate governance, by (00:04:17) of a Harvard Law School. Supermajority requirements can be used to block proposals supported by most shareholders, but opposed by management. Unfortunately, insiders own only 7% of Datadog's stock yet control 70% of the voting power. This is a sign that financial decisions can be made without having to bear anywhere near the full cost of potential disastrous results. This proposal topic won from 74% to 88% supported warehouser waste management, Goldman Sachs, FirstEnergy and Macy's. These votes would have been higher than 74% to 88% support if more shareholders had access did an independent proxy voting advice.
展开本节全部 25 条原文 (支持 3 / 反对 4 / 其他 18，按时间)
- 中性 Q3 FCF $214M (24% 利润率)，营业利润率 23% (上季 20%，去年同期 25%)
Our Q3 OpEx grew 30% -- 32% excuse me, year-over-year, down from 36% last quarter. We continue to grow our investments to pursue our long-term growth opportunities, and this OpEx growth is an indication of our execution on our hiring plan. Q3 operating income was $207 million for a 23% operating margin compared to 20% last quarter and 25% in the year ago quarter. And now turning to our balance sheet and cash flow statements. We ended the quarter with $4.1 billion in cash, cash equivalents and marketable securities and cash flow from operations was $251 billion in the quarter.
- 中性 毛利率框架：长期 80% 上下，不改
Yes, there's a couple of things. I think we've mentioned that we've been focused and have focused over the many years on the efficiency of our cloud platform. We have significant engineering efforts around cost of sales and delivery of value. And so we've been able to deliver on that. We also have a very broad customer base distributed in terms of volume. So as customers get larger and maybe get volume discounts, we have a number -- a lot of customers coming in, it's smaller, so that balance there. And then in terms of the sort of the future -- I'll repeat what we've always said that we've been running the company with a gross margin plus or minus 80%, we've given that range and not changed it, and we watch it. And it gives us signals in terms of efficiency, how we're operating, it gives us signals in pricing and things like that and I wouldn't change the comments we made over the many years about looking at that and then developing operations and strategies around that.
- 中性 FY26 营业利润率指引 21%，低于此前 mid-20s；CFO 解释为先投入、超额收入再部分流入利润率
Yes, the framework is we try to plan with more conservative revenues, understanding that if the revenues exceed above the targets that we give, it's difficult in the short term to invest incrementally. So, what we're trying to do is invest first in the revenue growth and then layer in additional investment as we see -- if we see excess of target. So generally, it reflects, one, the continued investment, which we think is paying off, both in terms of the platform and R&D as well as in -- and including AI as a go-to-market. And then as we've seen over the years in our beat and raise, we've tended to have some of that flow through into the margin line and then re-up again for the next phase of growth.
- 中性 毛利率目标 80% 上下不变
Yes. On the gross margin, I think what we said is plus or minus the 80% mark. We try to engineer when we see opportunities for efficiency, we've been quite good at being able to harvest them. At the same time, we want to make sure we're investing in the platform. So, I think what we're essentially -- where we are today is very much sort of in line with what we said we're targeting. There may be opportunities longer term, but we also are trying to balance those opportunities with investment in the platform. And in terms of AI, to date, we are using it in our internal operations. So far, it's -- with the first signs of what we're seeing is productivity and adoption. We will continue to update everybody as we see opportunities in terms of the cost structure. Oli, anything else you want to go over?
- 中性 Q4 FCF $291M (31%)，营业利润率 24%
Q4 OpEx grew 29% year-over-year versus 32% last quarter and 30% in the year ago quarter. And we continue to grow our investments to pursue our long-term growth opportunities, and this OpEx growth is an indication of our successful execution on our hiring plans. Our Q4 operating income was $230 million for a 24% operating margin compared to 23% last quarter and 24% in the year ago quarter. Now turning to the balance sheet and cash flow statements. We ended the quarter with $4.47 billion in cash, cash equivalents and marketable securities.
- 反对 FY26 adj OPM 指引 20.6–21.5%，同比降 100–200bps，因为在加大 AI 投入和扩张销售团队
* 然而，FY26 adj OPM指引为20.6-21.5%，**同比下降100-200bps，低于预期；管理层指出仍在扩大对AI的投入以及销售团队将持续扩张**
来源背景：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 - 中性 FCF $289M，FCF 率 29%
Turning to the balance sheet and cash flow statements. We ended the quarter with $4.8 billion in cash, cash equivalents and marketable securities. Our cash flow from operations was $335 million in the quarter, after taking into consideration capital expenditures and capitalized software, free cash flow was $289 million and free cash flow margin was 29%. And now for our outlook for the second quarter and for the fiscal year 2026. First, our guidance philosophy overall remains unchanged.
- 反对 费用增速 31% (快于上季 29%)，经营利润率 22% (上季 24%)，在继续加大投入
This compares to a gross margin of 81.4% last quarter and 80.3% in the year-ago quarter. As we've discussed in the past, our gross margin varies from quarter-to-quarter with investments into innovations for our customers offset by efficiency efforts. Our Q1 OpEx grew 31% year-over-year versus 29% last quarter and 29% in the year-ago quarter. As a reminder, we continue to grow our investments to pursue our long-term growth opportunities. And this OpEx growth is an indication of our execution of our hiring plans. Q1 operating income was $223 million or a 22% operating margin, compared to 24% last quarter and 22% in the year-ago quarter. Turning to the balance sheet and cash flow statements. We ended the quarter with $4.8 billion in cash, cash equivalents and marketable securities.
- 中性 工作负载跑在公有云，成本进 OpEx 而非 CapEx；自训模型投入加大但暂不影响数字
Thank you. Yeah. I mean, look, the sort of the investments we're making right now, we run most of our workloads on cloud, meaning you'll see all of that in OpEx, not in CapEx. So, we have low CapEx. If it changes, we'll tell you, like if for some reason we decide to make different kinds of investments and some of it more upfront and more CapEx, we'll tell you, but that's not the case today. We are definitely ramping up our investments in particular in R&D and in the scale of the models, we train ourselves on things like that. Right now, there's nothing that you can actually see in the numbers that moves any needle. But, if that changes also we'll tell you. We don't expect any change to our model.
- 支持 剔除 6 月 DASH 一次性费用，2Q 经营利润率指引同比扩 4pp；全年上调 1pp 至持平
* Excluding one-off DASH conference spending in June, 2Q adj OPM is guided to a 4ppt YoY expansion. Full year Adj OPM is also raised by 1ppt to be flat YoY.
来源背景：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 - 中性 R&D 约占收入 30%，总盘子不变但构成从人力转向 GPU/收购
Historically, that's been mostly about human labor and how many people are on the team. Now, maybe there's a bit of a mix shift. Maybe we'll have the same overall R&D investment, which right now is around 30% of the top line. But some more of it will go to tuck-ins or GPUs and less of it will go to labor. I think it's unclear, but we know we'll keep investing, and we also know that right now also we keep hiring. We think we can still scale.
- 中性 研发约占收入 30%，未来更多花在 token/GPU、更少花在人力
Maybe you will have the same overall R&D investment, which right now is around 30% of the top line, but some more of it will go to tokens or GPUs and less of it will go to labor.
来源背景：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 - 中性 R&D 内部在做 token vs 人力的 A/B 实验，构成会转向 token
No, constrain the people, let's see how fast you get. And we're doing – we're pretty nerdy engineering wonky so we have all this stuff going on and we're trying to see what's going to happen. We believe that the componentry of R&D will shift somewhat. We don't know exactly to – from – to tokens from people, but we don't know the exact amount. Have you been able to use AI internally with all that development that you have to help out with your own homegrown solutions in sales and marketing and G&A at all? So we have – AI is there's four – there's three or four major use cases. One, in the product itself, we talked about... Yeah. ...the models and the products, so those would be components of the product. Two is accelerate product release through software coding tools.
- 中性 双重股权结构：提案人称内部人持股约 7%、控制约 70% 投票权；简单多数提案未通过
The supermajority voting requirements, like those of Datadog's have been found to be one of six entrenching mechanisms that are negatively related to company performance according to what matters in corporate governance, by (00:04:17) of a Harvard Law School. Supermajority requirements can be used to block proposals supported by most shareholders, but opposed by management. Unfortunately, insiders own only 7% of Datadog's stock yet control 70% of the voting power. This is a sign that financial decisions can be made without having to bear anywhere near the full cost of potential disastrous results. This proposal topic won from 74% to 88% supported warehouser waste management, Goldman Sachs, FirstEnergy and Macy's.
- 中性 Adaptive ML 收购未披露财务条款
BofA views Datadog’s acquisition of Adaptive ML as a strategic addition that accelerates its AI roadmap by adding reinforcement learning capabilities to support proprietary AI agents and observability products. The firm believes the deal enhances Datadog’s ability to build differentiated AI offerings such as Bits SRE, Bits Code, and Bits Security Analyst, making it incrementally more confident the company can meet or exceed near-term targets while strengthening its long-term growth trajectory. No financial terms were disclosed.
来源背景：TMT Breakout (TMTB) 日报，转述 BofA (DDOG) 与 BTIG/BNP (PANW，涉及 Chronosphere) 报告摘要。二手转述，看不到底稿。 - 中性 FCF $279M，FCF 率 25% (Q1 29%)
We ended the quarter with $5 billion in cash, cash equivalents and marketable securities. Cash flow from operations was $316 million in the quarter. After taking into consideration capital expenditures and capitalized software, free cash flow was $279 million for a free cash flow margin of 25%. And now for our outlook for the third quarter and the fiscal year 2026. Our guidance philosophy overall remains unchanged.
- 支持 OpEx 增速降到 26% (Q1 31%)，经营利润率 23% (Q1 22%)，出现经营杠杆
As we've discussed in the past, our gross margin varies from quarter-to-quarter with investments into innovations for our customers, offset by efficiency efforts. There's no change in our expectations for gross margin, which has been in the 80% plus or minus range historically. Q2 OpEx grew 26% year-over-year versus 31% last quarter and 36% in the year ago quarter. We held our DASH user conference in June, and as expected, the event cost about $15 million. Q2 operating income was $257 million for a 23% operating margin compared to 22% last quarter and 20% in the year ago quarter. Turning to our balance sheet and cash flow statements. We ended the quarter with $5 billion in cash, cash equivalents and marketable securities.
- 反对 稀释股数指引上升：Q3 约 378M、FY26 约 376M (Q1 时 FY26 指引约 372M)
For the third quarter, we expect our revenue to be in the range of $1.135 billion to $1.145 billion, which represents a 28% to 29% year-over-year growth. Non-GAAP operating income is expected to be in the range of $260 million to $270 million, which implies an operating margin of 23% to 24%. And non-GAAP net income per share is expected to be in the $0.63 to $0.65 per share range based on approximately 378 million weighted average diluted shares outstanding. For the full fiscal year 2026, we expect revenue to be in the range of $4.45 billion to $4.47 billion, which represents a 30% year-over-year growth. Non-GAAP operating income is expected to be in the range of $1.01 billion to $1.03 billion, which implies an operating margin of 23% and non-GAAP net income per share is expected to be in the range of $2.50 to $2.54 per share based on approximately 376 million average diluted shares outstanding.
- 反对 毛利率跌破 80%，但经营利润率靠杠杆超预期
* 26Q2 revenue grew 35.6% YoY, in line with buy-side expectations of 35-36%, accelerating 3ppts QoQ. However, the US$46m beat was slightly below the US$50m last quarter. * 26Q3 revenue guidance was a touch light at 29.5% at the high end, where expectations were looking for a 3-handle. Applying a typical 5ppts beat would imply a marginal sequential growth deceleration in 26Q3. * **The miss appears mostly driven by usage optimization at OpenAI, DDOG’s largest customer. For 26Q2, we estimated a 20% QoQ decline in OpenAI’s spending, on which market consensus appeared too optimistic. Management confirmed it has renewed a 9-figure contract with OpenAI, but the latter’s usage will decline sequentially in 26Q3. We believe Chronosphere is gaining share within the OpenAI platform, as it prices 30-40% lower than DDOG.** * Core/non-AI usage growth continued to accelerate to high-20s from mid-20s, driven by strong cross-sell momentum and increasing usage from cloud migration and AI app development. * Full-year revenue guidance was raised by more than the 26Q2 beat flowed through, to 30-31% YoY, in line with expectations. * The margin trend remains strong — GPM dipped below 80%, but OPM beat expectations on operating leverage.
来源背景：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 - 支持 2Q FCF $279M (24.9% 利润率)，街 $221M；OPM 22.9%
FCF $279M, 24.9% margin vs. Street $221M
来源背景：TMT Breakout (TMTB) 日报，汇总 DDOG 2Q26 公布数字、街上与买方预期，并评论股价反应。二手转述，看不到底稿。 - 中性 利润率持平，R&D 构成向 token 倾斜，“是分配不是利润侵蚀”
We still have many areas, I can spend a lot of time on that. So, that we're going to continue to invest in. When it comes to R&D, I think we're going to continue to invest at a high rate, but it's very likely and it is happening that the percentage of that, the allocation of that is going to move more towards tokens than it is towards humans. We're playing around with it. But you asked about margins. You can see our margins haven't changed, right? But we are also – we're not tokenmaxxing. We're basically using tokens in order to create good software and products to release to clients.
- 中性 分析师概括：收入加速、利润率大致持平、FCF 利润率 high-20s
So, I mean, you're accelerating revenue. Yeah. You're holding margins roughly flat, free cash flow in the high-20s. Can you just give us a sense of how you plan on investing in this, again, the world-class R&D organization and how you also manage the pace of hiring sales and what's the, kind of, dynamic consumption environment? Very good question. So, I think for right now, there needs to be salespeople to sell, meaning we haven't gotten agents to be able to sell. So, we look at that as what's the TAM, what are we covering, what do we need to continue to be able to sell our software. And right now, we've done a good job. We have successfully ramped quota capacity on a global basis, and it's been roughly in line with revenues.
- 中性 R&D 维持约 30% 收入，构成转向 token 和训练模型的 GPU
So, we're heavily dogfooding all of that, so we're building the right products for our customers on that. And there's a lot of opportunity in each of those area, whether that's shipping to production, keeping it running on production, measuring the value with the end users, are we actually shipping the right things, understanding what to work on or optimizing the costs, the token costs of writing code and automating. In terms of how we allocate the spend, the mental model you should have is that the overall envelope doesn't change all that much, so we still – we invest 30% in R&D, but the makeup of those 30% might change. And so, we're obviously spending more and more on tokens. These engineers are using a lot more agents and doing a lot more there. And we also started spending a lot more on GPUs. I mentioned earlier that we're training models. The first of those models were pretty small, but now we're starting to scale them and you should see more spend there also that's related to that.
- 中性 token 成本将从第三方模型转向自研模型，前期需要投入 (GPU、推理、人)
And similar to what you're talking about at Goldman, we own the data. We're the experts in observability. We have the ability to create from a functionality and also a cost model that we most perform it, and we're starting to invest behind it, which I think is going to result in reallocation of the cost of our GPU – our token cost towards our proprietary models. Early on, it takes investment. It takes a lift to do that. But just like you're talking about at Goldman, we're seeing that in the metrics we're seeing in our observability, and also with Datadog is doing in its own research lab.
- 中性 R&D 持续约 30% 收入，超 $10 亿
But in this case, I think we're essentially investing behind, making sure that we are the leading innovator here so that the balance of trade between modern point solutions and the platform of Datadog is weighted in our favor, and that's why we're investing significantly. Now, I think it is a very big competitive advantage. Datadog – and if Oli were sitting here, he would say, the thing that investors should think about is Datadog has consistently invested 30% of its revenues in R&D. That's over $1 billion. That is out-investing everybody else in the market. And why that's important is exactly your point.
产品理解
工程师一天怎么用 Datadog
这一节回答"工程师每天实际怎么用 Datadog"。前 12 步是一次线上故障从告警到复盘的完整流程，后 6 步是平时的用法。每一步都对应官方文档原文 (点链接可看)。
| 步骤 | 发生了什么 | 谁在做 | 用到的产品 |
|---|---|---|---|
| 1. 告警触发 (故障流程) | 工程师事先给关键指标设好阈值或异常规则(monitor)。指标越界时 Datadog 自动发告警，通常推送到 Slack/邮件/手机。这是故障处理的起点。 | SRE / 服务负责的后端工程师 | Monitors (告警) |
| 2. 呼叫值班人 (故障流程) | 告警按团队、优先级路由给当班工程师(电话/短信/推送)。无人确认就按升级策略找下一个人。这一块过去是 PagerDuty 的地盘，Datadog 用 On-Call 产品切入。 | 当班 (on-call) 工程师 | On-Call |
| 3. 声明事故、拉群 (故障流程) | 影响较大时声明 incident，自动建 Slack 频道、指定指挥官、记录时间线。也可以配置告警自动声明 incident。 | 事故指挥官 (incident commander) + 相关团队 | Incident Management |
| 4. AI 先做一轮调查 (2025 年后的新步骤) | 告警一触发，Bits Investigation(原名 Bits AI SRE)就可以自动开始查：翻指标、日志、trace、变更记录，给出根因假设。工程师也可以在 Slack 里 @Datadog 让它查。 | AI 代理，工程师审核结论 | Bits Investigation / Bits AI SRE |
| 5. 看仪表盘，确认影响面 (故障流程) | 打开服务或业务的仪表盘，看错误率、延迟、流量从什么时候开始变化，判断影响多大、波及哪些服务。 | 当班工程师 / 指挥官 | Dashboards |
| 6. 看自动发现的异常 (故障流程) | Watchdog 用机器学习给系统建立正常基线，主动标出异常和可能的根因，帮工程师缩小排查范围。 | 当班工程师 | Watchdog |
| 7. 用 APM trace 定位到具体服务和调用 (故障流程) | 打开一条出错或很慢的请求 trace，用火焰图看这条请求经过了哪些服务、时间花在哪一步(例如某条 SQL 或某个下游 API)。 | 后端工程师 | APM / 分布式追踪 |
| 8. 从 trace 跳到同一时刻的日志 (故障流程) | trace 和日志用 trace ID 关联，在 trace 页面可以直接看到这次请求产生的日志，比如错误堆栈。这种跨数据类型的关联是 Datadog 平台化的核心卖点。 | 后端工程师 | APM + Log Management 关联 |
| 9. 在日志里搜索和聚合 (故障流程) | 在 Log Explorer 里按服务、错误码、版本等过滤和聚合日志，确认错误模式和开始时间。注意：只有被索引的日志才能这样快速搜索，索引按事件条数收费。 | 后端工程师 / SRE | Log Management (Log Explorer) |
| 10. 根因：是不是刚才那次发布? (故障流程) | 大部分故障和变更有关。服务打上 version 标签后，Datadog 能对比新旧版本，自动识别可能有问题的发布，并且可以对此设告警。 | 后端工程师 / 发布负责人 | Deployment Tracking / Faulty Deployment Detection |
| 11. 修复：回滚、重启或改代码 (故障流程，2026 年 AI 化) | 传统做法是工程师自己回滚或改代码。DASH 2026 推出 Bits Remediation(预览版)：在客户设定的权限内调用 API、执行脚本(如重启 K8s)，或者直接生成修复代码的 PR。 | 工程师审批，AI 执行 | Bits Remediation / Bits Code |
| 12. 复盘 (postmortem) (故障流程) | 事故结束后用模板自动生成复盘文档(时间线、影响、根因、后续改进项)，可以导出到 Notebooks、Confluence 或 Google Drive，并跟踪改进项。 | 事故指挥官 / 团队负责人 | Incident Management - Postmortems |
| 平时-A 发布监控 | 每次上线后对比新旧版本的错误率和延迟。DASH 2026 又推出 Bits Release(预览版)：从 PR 开始验证，一直跟到生产环境发布。 | 后端工程师 / 平台团队 | Deployment Tracking, Bits Release |
| 平时-B 云成本 | 把 AWS/Azure/GCP/OCI 账单导进来，和使用量指标放在一起看，查清楚成本为什么上涨。FinOps 团队和工程师都会用。 | FinOps / 平台工程 | Cloud Cost Management |
| 平时-C 控制 Datadog 自己的账单 (日志治理) | 团队会把日志按价值分到不同索引，设定保留期、配额和排除规则，控制索引费用。这是 Datadog 账单里最常被优化的一项。 | 平台 / 可观测性团队 | Log Indexes, Exclusion filters, Flex Logs |
| 平时-D 性能优化 | 用 Continuous Profiler 在生产环境持续采样，找到最耗 CPU、内存的函数和代码行，同时降低延迟和机器成本。 | 后端工程师 | Continuous Profiler |
| 平时-E 安全 | 安全团队把云审计日志、身份日志等送进 Cloud SIEM，用检测规则产生安全信号，再调查和响应。可以复用同一份日志管道。 | 安全运营 (SOC) / 检测工程师 | Cloud SIEM |
| 平时-F 在 AI 编码工具里直接查 Datadog (2026 新习惯) | 工程师在 Claude Code、Cursor、Codex 里通过 MCP Server 直接查日志、指标和 trace，不必打开 Datadog 网页。界面入口在变化，数据仍在 Datadog。 | 后端工程师 + AI 编码代理 | Datadog MCP Server |
从流程里看出的三件事：
- Datadog 的护城河在第 7–9 步：从一条出错的请求 (trace) 一键跳到同一时刻的日志和机器指标。这需要三类数据都在同一个平台、用同一套标签。客户把日志搬到 ClickHouse 以后，这种跳转就断了，所以 Datadog 推 Federated Logs 来补。
- AI 正在吃掉第 4–6 步和第 11 步：Bits 自动调查、Bits Remediation 自动修复。这是"人看界面"的那部分价值。如果客户用的是 Claude Code 通过 MCP 读 Datadog 的数据 (平时-F)，这部分价值就到了第三方 agent 手里。
- "控制 Datadog 自己的账单"是一项日常工作 (平时-C)：平台团队会持续给日志分级、设排除规则、设配额。这说明客户对"量"的控制是常态，不是例外。
真实客户故事
公开可查的客户故事。Datadog 官方案例经过挑选，第三方报道 (Pragmatic Engineer) 更能说明成本压力。
| 客户 | 发生了什么 | 数字 | 来源 |
|---|---|---|---|
| Coinbase | 2021 年 Coinbase 付给 Datadog 约 6,500 万美元，就是 Datadog 财报里那笔神秘的 6,500 万美元大单。2022 年 Coinbase 组建团队，准备迁到自建的 Grafana/Prometheus/ClickHouse，最后 Datadog 给了很优惠的条件，把客户留住了。 | 2021 年账单 6,500 万美元 | 链接 · Coinbase spent $65M with Datadog in 2021, and this was their due bill for that year. |
| Shopify | Shopify 曾决定自建可观测性平台、迁出 Datadog，后来裁掉了负责自建的团队，计划随之生变(全文在付费墙后)。 | 未披露金额 | 链接 · Shopify decided to build its own observability platform and migrate off Datadog. |
| Zendesk | Zendesk 工程师在 Datadog 博客上讲自家的降本项目：按服务建成本归因看板；trace 只保留根 span 来降低摄入量；日志去重。他们发现，即使 88% 的日志不进索引，摄入费照样要付。 | 最多 88% 日志不进索引但仍付摄入费；一个高量服务日志量降到 1/4；核心数据库 TCO 减半(借助 APM 找出慢查询) | 链接 · Even with up to 88% of logs excluded from indexing, we still paid ingestion costs, while the excluded logs made spikes harder to spot. |
| Rakuten Viber | Viber 从自建开源日志栈迁到 Datadog，只用 Flex Logs(低价存储层)，不用标准索引；用 Observability Pipelines 先过滤，其余日志直接归档到 S3。 | 冗余日志从每天 25 TB 降到 5 TB (-80%) | 链接 · We use Flex Logs exclusively, as indexing the full volumes of our logs is way too expensive |
| Nulab (Backlog/Cacoo) | SRE 团队很小，要管 300 多台主机。告警来了就从 Slack 触发 Bits Investigation：低优先级告警的排查从平均 30 分钟以上降到 5 分钟以内，一次深夜 DDoS 4 分钟查清。 | 50+ 服务、1,500+ 容器、300+ 主机、35+ AWS 账户；30 分钟以上→5 分钟以内 | 链接 · Many low-priority alerts that previously required at least 30 minutes on average to investigate can now be triaged with Bits in under 5 minutes. |
| iFood | iFood 的 SRE 在 Datadog 产品页上称，用 Bits Investigation 后 MTTR 下降 70%。这是厂商挑选的客户证言。 | MTTR -70% | 链接 · From day one, Bits Investigation started cutting our MTTR by 70%. It felt like adding a senior engineer to our team. |
| Energisa | 巴西电力公司 Energisa 称 Bits 4 分钟内给出准确根因，并指向该负责的内部团队。 | 根因 4 分钟以内 | 链接 · Bits immediately sped up our crisis room by delivering accurate root causes in under four minutes and directing us to the right team internally. |
| DelightRoom / Uber Freight | Bits AI SRE 正式发布(GA)的新闻稿引用了两家客户。Datadog 称发布前已在 2,000 多个客户环境中测试，跑过数万次调查。 | 2,000+ 客户环境测试；数万次调查 | 链接 · With Bits AI SRE being on-call 24/7 for us, MTTR for our services have improved significantly |
| Datadog 自身 (Q2 2026 电话会) | CEO Pomel 说 MCP 工具调用量又一次环比翻 4 倍；Agent 监控产品的用量在爆发；AI credits 计费模式正在推开。来源是第三方转录稿。 | MCP 工具调用环比约 4 倍(据称同比 22 倍以上，需核实)；Q2 客户约 33,400 家 | 链接 · a number of MCP tool calls quadrupling again quarter over quarter |
读法：Coinbase 2021 年一年付 $6,500 万，2022 年组队准备迁到自建的 Grafana/Prometheus/ClickHouse，Datadog 给了很优惠的条件才留住。这就是"大客户压价"在 2022 年的样子，和今天最大 AI 客户的故事几乎一样。区别是 Coinbase 留下了，最大 AI 客户只留下了一部分。
2026 年新产品
以 2026 年 6 月 DASH 大会为主。注意哪些是 GA (可以收钱)、哪些还是 Preview。
| 产品 | 时间/状态 | 是什么 | 怎么收费 |
|---|---|---|---|
| DASH 2026 Keynote (总述) | 2026-06-09 | 主题是 Bits AI 代理贯穿开发和运维全流程：自主检测、调查、修复、发布验证、测试。Datadog 称 DASH 期间发布了 170 多个新产品和功能。 | 多数新 Bits 功能消耗 AI Credits |
| Bits Detection | 2026-06-09 (Preview) | 根据服务、端点、依赖关系和部署历史，自动建立并维护告警覆盖，用历史行为做基线，不再依赖静态阈值。 | 预览版，未公布 |
| Bits Memories | 2026-06-09 (Preview) | 从调查、runbook、复盘、Slack 讨论中自动沉淀团队的运维经验，下次调查时复用。 | 预览版 |
| Bits Remediation | 2026-06-09 (Preview) | 调查得出根因后自动修复：调用 API、执行脚本、开修复 PR，并遵守客户设定的护栏。 | 预览版 |
| Bits Infrastructure Operations | 2026-06-09 (Preview) | 自主检测、调查、修复常见的基础设施问题(磁盘满、OOMKilled、证书过期等)，并在 PR 阶段标出有风险的基础设施变更。 | 预览版 |
| Bits Code (GA) | 2026-06-09 | 平台级编码代理，嵌在 Error Tracking、APM、Profiler、Code Security、Bits AI SRE 等产品中，基于生产遥测数据写修复代码。 | AI Credits，一次修复约 5 credits |
| Bits Release / Bits Testing | 2026-06-09 (Preview) | Release：从 PR 到生产环境逐个验证代码变更，针对 AI 生成代码的高频发布。Testing：自主探索应用，生成并维护合成测试。 | Bits Testing 11/30 前免费 (定价页) |
| Bits Agent Builder (GA) | 2026-06-09 | 在 Workflow Automation 之上加 AI 编排，让客户用自然语言定义专用 agent(告警响应、修复、报告)，并控制它能用哪些数据和工具。 | AI Credits，每次运行约 1 credit |
| Datadog MCP Server (GA) + MCP Apps + Pup CLI + Agent MCP | MCP Server GA 2026-03-09；MCP Apps/Pup/Agent MCP 2026-06-09 | MCP Server 让 Claude/Cursor/Codex 等外部 agent 直接查 Datadog 数据；MCP Apps 在这些工具里渲染交互图表；Pup CLI 面向命令行 agent，200 多个命令；Agent MCP(预览版)让 AI 通过 Private Action Runner 在主机上执行 shell 命令。 | 定价页未标价；每月 10 万次工具调用的合理使用限制 |
| Agent Console | 2026-06-09 | 统一查看企业内 Claude Code、Cursor、GitHub Copilot 和 Bits 代理的使用情况：谁在用、产出如何、钱花在哪、哪里有浪费。 | 未公布 |
| Agent Observability (原 LLM Observability)：Patterns + Bits Evals | 2026-06-09 (Patterns/Evals Preview) | 追踪 AI agent 的每次模型调用、成本、质量评估。Patterns 自动把生产环境的交互聚类；Bits Evals 自动做根因分析和提示词改进建议。 | 每 LLM span：免费 4 万/月；Pro $160/月含 10 万，超出 $3.50/万 |
| AI Guard | 2026-06-09 (Limited Availability) | 内联在客户自建 agent 前面的运行时护栏，拦截提示词注入、工具滥用、数据外泄，也防护编码代理的恶意 skills。 | $0.80/百万 token (按年)，按需 $1.20 |
| BYOC Logs (Bring Your Own Cloud，即 CloudPrem) | 2025-10-31 发布博客；DASH 2026 再次重点介绍 | 日志存储和检索跑在客户自己的云或机房(基于收购来的 Quickwit 引擎，存算分离、数据放对象存储)，同时保持与 Datadog SaaS 的关联分析和 AI 能力。对象是数据主权需求和超大日志量客户。 | 未在定价页找到 |
| Federated Logs | 2026-06-09 (Preview) | 在 Datadog Log Explorer 里直接查询 Databricks、ClickHouse 等外部存储里的日志，不必先把日志导入 Datadog。 | 预览版 |
| Infinite Cardinality Metrics (GA) | 2026-06-09 | 新的自定义指标计费选项：按指标名计费，不按标签组合数，面向 agent 式查询。 | 按指标名、随数据量计费 |
| GPU Monitoring (GA) | 2026-04-22 (2025-07 已有官方演示视频) | 统一查看云、自建机房、neocloud 上 GPU 集群的健康、利用率和闲置成本，并按团队分摊费用。 | $20/活跃 GPU/月 (按年) |
| Bits AI SRE → Bits Investigation | DASH 2025 发布；2025-12-02 GA 新闻稿；2026-03-05 新一代；目前已更名 Bits Investigation | 工作方式：告警触发(或在 Slack 里 @Datadog)后，读 runbook 和遥测数据，迭代提出根因假设并自我验证，把结论发到 Slack。2026-03 新版速度约快一倍，单次 3-4 分钟，数据源扩展到源代码、RUM、DBM 等，可以触发分诊和修复工作流。DASH 2026 起可从 incident 直接拉起，并作为具名响应人出现在事故记录里。 | AI Credits：一次自主调查约 6.5 credits，按 $1-1.30/credit 推算约 $6.5-8.5/次 |
| AI Credits 计费模式 | 2026 (Q2 2026 电话会称正在推开) | 把 Bits Chat、Investigation、Code、Agent Builder 统一成按 credit 计费的用量模式，每月重置，可设组织级和个人级上限。 | $500/500 credits/月，按需 $1.30/credit |
| Bits Security Analyst / Bits Threat Hunting | 2025-07 (Security Analyst 视频)；2026-06-09 (Threat Hunting，可独立用于任意 SIEM) | AI 自动分诊和调查 SIEM 安全信号，主动猎杀威胁。DASH 2026 起可独立用于其他厂商的 SIEM。 | 随 Cloud SIEM 每 TB 附送 100 次调查 |
视频导读
20 个视频都已核实 (标题、频道、日期)。建议顺序：先看平台总览和 AWS/APM 演示建立概念，再看事故流程，最后看 2026 年的 AI 和 MCP。时长大多因 YouTube 反爬没取到。
| 视频 | 频道 · 日期 | 看它是为了回答 | 注意 |
|---|---|---|---|
| DASH 2026 Keynote | Datadog · 2026-06-09 · 约 1:50 (章节列表结束于 1:49:33) | Datadog 2026 年的产品方向是什么？Bits 代理从调查扩展到检测、修复、发布验证，自主化到了哪一步？ | 发布会演示，很多功能是 Preview，要对照 GA 状态和定价页看哪些已经能收钱。 |
| DASH 2026, featuring Bits Detection, Bits Evals, Federated Logs, and more | This Month in Datadog | Datadog · 2026-06-24 | 没时间看 2 小时 keynote 时，DASH 2026 最重要的几项发布是什么？ | 官方精选，只挑亮点，不提限制和价格。 |
| Datadog Overview: See Inside the Platform | Datadog · 2021-04-30 | Datadog 的界面长什么样？基础设施、APM、日志、安全如何在一个平台里串起来？ | 2021 年的旧视频，没有 Bits AI 等新产品，只适合建立基本概念。 |
| AWS Monitoring Demo | Datadog · 2023-08-23 | 基础设施监控(按主机收费的入门产品)实际给工程师看什么：主机、容器、AWS 服务指标和地图？ | 营销演示用的是干净的样例环境，真实客户的标签和仪表盘要乱得多。 |
| Datadog Application Performance Monitoring (APM) | Datadog · 2021-05-03 | 什么是 trace 和火焰图？为什么 APM 是从告警到根因的关键一环？ | 2021 年的版本，界面已更新，但概念不变。 |
| Datadog Log Management: Analyze complex data sets | Datadog · 2025-07-24 | 日志在 Datadog 里除了搜索，还能怎样像表格一样分析(Sheets)？日志产品为什么越来越像数据分析产品？ | 标题是日志管理，内容主要是 Datadog Sheets 功能，不讲日志计费(摄入/索引/Flex)，计费要看定价页。 |
| Watchdog: Detect performance anomalies automatically | Datadog · 2018-07-13 | Datadog 在 LLM 之前的 AI 是什么样(基于统计基线的异常检测)？ | 2018 年视频，可以和 2026 年的 Bits Detection 对照，看自动检测这几年的演进。 |
| Datadog Incident Management Demo | Datadog · 2024-03-05 | 声明事故、拉 Slack 群、时间线、复盘在 Datadog 里怎么走？和 PagerDuty、incident.io 的重叠在哪里？ | 演示流程很顺，真实事故里大量协作仍然发生在 Slack 和 Zoom。 |
| This Month in Datadog: Datadog On-Call is now generally available | Datadog · 2025-01-29 | Datadog 什么时候、以什么功能进入值班呼叫(PagerDuty 的核心市场)？ | 月度新闻节目，只讲功能，不讲迁移成本和客户采用率。 |
| How SeatGeek uses Datadog Incident Management to resolve issues faster | Datadog · 2024-12-11 | 真实客户(票务平台 SeatGeek)为什么把事故流程放进 Datadog？ | 官方客户故事，经过挑选和包装，没有负面信息。 |
| Introducing Bits AI SRE, your AI on-call teammate | Datadog · 日期未取得 | Bits AI SRE 收到告警后具体做哪些步骤(假设、取证、结论、推送 Slack)？ | 产品已更名 Bits Investigation，计费改为 AI Credits；演示场景是预先挑好的，真实准确率要看客户反馈。 |
| Datadog at AWS re:Invent, Bits AI SRE, MCP Server, CloudPrem, and more | This Month in Datadog | Datadog · 2025-12-10 | Bits AI SRE GA、MCP Server、CloudPrem(BYOC Logs)这三条线在 2025 年底各进展到哪一步？ | CloudPrem 后来改名 BYOC Logs，MCP Server 2026-03 才 GA，要注意时间线。 |
| Datadog LLM Observability: Monitor and secure your AI workloads | Datadog · 2025-07-25 | AI 应用团队用 Datadog 看 agent 的执行路径、token 成本、质量评估时，界面上看的是什么？ | 产品已更名 Agent Observability，按 LLM span 计费；竞品(LangSmith、Arize、Braintrust)不会出现在官方视频里。 |
| How Appfolio uses Datadog LLM Observability to deliver exceptional GenAI experiences | Datadog · 2024-10-29 | 真实企业客户为什么在已有 Datadog 的情况下用它监控 GenAI，而不是买专门的 LLM 工具？ | 官方客户故事，没有用量和花费数字。 |
| Datadog GPU Monitoring: Optimize and troubleshoot AI infrastructure | Datadog · 2025-07-25 | GPU 监控解决什么问题(闲置 GPU、按团队分摊成本、训练卡顿)？谁会买？ | 视频早于 2026-04 的 GA；按活跃 GPU 每台每月 $20 计费，要和 NVIDIA DCGM 等免费工具对比。 |
| Datadog MCP Server, Experiments, Bits AI Security Analyst, and more | This Month in Datadog | Datadog · 2026-04-29 · 约 8-9 分钟 (章节到 8:39) | 外部 AI agent(Claude Code、Cursor)怎样通过 MCP 用 Datadog 数据？这对 Datadog 的界面入口和计费意味着什么？ | 官方月度节目；MCP 本身不单独收费，变现靠底层数据。 |
| Building the Datadog MCP Server | Datadog Community · 日期未取得 | Datadog 工程师怎样设计给 agent 用的工具接口(返回截断、token 预算等)？ | 是 Datadog Community 频道，不是主频道；偏技术，PM 可以只看开头和结论。 |
| Datadog Cloud SIEM Demo | Datadog · 2024-01-11 | Cloud SIEM 如何复用同一份日志做威胁检测？和 Splunk、CrowdStrike 的 SIEM 有什么区别？ | 2024 年版本，没有 Bits Security Analyst；SIEM 是 Datadog 在安全市场的追赶型产品。 |
| Bits AI Security Analyst: Automate Cloud SIEM investigations | Datadog · 2025-07-23 | AI 怎样自动分诊 SIEM 告警？为什么 Datadog 把调查次数附送在 SIEM 里？ | 营销演示；每 TB 附送 100 次调查，说明它目前是 SIEM 的配套功能，不是独立收入来源。 |
| Datadog vs Grafana vs ClickHouse: Build or Buy? | Devsplainers · 2026-07-10 · 7:05 | 客户为什么考虑从 Datadog 迁到 Grafana 或 ClickHouse 自建？核心是账单还是能力？ | 小型第三方频道，观点未经核实；可以和 Coinbase、Shopify 的真实迁移故事对照着看。 |
反过来想 (Invert)
芒格说"反过来想，总是反过来想"。这一节不问"DDOG 怎么赢"，而问"怎样才会让持有 DDOG 的人在 2 年后亏掉一半"，再看每条路径今天有多少证据。
| 亏一半的路径 | 机制 | 今天的证据 | 可能性 |
|---|---|---|---|
| 1. 非 AI 增速走平 + 估值回到历史中枢 | 上云这一轮是 2022–24 年优化周期结束后的反弹，2027 年回到 20% 左右；EV/Sales 从 21 倍回到 12–14 倍 (2024–25 年的大致区间) | Bernstein 预计 Q4 降到约 29%、AWS 非 AI 负载可能再弱 100bp；全年指引隐含 Q4 +25%；CEO 一年前自己说"环境不再逆风" | 中 |
| 2. 前十大 AI 客户里再有一两家走 OpenAI 的路 | "规模大、应用少"的客户能自建；Chronosphere/ClickHouse 已经证明能接走指标和日志 | Chronosphere 9 月拿下第二家 AI 推理客户 (约 $2,000 万 ACV)；Anthropic 底层存储也用 ClickHouse (Funda 渠道) | 中 |
| 3. Datadog 退化成"实时数据层" | 第三方 agent 通过 MCP 读数据做排障，Bits 收不到钱；日志等大量数据存在客户自己的湖仓，Datadog 只收实时部分 | CEO 承认通用模型 + MCP 能做类似推理；CPO 说是"有意思的争夺"；Federated Logs、BYOC 是 Datadog 自己推的 | 中低，2–3 年以上 |
| 4. AI 成本合理化波及观测预算 | 客户 AI 支出暴涨后 CFO 开始砍，观测是最容易优化的一项 (日志) | CEO 5 月自己预警"可能比想象来得早"；2022 年 Coinbase 先例；北美渠道专家预计 2026 年定价能力下降 | 中 |
| 5. 什么都没发生，但股价不涨 | 收入每年 +25%，SBC 21%、稀释 2–3%/年，估值从 21 倍收敛到 15 倍，2 年回报接近 0 | 这是我的基本情形：按每股 FCF 口径，$258 的 5 年年化回报约 2% | 高 |
反过来，什么会让我错：
- agent 的遥测真正按量进收入：如果 2027 年 AI credits、Agent Observability、GPU 监控合计占到收入的 10% 以上，而且老产品单价不再下降，那"量涨"就真正变成了"收入涨"，25–30% 可以维持更久。
- 传统企业上云从 25–30% 走到 50%：CFO 的说法如果成立，未来 5 年的主驱动不是 AI 而是上云，这比 AI 更可预测。这是 bull case 最扎实的一条。
- SBC 开始下降：如果 SBC 从收入的 21% 降到 12–14%，扣 SBC 的 FCF 率会从 6% 升到 15% 以上，估值逻辑会大不一样。目前没有任何管理层表态。
框架打分：计费单位三问和赢家五条
计费单位三问 (我们在 TEAM 之后定的 AI 定价筛选)
| 问题 | DDOG 的答案 | 和 TEAM 比 |
|---|---|---|
| 1. AI 让计费单位变多还是变少？ | 变多，但有闸门 | TEAM 按 seat 收费，agent 替人以后 seat 可能变少。DDOG 按主机、数据量、span、模型调用收费，AI 让应用变多 (400 → 650)。闸门是客户采样、分层，以及预算锚定云支出 |
| 2. 被计量的东西是不是它自己拥有的？ | 部分是，而且在主动让出存储 | DDOG 拥有采集 (agent)、实时处理和关联分析；但日志存储正在通过 BYOC、Federated Logs 让给客户的云和 ClickHouse/Databricks。它保住的是"实时 + 关联"这一层 |
| 3. 产出能不能数？ | 能 | 主机、GB、span、调查次数都能数，AI credits 按次扣。但能数也意味着客户能精确优化 |
赢家五条 (从 CRWD/DDOG/NET/SNOW 的反转里总结的)
| 条件 | DDOG 现在 | 判断 |
|---|---|---|
| ① AI 让计费单位变多 (不是 seat) | 用量制，几乎没有 seat 收入；AI 应用、MCP、GPU 都是新增单位 | 满足 |
| ② 报告增速连续 ≥2 个季度加速，指引上调 | 收入 25% → 28% → 28% → 29% → 32% → 36%；全年指引 18–20% → 25–27% → 30% | 满足 |
| ③ 不是靠一个 AI 客户 | 剔除最大客户后连续 5 季加速；非 AI 约占 85% | 满足 |
| ④ 拥有被计量的东西 (毛利率漂移 ≤3–4pt) | 毛利率 81.7% → 79.6%，漂移约 2pt；但存储层在让出 | 基本满足 |
| ⑤ 和板块一起被错杀，而 1–2 条已经成立 | 2026 年 2–4 月确实如此 ($191 → $105)，当时 ①③ 已经成立；现在已涨 2.5 倍 | 已过时点 |
| 价值检验：FCF-SBC 和估值 | FCF-SBC 约收入 6%，EV/Sales 约 21 倍 | 不满足 |
结论：DDOG 是这个框架下"赢家"的标准样本，满足前四条，所以 4 月低点之后涨了 2.5 倍。现在它和 CRWD、NET、SNOW 一样，卡在最后一条：好公司已经被定价成好公司了。我们 mapping 里的结论"winners now fail value test"对 DDOG 同样成立。
估值和买点
现在的价格：$258.51 (2026-10-20 收盘)，稀释股数约 376M，市值约 $972 亿；净现金约 $40 亿，EV 约 $932 亿。EV/2026 收入 (指引中值 $44.6 亿) 约 21 倍；如果 2027 年增长 25%，EV/2027 收入约 17 倍。过去 12 个月 FCF 收益率 1.1%，扣 SBC 后 0.3%。
两种口径
- 口径 A (你的默认口径)：把 SBC 当成真实成本。2031 年的 (FCF − SBC) × 终值倍数，按 10% 折现回今天，加净现金，除以今天的股数。
- 口径 B (市场在用的口径)：不扣 SBC，但把它当稀释：2031 年的每股 FCF (股数每年 +2.5%) × 终值倍数，再折现。这个口径偏乐观：它把约占收入 21% 的 SBC 从成本里拿掉，只用每年 2.5% 的股数增长来近似 SBC 的代价。
| 情景 | 2027–31 年增速 | 2031 收入 | FCF 率 / SBC 率 | 终值倍数 | 口径 A 价值 | 口径 B 价值 | 现价 5 年年化 (A / B) |
|---|---|---|---|---|---|---|---|
| 悲观 | 20/16/14/13/12% | $90 亿 | 28% / 16% | 30x | $64 | $120 | -19% / -7% |
| 基本 | 25/22/20/18/16% (复合 20%) | $112 亿 | 30% / 14% | 35x | $114 | $182 | -8% / +2% |
| 乐观 | 30/28/26/24/22% (复合 26%) | $141 亿 | 33% / 12% | 45x | $231 | $317 | +8% / +15% |
2026 年收入取指引中值 $44.6 亿；折现率 10%；终值倍数是对 2031 年 (FCF − SBC) 或每股 FCF 的倍数。乐观情景的 26% 复合增长相当于 Wolfe 上行情景 (2025–28 年 32%) 之后逐年减速。
反推：现价隐含了什么
要在 $258 买入还拿到 10% 年化：
- 口径 A，FCF − SBC 率 16%、35 倍：2031 年收入要到约 $268 亿，5 年复合 43%。
- 口径 A，SBC 降到 10% (FCF − SBC 率 20%)：收入要到约 $214 亿，复合 37%。
- 口径 B，FCF 率 30%、35 倍：收入要到约 $162 亿，复合约 29%。
所以市场是按口径 B、5 年复合约 29% 在定价。这比 Wolfe 的上行情景还要再持续两年。考虑到最大客户在 2027 年还会拖累 4–5 个点、全年指引隐含 Q4 只有 +25%，这个预期偏满。
两年一倍检验
$517，市值约 $1,940 亿。2028 年收入按每年 +25–30% 算是 $68–74 亿，需要 26–28 倍 EV/Sales。DDOG 历史上只有 2021 年泡沫期到过这个水平。不满足。
买点
| 价格 | 口径 A 基本情形 5 年年化 | 口径 B 基本情形 5 年年化 | 怎么看 |
|---|---|---|---|
| $100 | 13% | 25% | 两种口径都便宜，2023 年初的 $64 和 2026 年 4 月的 $98 都曾出现 |
| $130 | 7% | 18% | |
| $150 | 4% | 15% | 买点区间上沿以下 |
| $170 | 1% | 12% | 买点区间 |
| $200 | -3% | 8% | |
| $258 (现价) | -8% | 2% | 不买 |
| $300 | -11% | -1% |
结论：买点 $150–170，对应口径 B 基本情形 12–15% 年化，和 primer 的区间一致。口径 A 下基本情形最多值约 $114，只有在你相信乐观情形 (5 年复合 26%、SBC 降到 12%) 时，$150–170 才同时在口径 A 下有回报。所以这个买点隐含的判断是：SBC 会下降，或者增长会好于基本情形。
自己算
置信度检查
| 来源 | 检查范围 | 身份与视角 | 影响判断的问题 | 采用边界 |
|---|---|---|---|---|
| 4 场业绩会 (Q3 2025–Q2 2026) | 全文 | CEO Pomel、CFO Obstler、IR | 最大客户减量幅度、单价变化、AI 产品收入都拒绝给数；Q1、Q2 2025 原文没取到 | 可作主要证据 (限"管理层怎么说") |
| 8 场会议 (JPM、Bernstein、BofA、RAISE、Canaccord、Citi、GS、股东大会) | 全文 | CEO 或 CFO | FactSet 近实时转录有错字 ("known AI" 应为 non-AI 等) | 可作主要证据 |
| Merit 音频 4 场 + MeritForum | 全文 | 渠道商、SI、SHI 前员工、Datadog 印度前员工 | MeritForum 整理稿和音频有矛盾 (以音频为准)；印度前员工偏乐观、可能涉及内部信息；渠道商只看到自己的客户 | 需交叉验证 |
| GuidePoint / ThirdBridge 7 场 | 全文 | 独立顾问、渠道 CTO、前 Datadog 产品营销/工程/销售、银行平台负责人 | 前员工有立场；PNC 专家现任/前任表述矛盾 | 需交叉验证 |
| 中文专家纪要 (MeritForum、Acecamp) | 全文 (一篇只读 DDOG 章节) | 北美/欧洲渠道、GSI、单一用户 | 整理稿非逐字；汇总稿和原访谈数字不一致 (Bits POC 比例、日志迁移比例) | 仅作线索或需交叉验证 |
| Funda、TMTB | Funda 全文；TMTB 只检索关键词段落 | 二手转述卖方、渠道和客户 | 看不到底稿；Funda 两篇对同一专家 Q1 增速说法不一 (21% vs 26%) | 需交叉验证 |
| 周边材料 (安全、Snowflake、AI infra 调研中的顺带提及) | 关键词段落 | 各类专家 | 多为顺带评论；三组 Merit/MeritForum 是同一通话 | 仅作线索 |
| 产品和定价 | 官方定价页、文档、博客、YouTube | Datadog 官方 | 官方案例经过挑选 | 定价可作主要证据；客户故事仅作线索 |
影响判断的口径问题
| 原话或数字 | 问题 | 对结论的影响 |
|---|---|---|
| "8 位数新 logo 是最大的基础模型公司之一" → 推测 Anthropic | 推测 | E02 里"补上 OpenAI 缺口"的说法打折 |
| >$1M/>$10M AI 客户 22/5 → 31/8 | Q2 起包括超大厂的内部 AI 实验室 | 不能完全算作 AI 原生的自然增长 |
| Bits 定价 $500/20 次 → 6.5 credits/次 | 两次计费单位可能不同 | 单价降幅不能直接算 |
| Jason："OpenAI 占收入低到高个位数"、"Bits 占收入低到中个位数" | 前员工，可能是内部信息，也可能是推测 | 只作为乐观端的参照 |
| MCP "22 倍" | 只有增速，没有绝对值 | 不能判断规模 |
| 渠道增速 (+31–35%) vs 公司 (+32–36%) | 渠道的客户群和公司不同 | 方向一致，可作印证 |
附：来源与阅读记录
共 95 份有效材料：管理层发言 12 场 (4 场业绩会、8 场会议)，其余为专家访谈、渠道调研、卖方和 Substack 摘要。分级：A 核心 33 / B 实质 37 / C 顺带提及 25 / X 未取到 2。证据 746 条，另有反方原话 122 条，全部经脚本逐字比对原文。三组 Merit 音频和 MeritForum 整理稿是同一通话，Funda 8/9 周报的两条客户摘要可能和 8/3 是同一批人，计数时已注意，但没有从总数里剔除。
专家与外部来源背景
| 日期 | 来源 | 背景与局限 | 分级 | 可信度 |
|---|---|---|---|---|
| 2026-01-07 | Acecamp | Acecamp 作者整理的卖方报告和买方讨论 (不是专家访谈)；涉及 DDOG 的是转述的公开数字和高盛的观点。 | C | 仅作局部线索 |
| 2026-01-14 | Acecamp | 未具名；对 PANW 的内部云部署、并购节奏、团队构成很熟，推测为 Palo Alto Networks 现任或前任员工 (没有明示)。局限：安全厂商视角，对 Datadog 只是顺带一提。 | C | 仅作局部线索 |
| 2026-01-29 | MeritForum | 未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO ，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。 | A | 仅作局部线索 |
| 2026-01-30 | MeritForum | 未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。 | A | 需交叉验证 |
| 2026-02-03 | MeritForum | MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。 | A | 需交叉验证 |
| 2026-02-03 | MeritForum | 未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。 | A | 需交叉验证 |
| 2026-02-11 | MeritForum | MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。 | A | 需交叉验证 |
| 2026-03-13 | Acecamp | 未具名；正文明确是埃森哲的人 ('埃森哲2026财年Q2'、'埃森哲倾向于使用自有的工具')，举的案例 (美孚中国区验证) 和国内模型对比说明主要是中国/亚太视角，职位不详。局限：咨询公司有推销自身方案的利益偏向；Datadog 只是被提问者点名，回答没有具体数据。 | B | 仅作局部线索 |
| 2026-04-17 | GuidePoint | Dave Sidwell，Improving Holdings 芝加哥区总裁 (President for Chicago)，多云咨询/实施伙伴 (Industry Consultant，标题明示)；主持人 Brian Oh。视角局限：只看自己公司的客户，增速是他们自己 AWS 业务的数字，不是 AWS 整体。 | B | 需交叉验证 |
| 2026-04-25 | ThirdBridge | Elastic NV 前 Senior VP, Channel Sales (Americas 渠道，标题明示)，在企业基础设施软件和云生态有 30 多年经验；访谈实际日期 2026-02-25 (Client-led)。视角局限：渠道销售视角，不碰技术；离职时间未说明。 | C | 仅作局部线索 |
| 2026-05-04 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 JPM (渠道检查及一位行业联系人) 与 UBS (客户检查) 的报告摘要，二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-05-05 | Funda (Substack) | FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。 | A | 需交叉验证 |
| 2026-05-07 | Funda (Substack) | FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。 | A | 可作主要证据 |
| 2026-05-07 | TMTB (Substack) | TMT Breakout (TMTB) 日报，作者市场评论 + TMTB Slack 成员 Josh S 的评论，转述业绩会要点。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-05-07 | TMTB (Substack) | TMT Breakout (TMTB) 日报，汇总 DDOG 1Q26 公布数字与街上预期 (Street 口径来源未注明)，转述 CEO 评论。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-05-11 | Funda (Substack) | FundaAI 周报，DDOG 部分是 5/5 preview 和 5/7 review 的摘要重述；仓位部分引用 Coatue 演示材料。二手转述，看不到底稿。 | C | 仅作局部线索 |
| 2026-05-13 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Sohn 投资大会上 Whale Rock (Alex Sacerdote) 和 Maplelane (Leon Shaulov) 的 AI panel 观点。买方观点，TMTB 摘要。二手转述，看不到底稿。 | C | 仅作局部线索 |
| 2026-05-19 | TMTB (Substack) | TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。 | A | 可作主要证据 |
| 2026-05-27 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (渠道检查对象为 Cognizant)。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-05-28 | TMTB (Substack) | TMT Breakout (TMTB) 日报，作者对 SNOW 财报的评论中顺带提及 DDOG 仓位。二手转述，看不到底稿。 | C | 仅作局部线索 |
| 2026-05-29 | TMTB (Substack) | TMT Breakout (TMTB) 日报，MDB 财报点评中顺带提及 DDOG。二手转述，看不到底稿。 | C | 仅作局部线索 |
| 2026-05-31 | Acecamp | 未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。 | B | 需交叉验证 |
| 2026-05-31 | Acecamp | Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。 | B | 仅作局部线索 |
| 2026-06-04 | Merit 专家电话 (音频转写) | 未具名专家，Accenture (埃森哲) 前任人员 (身份由同场 MeritForum 06-04 标题'埃森哲离职专家'推得；本篇无自我介绍)。以'my portfolio'口吻谈客户，推测为咨询/集成侧的安全业务负责人。视角局限：咨询商视角，只见自己客户组合；主题是 Palo Alto/CyberArk 安全，DDOG 为顺带；Chronosphere 数字为举例性量级。 | B | 需交叉验证 |
| 2026-06-04 | MeritForum | 埃森哲 (Accenture) 离职专家 (标题明示)，具体职位未说明。同 Merit 06-04 音频那篇。视角局限同前：咨询商的客户组合视角。 | C | 仅作局部线索 |
| 2026-06-05 | Merit 专家电话 (音频转写) | 未具名专家，Capgemini 前任人员 (由同场 MeritForum 06-05 标题'Capgemini 离职专家'推得)；以'we/our customers'口吻谈 Palo Alto 增长预期，推测为 Palo Alto 渠道/集成伙伴侧负责人。视角局限：伙伴视角只看到自己的客户群；对 Datadog 仅顺带评论。 | C | 仅作局部线索 |
| 2026-06-05 | MeritForum | Capgemini 离职专家 (标题明示)，具体职位未说明。同 Merit 06-05 音频那篇。 | C | 仅作局部线索 |
| 2026-06-06 | Acecamp | 未具名；以 Elastic (ES) 内部视角回答 (清楚 ES 组织架构调整、API 文档策略、Jina 整合)，推测为 Elastic 现任或前任员工，带明显中国区视角 (国内定价、阿里/腾讯 ES 服务)。局限：竞争对手视角；对 Datadog 的判断是间接的。 | B | 仅作局部线索 |
| 2026-06-09 | TMTB (Substack) | TMT Breakout (TMTB) 日报，日程提示 + 转述 Wells Fargo 对 SNOW 的观点。二手转述，看不到底稿。 | C | 仅作局部线索 |
| 2026-06-10 | ThirdBridge | 匿名专家 ([Expert])。Third Bridge 主持人介绍为 'former Head of CCoE [Cloud Centre of Enablement] and Platform Engineering at PNC Financial Services'；但专家自述 '我是' PNC Cloud CoE 和平台工程负责人、在该行两年 。负责容器、中间件、数据平台及平台安全 (本地和云)；上一家机构曾把 Splunk 整合迁移到 Google Chronicle 。视角局限：受严格监管的大型银行买方，偏 '最佳单品 + 多厂商冗余' 策略，不代表一般企业；角色偏平台/安全，不直接管观测账单细节；对 Datadog 的评价较笼统、无数字。 | B | 仅作局部线索 |
| 2026-06-10 | Acecamp | 未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。 | A | 仅作局部线索 |
| 2026-06-10 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Wedbush (DASH 2026 后) 与 BofA (基础设施软件 'Fab Five') 报告摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-06-11 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 BNP Paribas DASH 2026 后客户调研摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-06-15 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Truist 升级报告摘要 (基于客户与伙伴交流)。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-06-16 | Merit 专家电话 (音频转写) | 未具名专家。音频原文没有专家介绍 (开头只有合规声明，)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。 | A | 需交叉验证 |
| 2026-06-16 | MeritForum | 标题写明 '北美软件渠道商离职专家' ，作者 Lantian；内容与 Merit Datadog 产品进展调研 I (2026-06-16 音频) 逐段对应，判断为同一专家同一场的英文转写再翻译版。局限同音频 I：只看到渠道客户，无样本量，回答偏框架化。 | A | 仅作局部线索 |
| 2026-06-16 | ThirdBridge | ServiceNow 前 Senior Executive, Alliances & Channels (标题明示)，有约 25 年软件行业经验；访谈实际日期 2026-04-17 (Client-led)。视角局限：ServiceNow 一方，自然强调 ServiceNow 的价值；不熟悉 Datadog 产品细节。 | C | 仅作局部线索 |
| 2026-06-17 | Merit 专家电话 (音频转写) | 未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' ，以及 '客户谈判时有时找我们帮忙' 。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。 | A | 需交叉验证 |
| 2026-06-17 | MeritForum | 标题写明 '北美软件渠道商离职专家' ，作者 Lantian；内容与 Merit Datadog 产品进展调研 II (2026-06-17 音频) 逐段对应，判断为同一专家同一场。专家自述 '我们主要与企业客户合作……后者通常不需要系统集成商的支持' ，实为系统集成商视角；'离职' 与原话现在时矛盾。 | A | 仅作局部线索 |
| 2026-06-17 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Citi (DASH 后与客户检查) 报告摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-06-18 | Merit 专家电话 (音频转写) | 未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。 | B | 仅作局部线索 |
| 2026-06-18 | MeritForum | 标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。 | A | 仅作局部线索 |
| 2026-06-19 | ThirdBridge | Christine Malloy：Third Bridge 介绍为 Elastic 前 Director of National Strategic Partners ；页头写的是 Perforce Software 现任 Partner Ecosystem Manager；自述曾在 New Relic 工作数年。访谈日期 2026-06-10 (Analyst-led，主持人 Luke Fragnito)。视角局限：伙伴/联盟视角；离开 Elastic 的时间未说明；曾在 New Relic 工作可能影响她对 Datadog 的看法。 | B | 需交叉验证 |
| 2026-06-22 | Merit 专家电话 (音频转写) | 未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3)，且 M3 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。 | A | 需交叉验证 |
| 2026-06-22 | MeritForum | 标题写明 'Datadog离职专家' ，作者 Lantian； 自述 '我来自印度'；内容与 Merit Datadog 产品进展调研 IV (2026-06-22 音频) 逐段对应，判断为同一专家同一场。局限同音频 IV：前员工，口头区间，可能 MNPI，偏看多。 | A | 仅作局部线索 |
| 2026-07-01 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 BofA (DDOG) 与 BTIG/BNP (PANW，涉及 Chronosphere) 报告摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-07-06 | Acecamp | 未具名；用 Cloudflare 内部视角回答 ('Cloudflare内部'、'我们能看到merge request的数量')，推测为 Cloudflare 现任员工，职位不详。局限：对 Datadog 只是顺带一提。 | C | 仅作局部线索 |
| 2026-07-06 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Bernstein 降级报告摘要。二手转述，看不到底稿。 | B | 需交叉验证 |
| 2026-07-09 | GuidePoint | Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI 。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。 | A | 需交叉验证 |
| 2026-07-16 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 BNP Paribas 季度经销商调查与 UBS 客户检查 (7 位客户 + 3 位伙伴) 摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-07-24 | GuidePoint | Kira Busman，Cognition 前顾问，前 Index Ventures 全球人才网络负责人；更早在 Datadog 负责全球人才招聘 (员工从约 300 人到 5,500 多人，经历 IPO) 。视角局限：招聘/人才视角，谈 Datadog 只是回忆，与当前业务无关。 | C | 仅作局部线索 |
| 2026-07-28 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 BTIG (伙伴检查) 与 Morgan Stanley (企业检查) 报告摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-07-30 | ThirdBridge | Terry Lucy (TL)。Third Bridge 页眉 'Executive, Go-to-market at CodeRabbit Inc'，介绍为 'former Strategic Account Executive at Monte Carlo Data' (明示)；自述在 Monte Carlo 多年，是伦敦办公室区域第二号员工，负责企业/战略客户 。视角局限：数据观测 (非 APM) 销售视角，对 Datadog 的看法来自竞争一侧；主题是 Monte Carlo，涉及 Datadog 的内容只有几段。 | C | 仅作局部线索 |
| 2026-07-30 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (AWS 收入分析)。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-07-31 | Merit 专家电话 (音频转写) | 未具名专家，LiteLLM 前任人员 (由同场 MeritForum 07-31 标题'LiteLLM 离职专家'推得)。视角局限：AI 网关/路由厂商视角，对可观测性采购为推测性判断，没有具体账单数据。 | C | 仅作局部线索 |
| 2026-07-31 | MeritForum | LiteLLM 离职专家 (标题明示)，具体职位未说明。同 Merit 07-31 那篇。 | C | 仅作局部线索 |
| 2026-08-03 | Funda (Substack) | FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。 | A | 需交叉验证 |
| 2026-08-03 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 FundaAI 预览 (同 8/3 Funda 原文) 和 Wells Fargo 2Q 客户调查 (约 $95M 年 DDOG 支出) 摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-08-06 | Funda (Substack) | FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。 | A | 可作主要证据 |
| 2026-08-06 | TMTB (Substack) | TMT Breakout (TMTB) 日报，汇总 DDOG 2Q26 公布数字、街上与买方预期，并评论股价反应。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-08-09 | Funda (Substack) | FundaAI 周报：DDOG 段落为 2Q review 摘要 + Funda 当周两次专家访谈的一句话摘要 (某全球云通信平台的工程高管——疑似 Twilio，因标题标 TWLO；某全球金融信息服务商经理)。二手转述，看不到底稿，访谈全文在 Funda 平台未附。 | B | 仅作局部线索 |
| 2026-08-10 | TMTB (Substack) | TMT Breakout (TMTB) 日报，市场评论。二手转述，看不到底稿。 | C | 仅作局部线索 |
| 2026-08-10 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Barclays 与 DDOG CEO/CFO 财报后 fireside 会面的报告摘要。二手转述，看不到底稿。 | B | 需交叉验证 |
| 2026-08-12 | Merit 专家电话 (音频转写) | 未具名专家，从'我们这边的客户''我们围绕 Snowflake 的可观测性平台开展了大量市场推广培训'推测为 Snowflake 内部 GTM 人员或紧密合作伙伴 (未明示)。视角局限：站在 Snowflake 一方；只看自己客户，未必了解 Datadog 内部数据。 | B | 需交叉验证 |
| 2026-08-13 | Merit 专家电话 (音频转写) | 未具名专家，从'我们已经进行了大量优化''我们还推出了自己的 AI Gateway'推测为 Databricks 现任员工 (产品/解决方案方向，未明示)。视角局限：谈 Databricks；Datadog 只出现在研究员提问里，专家没有回应 Datadog 部分。 | C | 仅作局部线索 |
| 2026-08-17 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Mizuho 与 DDOG 管理层 NDR 后报告摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-08-19 | ThirdBridge | Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。 | A | 需交叉验证 |
| 2026-08-20 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 BNP Paribas 招聘/销售生产力数据摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-08-27 | Merit 专家电话 (音频转写) | 未具名专家，现任某保险公司 (金融业) IT 负责人，此前在软件公司工作 (原文 'I worked in software companies and also uh now I'm working in ... the financial sector for an insur...')。视角局限：单一保险企业样本，AI 支出以账号和 Token 为主。 | C | 仅作局部线索 |
| 2026-08-28 | Acecamp | Acecamp 专栏文章，作者未署名 (系列作者)；不是专家访谈，是作者自己的分析。 | C | 仅作局部线索 |
| 2026-08-31 | Merit 专家电话 (音频转写) | Merit 研究团队 (路演主讲人，未具名)，转述渠道商和 Snowflake 专家的观察并加上自己的判断。视角局限：二手汇总，不是一手专家；主题是 Snowflake/MongoDB。 | C | 仅作局部线索 |
| 2026-09-02 | TMTB (Substack) | TMT Breakout (TMTB) 日报，PANW 财报点评，引用 PANW 管理层原话 (CEO/CFO 未具名) 与 TMTB 自己的分析。关于 '大型 LLM 客户' 未点名 (市场普遍认为是 OpenAI)。二手转述，看不到底稿。 | B | 可作主要证据 |
| 2026-09-05 | Acecamp | Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。 | B | 需交叉验证 |
| 2026-09-08 | GuidePoint | Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 ；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。 | A | 需交叉验证 |
| 2026-09-10 | Merit 专家电话 (音频转写) | 未具名专家，LVMH 集团内部人员 (原文多次以'集团''我们'谈 LVMH 与 Google Cloud 共建 MAIA，具体职位未明示，推测是数字化/IT 方向)。视角局限：奢侈品集团单一样本；本人承认不清楚 Datadog 等的具体采购金额。 | C | 仅作局部线索 |
| 2026-09-11 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Wedbush 报告摘要。二手转述，看不到底稿。 | C | 仅作局部线索 |
| 2026-09-15 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Bernstein (AWS 网站流量数据) 与 Citi (DT) 报告摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-09-16 | GuidePoint | Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 ，当期一手观察有限；全文无具体客户账单数字。 | A | 需交叉验证 |
| 2026-09-16 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Truist 与 DDOG IR 会面后的报告摘要。二手转述，看不到底稿。 | B | 需交叉验证 |
| 2026-09-23 | GuidePoint | Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) 。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 ；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 。 | A | 需交叉验证 |
| 2026-09-25 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Wolfe Research 生态检查报告摘要。二手转述，看不到底稿。 | B | 仅作局部线索 |
| 2026-10-09 | TMTB (Substack) | TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。 | B | 需交叉验证 |
| 2026-10-12 | TMTB (Substack) | TMT Breakout (TMTB) 日报，转述 Jefferies 报告摘要 (DDOG 与 DT)。二手转述，看不到底稿。 | B | 仅作局部线索 |
KB/ 开头的路径相对于你的 Knowledge Base 根目录；web/ 开头的是从 Datadog IR 网站取的 S&P 转录和新闻稿。
X 2025-05-06 · S&P 业绩会转录 (Datadog IR) · Datadog Q1 2025 业绩会 transcript：未取到 0 条
文件：未取到
发言人：
阅读范围：未取到全文 可信度：仅作局部线索
IR 页面没有 Q1 2025 transcript；搜索到的 static-file (6019d4df-...) 已 404；第三方 transcript 网站全部被代理拒绝。8.5% 的 AI 原生占比无法用原文核实；后续业绩会回溯 Q1 2025 非 AI 增速为 19% (Q1 2026 业绩会 L44)。
问题：全文不可得，没有编造；8.5% 未核实
X 2025-08-07 · S&P 业绩会转录 (Datadog IR) · Datadog Q2 2025 业绩会 transcript：未取到 0 条
文件：未取到
发言人：
阅读范围：未取到全文 可信度：仅作局部线索
Datadog IR 季度结果页 Q2 2025 一栏没有 transcript 链接；搜索结果中的 IR static-file (4b5b9407-...) 已 404；fool.com、Seeking Alpha、Insider Monkey、AOL/Yahoo、bullfincher、platformaeronaut 等均被代理拒绝 (403 CONNECT)。只有新闻稿 (DDOG_2025-08-07_Q2-2025_press_release.txt) 和后续业绩会中的回溯数字 (Q2 2025 非 AI 增速 18%、AI 原生占比 11%)。
问题：全文不可得，没有编造
A 2025-11-06 · S&P 业绩会转录 (Datadog IR) · Datadog Q3 2025 业绩会：非 AI 客户增速 18%→20% 拐点、AI 原生占收入 12%、最大客户提前续约换更低单价 24 条
文件：web/DDOG_2025-11-06_Q3-2025_call.txt
发言人：管理层：Olivier Pomel (联合创始人兼 CEO)、David Obstler (CFO)、Yuka Broderick (IR SVP)；分析师：Kash Rangan (GS)、Sanjit Singh (MS)、Raimo Lenschow (Barclays)、Mark Murphy (JPM)、Fatima Boolani (Citi)、Eric Heath (KeyBanc)、Gray Powell (BTIG)、Koji Ikeda (BofA)、Ittai Kidron (Oppenheimer)、Andrew Sherman (TD Cowen)、Alex Zukin (Wolfe)、Ryan MacWilliams (Wells Fargo)、Mike Cikos (Needham)、Karl Keirstead (UBS)、Jake Roberge (William Blair)。身份来自 transcript 参会名单 (明示)。局限：管理层自述，有利益偏向；S&P 转录有明显语音识别错误 (如 "$251 billion" 应为 million、"good security" 应为 code security)。
阅读范围：全文 L1-L293 (正文 L49-L290) 可信度：可作主要证据
Q3 2025 收入 $886M (+28%)，非 AI 原生客户增速从 Q2 的 18% 加速到 20%，AI 原生客户占收入 12% (Q2 11%、一年前约 6%)。最大 AI 客户提前续约，"更高承诺换更好价格"，管理层承认续约会让同样用量下收入下降 (价格下行的直接证据)，拒绝给该客户贡献变化。管理层把非 AI 加速归因于需求环境不再逆风、销售产能扩张和新产品，但承认没看到上云大幅加速。
问题：S&P 转录稿有较多识别错误 (L97 "$251 billion" 应为 $251 million；L67 "good security" 疑为 code security；L119 "large surprises to adult Flex Logs" 疑为 large enterprises to adopt)，引用时以数字为准并与新闻稿核对；AI 原生收入占比 6% 指 Q3 2024；本会没有给 Q4 2024 / Q1 2025 的占比 (8.5% 那一档需要 Q1 2025 业绩会原文，本次未取到)；"better economics" 的方向 (对客户还是对 Datadog) 原文有歧义，Mark Murphy 追问后 CEO 的解释是客户拿到更低价格
C 2026-01-07 · Acecamp · 美股软件十问十答 (Acecamp 作者综合卖方报告)：DDOG 只是顺带提及 2 条
文件：KB/Acecamp/2026-01/2026-01-07-mei-gu-ruan-jian-shi-wen-shi-da-26nian-aiying-yong-fa-zhan-qian-jing-zhan-wang-ji-yu-mai-fang-bao-gao-mai-fang-tao-lun-c.md
发言人：Acecamp 作者整理的卖方报告和买方讨论 (不是专家访谈)；涉及 DDOG 的是转述的公开数字和高盛的观点。
阅读范围：全文 L1-L248 可信度：仅作局部线索
主题是软件板块。DDOG 只出现两处：AI 原生客户约占收入 11% (Q2'25)；高盛提到 DDOG 等公司业绩好但股价下跌，反映市场担心 AI 冲击 SaaS。
问题：作者综合稿，卖方引文为机器翻译
C 2026-01-14 · Acecamp · PANW 相关专家：Chronosphere 9 月末 ARR 1.6 亿美元，收购后三年独立运营，PANW 在观测领域的前景难说 2 条
文件：KB/Acecamp/2026-01/2026-01-14-wizbei-gu-ge-shou-gou-hou-de-ke-hu-qian-yi-jie-zou-da-han-zheng-he-chronospherejin-zhan-he-zhan-lue-kao-lu-bei-mei-csphe.md
发言人：未具名；对 PANW 的内部云部署、并购节奏、团队构成很熟，推测为 Palo Alto Networks 现任或前任员工 (没有明示)。局限：安全厂商视角，对 Datadog 只是顺带一提。
阅读范围：全文 L1-L100 可信度：仅作局部线索
Chronosphere 9 月末 ARR 1.6 亿美元，增长很快，被 PANW 收购后三年内独立运营。他认为 PANW 的销售团队不是为可观测性搭的，市场目前不太买账，不过整合在技术上没有障碍。他还提到模型厂一般自建安全体系，可类比 AI 实验室自建可观测性。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；身份为推测
A 2026-01-29 · MeritForum · 北欧渠道 (GSI) 专家：其 Datadog 业务 2025 年约 +30%，AI 推动需求，但客户对价格和锁定的担忧在上升 11 条
文件：KB/MeritForum/2026-01/2026-01-29-ddog-bei-ou-qu-dao-zhuan-jia-ai-qu-dong-xu-qiu-zeng-chang.md
发言人：未具名；全球系统集成商 (GSI) 在北欧/比荷卢地区的资深人员 (自述 '我在公司里也属于资深级别')，负责外包协议中的 IT 运维和 AIOps 咨询，Datadog 是其中一部分。公司约 30–32 万员工、大量员工在印度、约 10 万人做 BPO (L216-L223)，推测为印度系大型 GSI (没有点名)。汇总稿 (2026-02-03 DDOG 4Q调研总结) 把他列为 '欧洲渠道专家C'，说 '该专家DDOG业务体量较小'。他直接负责的 IT 预算约 75–150 万美元，所管可观测性总支出约 1500 万美元 (含人员)，其中平台软件支出 300–400 万美元。局限：区域小、样本小；30% 是他所在 GSI 在本地区的 Datadog/可观测性业务增速，不是 DDOG 收入；他本人承认 Datadog 只是其业务的一小部分；身份是推测的。
阅读范围：全文 L1-L249 可信度：仅作局部线索
欧洲小样本渠道：2025 年 Datadog 业务约 +30%，Q4 比 Q3 (约 25%) 加速，2026 年目标 30%；他说驱动是 AI 让系统更复杂加上 AIOps。不过客户的生成式 AI 应用还在试验阶段，'应用数量 +10%/年' 是他自上而下估的。他明确说 Datadog 比别家贵，客户在看 Elastic 等开源替代、对锁定敏感，但还没看到迁出。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；30% 是该 GSI 本地 Datadog/可观测性业务的增速，不是 DDOG 的收入或消费增速；专家的 Datadog 业务体量小；L141-L155 讨论的 '去年晚些时候在美国正式发布的产品' 没有点名，结合议题推测是 Bits AI SRE；L177 'Palo Alto Networks 收购 Atmosphere' 应为 Chronosphere (转录错误)；'Altitude'/'Altana' 平台名前后不一致；员工数 32 万 (L216) 和 30 万 (L223) 前后不一；'约 25% 增长' 来自 AI SRE 降低人力 (L116) 的说法逻辑不清，属推测
A 2026-01-30 · MeritForum · 北美 GSI 专家：Datadog 业务 4Q25 +20.8% (在区间下沿)，2025 全年约 +22.8%；OpenAI 迁往 Chronosphere；Bits AI 采用率低于 10% 19 条
文件：KB/MeritForum/2026-01/2026-01-30-ddogbei-mei-zhuan-jia-4q-biao-xian-ping-wen-26nian-zhan-wang-le-guan.md
发言人：未具名；北美全球系统集成商 (GSI) 可观测性业务 (observability practice) 的负责人 (自述 'my observability practice'、'As a GSI')。汇总稿 (2026-02-03) 把他列为 '北美渠道专家B'：DDOG 相关业务约 1.1–1.15 亿美元，其中 85% 是实施部署、15% 是转售。他说曾为 OpenAI、Freepik 等 AI 原生公司做实施 (L84-L88)。局限：渠道收入多数是实施服务，不等于 DDOG 的消费；他对 OpenAI 的判断自称 'industry chatter'；Bits AI 的价格数字自称只是 'indicative'；英文记录稿，有转录错误。
阅读范围：全文 L1-L435 可信度：需交叉验证
渠道 4Q25 +20.8%，在 20.5–22.75% 区间的下沿；年底预算冲量没有落到可观测性，客户在适应 AI/agent 带来的日志暴增。2025 全年约 22.8%，1H26 预计 22–24.5%，24% 里至少 21 个百分点来自非 AI。AI 原生客户里除 OpenAI 外没有减量；OpenAI 已转向 Chronosphere 做 metrics，预计 2026Q2 前后迁完。Bits AI 在他的客户里采用率低于 10%，多数在 PoC，SRE 每周用不到 3 次，只能把 MTTU 缩短约 25%。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；英文记录稿 (非中文翻译)，有转录错误，如 'Bits AI SREM'；摘要 L32 和正文 L57 写的是 '2024 年第四季度/2024 年后两个季度'，从上下文看应为 2025 年 (口误或转录错误)；L128-L135：大单 TCV 写的是 4.2–7.6 (百万)，下文又说 '$30 million'，金额不一致；Bits AI 价格 ($4–8/host、Security $2–5/host、LLM Obs '$8-12 per 10k/50k') 专家自称只是 indicative，和官网口径不一致，不可靠；渠道增速包含实施服务收入 (汇总稿称 85% 是实施)，不等同 DDOG 消费增速；关于 OpenAI 迁移，专家自称是 'industry chatter'；L366-L371 说 OpenAI 关系 'still continuing'，又说 'they ended it' (指合作伙伴关系)，前后不一
A 2026-02-03 · MeritForum · MeritForum DDOG 4Q 渠道调研汇总：三位渠道专家 + OpenAI 迁移跟踪，预计 FY26 指引 18–19% 15 条
文件：KB/MeritForum/2026-02/2026-02-03-ddog-4qdiao-yan-zong-jie.md
发言人：MeritForum 作者 Marsh Lie 的汇总稿。综合三位渠道专家：北美专家A (DDOG 业务约 3500 万美元，客户以传统行业大企业为主；对应 2026-02-03 那篇北美渠道专家访谈)、北美专家B (业务 1.1–1.15 亿美元，85% 实施、15% 转售；对应 2026-01-30 那篇)、欧洲专家C (业务体量小；对应 2026-01-29 那篇)。'OpenAI 更新' 一节是作者自己的调研结论，没有披露来源。局限：汇总时有转述偏差 (见 issues)；OpenAI 的具体比例是作者估计。
阅读范围：全文 L1-L138 可信度：需交叉验证
三家渠道 4Q25 增速分别为 25% (A，区间低端)、20.8% (B，低端) 和 >30% (C，超预期)，北美两家都提到客户把日志转向自建湖仓、不愿重复付费。LLM Observability 起量快，Bits AI SRE 反馈平淡。作者估计 OpenAI 的 metrics 约占其 DDOG ARR 的 40%，2026 年中迁到 Chronosphere，OpenAI 的 DDOG ARR 到 2026 年底降 20% 以上；预计 FY26 首次指引 18–19% (实际给的是 18.5–19.5%)。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；汇总和原访谈不一致：专家A '5-60%客户在POC' (L78)，原访谈 (2026-02-03 北美渠道专家 L281) 是 '大约 50–60% 的客户在大规模试验 AI'，指的是泛 AI 试验，不是 Bits AI POC；专家A 日志迁移比例 '7-8%' (L44)，原访谈 L85 是 '5% 到 7%'；专家A 历史增速 '2023 16%' (L32)，原访谈 L151 是 '我们在 2022–23 年的结果为 25%'；专家B '$4-7m 大单' (L56) 和原访谈的 TCV 4.2–7.6 及 '$30 million' 口径都对不上；OpenAI 的 40% 和 -20% 是作者估计，来源未披露；'DDOG已经在3Q将Log management迁移到…OLog' 主语应为 OpenAI (表述有误)；Anthropic 部署 'database monitoring及infra monitoring' (L135) 与 Acecamp 2026-05-31 专家说的 'Anthropic和Datadog签的是RUM部分' 不一致；事后看：首次指引预测 18–19% 与实际 18.5–19.5% (见 2026-02-11 点评) 接近
A 2026-02-03 · MeritForum · 北美渠道专家：4Q 约 +25% (预期下限)，日志支出转向湖仓，2026 预期从 28–30% 下调到约 25% 17 条
文件：KB/MeritForum/2026-02/2026-02-03-ddogbei-mei-qu-dao-zhuan-jia-4qfu-he-yu-qi-yu-ji-2026lue-jia-su.md
发言人：未具名；北美渠道/系统集成方，其公司与 OpenAI 紧密合作 (L207：上次谈到 ServiceNow 时提到与 OpenAI 合作可带来约 2 亿美元)。汇总稿 (同日 DDOG 4Q调研总结) 称为 '北美渠道专家A'：DDOG 业务约 3500 万美元，客户以金融、消费、医药等传统行业大企业为主。局限：样本限于他自己的客户，偏传统行业和大企业；中文整理稿，'BCI' 等术语有转录错误；渠道收入含服务。
阅读范围：全文 L1-L308 可信度：需交叉验证
渠道 Q4 约 +25%，落在预期下限。主因是客户觉得日志被重复收费，开始把日志卸到 Snowflake/Databricks 湖仓或改用 OTel 方案，已看到 5–7% 的转移；同时基于 agent 的采集架构和存储计费也被质疑，他预计 2026 年 DDOG 定价能力下降。基础设施和 APM 仍强，Grafana 到生产规模就难用。LLM 可观测性带来的单子每笔几十万美元，新 AI 应用预算中约 8–10% 用于监控。BCI (Bits AI) 的 200 个用例只有 20–30% 开箱可用。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；'BCI' 应为 Bits AI (转录错误)；'Cloud RAM' 可能是 CloudPrem，'Active Search' 可能是 Archive Search (推测)；L151 '我们在 2022–23 年的结果为 25%' 与汇总稿的 '2023 16%' 矛盾；L281 '50–60% 的客户在大规模试验 AI' 被汇总稿写成 Bits AI POC 比例；25% 是渠道业务增速，含服务；L229 '如果我们有 2000 万美元…100 万…200 万' 的推算逻辑不清
A 2026-02-10 · S&P 业绩会转录 (Datadog IR) · Datadog Q4 2025 业绩会：非 AI 增速到 23%、FY26 只指引 18-20% (剔除最大客户 ≥20%)、不再披露 AI 原生收入占比 27 条
文件：web/DDOG_2026-02-10_Q4-2025_call.txt
发言人：管理层：Olivier Pomel (联合创始人兼 CEO)、David Obstler (CFO)、Yuka Broderick (IR SVP)；分析师：Sanjit Singh (MS)、Raimo Lenschow (Barclays)、Gabriela Borges (GS)、Ittai Kidron (Oppenheimer)、Todd Coupland (CIBC)、Mark Murphy (JPM)、Matt Hedberg (RBC)、Koji Ikeda (BofA)、Brad Reback (Stifel)、Howard Ma (Guggenheim)、Peter Weed (Bernstein)、Brent Thill (Jefferies)。身份来自 transcript 参会名单 (明示)。局限：管理层自述；两天后有 Investor Day，部分长期目标可能留到那里。
阅读范围：全文 L1-L276 (正文 L41-L273) 可信度：可作主要证据
Q4 2025 收入 $953M (+29%)，非 AI 原生客户增速加速到 23% (Q3 20%)，1 月延续；bookings $1.63B (+37%)，新签一家头部 AI 模型公司的 8 位数 land。FY26 收入指引仅 $4.06-4.10B (+18-20%)，管理层说明是把剔除最大客户的业务按至少 20% 建模，对最大客户做了"非常保守"的假设；被问 AI 原生占收入比例时拒绝披露。
问题：本会未披露 AI 原生客户占收入比例 (分析师追问被拒)，Q4 2025 比例空缺；"excluding our largest customer, grows at least 20%" 是指引建模假设，不是实际增速；CFO 在 L153 把 "non-AI or heavily diversified business" 与 "business excluding largest customer" 混用，口径不完全一致；L226 "That product is scaled at about $10 billion" 疑为转录错误 (应为 APM 市场规模或 ARR $1B)
A 2026-02-11 · MeritForum · MeritForum 4Q25 业绩点评：core (非 AI) 加速到 23%，FY26 指引 18.5–19.5% 好于担忧，利润率指引偏低 10 条
文件：KB/MeritForum/2026-02/2026-02-11-ddog-4q25ye-ji-dian-ping-corechi-xu-jia-su-2026shou-ru-zhi-yin-hao-yu-dan-you.md
发言人：MeritForum 作者 Marsh Lie 的业绩点评 (研究机构观点)；其中的公司数字转述自 4Q25 业绩和公司说法，没有附原始出处。局限：二手转述，口径以公司业绩会为准；Anthropic 是作者推测。
阅读范围：全文 L1-L47 可信度：需交叉验证
4Q25 收入 +29%，非 AI 的 core cohort 从 20% 加速到 23%，1 月继续加速；作者推算 AI 原生 cohort 环比只增长个位数。FY26 指引 18.5–19.5%，作者认为对 OpenAI 优化假设保守，并把自己的预期上调到 +26%；但 adj OPM 指引 20.6–21.5%，同比降 100–200bps。公司签了一家模型公司的八位数单子 (推测为 Anthropic)；过去一个月有超过 2000 个客户在测试或部分付费使用 Bits AI SRE。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；公司数字由作者转述，需要与 4Q25 业绩会原文核对；L28 '(3Q25%: 20%)' 排版错误；Anthropic 是作者推测，公司没有点名；APM mid-30% 的来源没有说明
B 2026-03-13 · Acecamp · 头部 IT 咨询公司 (埃森哲) 专家：AI 项目加速；称客户有用埃森哲方案替代 DataDog 等工具的趋势 4 条
文件：KB/Acecamp/2026-03/2026-03-13-tou-bu-itzi-xun-gong-si-aiye-wu-diao-yan-fy2026-q2da-ke-hu-qian-yue-yi-yuan-xiang-mu-jin-e-jia-su-zeng-chang-openaihe-zu.md
发言人：未具名；正文明确是埃森哲的人 ('埃森哲2026财年Q2'、'埃森哲倾向于使用自有的工具')，举的案例 (美孚中国区验证) 和国内模型对比说明主要是中国/亚太视角，职位不详。局限：咨询公司有推销自身方案的利益偏向；Datadog 只是被提问者点名，回答没有具体数据。
阅读范围：全文 L1-L176 可信度：仅作局部线索
主题是埃森哲的 AI 业务。涉及 DDOG 的部分：企业开始关注 token 消耗的性价比，埃森哲在项目里倾向用自有工具做成本和监控；他说存在客户用埃森哲方案替代 DataDog 等工具的趋势，但没给证据。数字员工会让已有软件 (如 ServiceNow) 的调用频率更高，但按人头的许可需求可能下降。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；Datadog 由提问者点名，专家回答很笼统，没有具体客户或金额；利益偏向：咨询商有动力夸大自有工具；中国/亚太视角；'GLM系统' 等术语可能是转录错误
B 2026-04-17 · GuidePoint · 多云伙伴渠道调研：AWS 增长部分来自观测基础设施，客户把高量数据从 Datadog 分流到 AWS 存储 4 条
文件：KB/GuidePoint/2026-04/2026-04-17-hyperscalers-aws-azure-and-gcp-consumption-2026-channel-check-with-a-multi-cloud-partner.md
发言人：Dave Sidwell，Improving Holdings 芝加哥区总裁 (President for Chicago)，多云咨询/实施伙伴 (Industry Consultant，标题明示)；主持人 Brian Oh。视角局限：只看自己公司的客户，增速是他们自己 AWS 业务的数字，不是 AWS 整体。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：需交叉验证
伙伴称 3 月季度 AWS 业务好于预期 (预测 8%，实际 10%)，部分来自为客户搭建观测基础设施。客户不想把全部数据放进 Datadog，而是把高量、不太关键的数据分流到 AWS 长期存储，'因为要付钱给 Datadog'。这是客户主动压低 Datadog 摄入量的直接证据。
问题：增速口径未说明 (同比还是环比)；说的是伙伴自己的客户，样本小
C 2026-04-25 · ThirdBridge · Elastic 前渠道销售 SVP：可观测性是竞争最激烈的市场 1 条
文件：KB/ThirdBridge/2026-04/2026-04-25-american-enterprise-infrastructure-software-space-elastic-former-senior-vp-channel-sales-at-elastic-nv.md
发言人：Elastic NV 前 Senior VP, Channel Sales (Americas 渠道，标题明示)，在企业基础设施软件和云生态有 30 多年经验；访谈实际日期 2026-02-25 (Client-led)。视角局限：渠道销售视角，不碰技术；离职时间未说明。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
专家称 Elastic 的竞争挑战主要在可观测性 (增长最快、竞争最激烈)，对手包括 Splunk、Datadog 等；Splunk 被 Cisco 收购后的替换机会 Elastic 基本错过。对 Datadog 没有具体判断。
问题：文件日期 04-25，访谈实际在 2026-02-25
B 2026-05-04 · TMTB (Substack) · TMTB 晨报 (5/4)：JPM 渠道检查称 AI 带来遥测量阶跃；UBS 财报前看多 3 条
文件：KB/Substack/2026-05/2026-05-04-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 JPM (渠道检查及一位行业联系人) 与 UBS (客户检查) 的报告摘要，二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
财报前两家卖方偏多：JPM 渠道称 1Q 遥测摄入环比加速，AI 带来 '阶跃式' 遥测量增长，但增量主要用于平台内 AI 功能；UBS 客户检查显示支出稳定增长约 8–15%、无份额流失，AI 挤压的是非核心支出而非 DDOG，也没看到 OpenAI/Anthropic 自己做可观测性。
问题：UBS '覆盖约 88% ARR' 字面不合理 (可能指受访客户支出占其 ARR 比例)，口径未明；JPM 所说 '增量摄入主要用于平台内 AI 功能' 意味增长可能来自 DDOG 自身 AI 功能而非客户负载，方向含糊
A 2026-05-05 · Funda (Substack) · Funda 1Q26 财报前渠道调研：核心稳、OpenAI 迁移推迟到 4Q26、Bits AI 仍早期 23 条
文件：KB/Substack/2026-05/2026-05-05-preview-ddog-1q26-stable-core-openai-migration-pushed-out-bits-ai-contribution-still-early.md
发言人：FundaAI (Substack 独立卖方式研究) 的季度渠道调研摘要：3 位前/现 Datadog 销售与渠道专家 (北美渠道专家 A；前北美 Commercial 销售经理 B，管 15 AE + 6 SDR；前 UK/Ireland 销售 C，约 200 个新客户 onboarding) + 1 家北美大客户 D (2025 年 DDOG 支出约 $1.2m)。身份为 Funda 明示，均匿名。二手转述，看不到底稿。局限：渠道专家口径是自己 book 的增速而非 DDOG 整体；前员工 (B、C) 信息可能滞后、且 B 对 Anthropic ARR 的说法属听闻；D 是单一客户样本；OpenAI/Anthropic 收入占比是 Funda 推算，未披露推算方法。
阅读范围：全文 L1-L201 可信度：需交叉验证
财报前 Funda 认为核心业务稳但分化：AI 原生 (含 AWS 生态 AI 初创) 超预期，传统/中型企业因 AI 挤占预算而谨慎；Bits AI 企业端 ROI 算不过来。OpenAI 迁往 Chronosphere 推迟到约 4Q26，Funda 估 OpenAI 占 2026 收入 7–8%、2027 降到低个位数；Anthropic 在 DDOG 的 ARR 季增 15–20%，年底有望超 $100m，但也用 ClickHouse，长期可能走 OpenAI 的路。预计 1Q 实际增速 31–32%、2Q 指引 24–25% (卖方约 20%)。
问题：渠道专家 A 的 '1Q26 ~21%' 是他自己 book 的增速，不是 DDOG 整体；8 月 preview 中同一专家改称 1Q 最终约 26%；OpenAI 占收入 7–8% (2026)/低个位数 (2027) 是 Funda 推算，未披露推算方法或数据源；前销售 B 说 Bits AI ARR ~$100m 可贡献 '5–15% incremental growth'，以 DDOG 约 $4.3B 收入计 $100m 约 2pp，口径对不上 (篇内矛盾)；大客户 D '~300+ investigations / month at ~US$500 each' 与专家 C 的 '$500/月 20 次调查' 定价口径冲突，可能是转述错误；专家 C 'AI 原生客户 ARR 翻倍/3–4x' 是理想情形 (idealized scenario)，不是观察
A 2026-05-07 · FactSet 业绩会转录 · Datadog Q1 2026 业绩会：收入 +32%、非 AI 加速到 mid-20s，对最大客户加倍保守，FY26 指引上调到 +25–27% 38 条
文件：KB/FactSet/2026-05/2026-05-07-corrected-transcript-datadog-inc-ddog-us-q1-2026-earnings-call-7-may-2026-8-00-am-et.md
发言人：Olivier Pomel (联合创始人兼 CEO，现任)、David Obstler (CFO，现任)、Yuka Broderick (IR 高级副总裁)；身份由开场白明示。卖方：Mark Murphy (JPM)、Sanjit Singh (MS)、Raimo Lenschow (Barclays)、Gabriela Borges (GS)、Karl Keirstead (UBS)、Fatima Boolani (Citi)、Kirk Materne (Evercore)、Patrick Colville (Scotiabank)、Peter Weed (Bernstein)、Gregg Moskowitz (Mizuho)。视角局限：公司自己口径，倾向正面叙述；用量数字只给增速；最大客户不点名；转录无发言人标注
阅读范围：全文 L1-L365 可信度：可作主要证据
Q1 收入 $1.01B (+32%，上季 29%)，非 AI 客户增速从 19%→23%→mid-20s，NRR 回升到 low-120s、GRR mid-to-high 90s，新客户 bookings 翻倍，支持 '加速是广泛的' 说法；AI native 组有 22 家 >$1M、5 家 >$10M，新签两家 hyperscaler AI 实验室。但公司连续两季对最大客户单独加大折扣，AI 产品只给用量倍数不给收入，费用增速 31% 吃掉经营杠杆，分析师也没问到具体竞品。
问题：转录稿无发言人标注 (FactSet 文本被压成连续段落)，speaker 按上下文推定；Lenschow 追问的回答中 Pomel 与 Obstler 交替，L237-242 归属不完全确定；非 AI 增速 'mid-20s' 只给区间不给精确数；Q1 用 'excluding the AI natives' 口径，Q2 改说 'excluding AI customers'，AI 组定义在 Q2 扩大 (含 hyperscaler 内部 AI 实验室)，两季口径可能不完全可比；AI 产品用量只给增长倍数 (2x/3x/4x/12x)，没有绝对值、没有收入/ARR 贡献，也没有 Bits AI 定价；Keirstead 提到 '其中一个研究实验室 Q4 已续约'，Pomel 说 'Q4 签下的客户在 Q1 贡献收入最多'——是否就是最大客户未说明；分析师没有直接问 Chronosphere、OTel、ClickHouse 等具体竞品；只有 Lenschow 泛问开源、Weed 问 hyperscaler 自建；管理层回答也未点名竞品；SBC、回购、GAAP 利润在电话会里均未提及；只有稀释股数指引；AI native >$1M/>$10M 客户数为收入/年化消费口径，未给 AI 组整体收入占比或最大客户占比
A 2026-05-07 · Funda (Substack) · Funda 1Q26 财报点评：指引大超预期，2H 增长断崖担忧解除 11 条
文件：KB/Substack/2026-05/2026-05-07-review-ddog-1q26-big-guidance-upside-surprise-no-growth-cliff-concern-anymore.md
发言人：FundaAI 自有财报点评 (Substack)，混合转述公司披露数据 (收入增速、非 AI 增速、AI 原生 >$1M 客户数) 与 Funda 自己的推断 (Anthropic 驱动、AI 原生 QoQ 17%、折扣导致毛利率下降)。二手转述，看不到底稿，看不到公司原始 transcript；买方预期数字是 Funda 口径。
阅读范围：全文 L1-L52 可信度：可作主要证据
1Q26 收入 +32%，比买方预期高 1pp，核心非 AI 增速从 4Q25 的 23% 升到 mid-20s；2Q 指引 +30% 比市场高 6–7pp，FY26 指引 26–27%，按惯常 4–5pp beat 推算 2Q 约 35%，解除 OpenAI 迁移带来的 2H 断崖担忧。Funda 认为 AI 原生动能受 Anthropic ARR 加速带动，并推断毛利率下降来自给大客户的折扣。
问题：L34 写 '4Q adj OPM'，上下文是 1Q26，疑为笔误；'2Q 实际可达 35%' 是按惯常 beat 外推，不是指引；Anthropic 驱动是 Funda 推断，管理层原话是 '不集中于某一客户'
B 2026-05-07 · TMTB (Substack) · TMTB 收盘 (5/7)：DDOG 1Q 后 +30%，FY26 指引 26% 远超街 20% 3 条
文件：KB/Substack/2026-05/2026-05-07-tmtb-eod-wrap.md
发言人：TMT Breakout (TMTB) 日报，作者市场评论 + TMTB Slack 成员 Josh S 的评论，转述业绩会要点。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
DDOG 1Q 后当日 +30%：收入增速从 29% 加速到 32%，FY26 指引 26% 远高于街上约 20%，指引上调幅度大于 1Q beat 加 2Q 指引超额之和，缓解 OpenAI 断崖担忧；管理层称训练正成为市场，hyperscaler 在已有自建工具的情况下也用 DDOG。TMTB 认为 DDOG 和网络安全因 AI 需求受益，和多数应用软件不同。
问题：Josh S 评论 '只差把 anthropic 说 50 遍' 暗示市场认为加速主要来自 Anthropic，但这是聊天室观点
B 2026-05-07 · TMTB (Substack) · TMTB 晨报 (5/7)：DDOG 1Q26 数字与盘前 +28% 5 条
文件：KB/Substack/2026-05/2026-05-07-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，汇总 DDOG 1Q26 公布数字与街上预期 (Street 口径来源未注明)，转述 CEO 评论。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
1Q26 收入 $1.006B (+32%) vs 街 $960M，billings $1.03B (+37%)，OPM 22.2%，FCF $289M；2Q 指引 $1.07–1.08B vs 街 $994M，FY26 $4.30–4.34B vs 街 $4.12B；NRR 从约 120% 回到 low-120s，$100K+ 客户 4,550 家。CEO 称加速在 AI 与非 AI 都广泛。TMTB 估若继续 5–6pp beat，年底可到 high-30s。
C 2026-05-11 · Funda (Substack) · Funda 周报 (5/11)：重述 DDOG 1Q26 点评 + 资金从软件转向半导体 2 条
文件：KB/Substack/2026-05/2026-05-11-weekly-why-memory-cpu-rally-accelerated-and-outperformed-optics-smtc-amd-deep-dive-ddog-app-lite-cohr.md
发言人：FundaAI 周报，DDOG 部分是 5/5 preview 和 5/7 review 的摘要重述；仓位部分引用 Coatue 演示材料。二手转述，看不到底稿。
阅读范围：全文 L1-L183 (DDOG 相关仅 L68-L108，其余为半导体/其他公司) 可信度：仅作局部线索
DDOG 段落只是重述 1Q 点评：2Q 指引 +30% (按 5pp beat 约 35%)、FY26 上调到 26–27%，'OpenAI 断崖' 叙事退场。开头提到美国基金从软件转向半导体 (Coatue 半导体+基础设施多头敞口从 35% 升到 58%)，是 DDOG 所在软件板块的仓位背景。
问题：DDOG 内容与 5/7 review 重复，非新增信息
C 2026-05-13 · TMTB (Substack) · TMTB Sohn 大会纪要：Whale Rock 与 Maplelane 均称 DDOG 是赢家 2 条
文件：KB/Substack/2026-05/2026-05-13-tmtb-sohn-conference-tech-pitches-app-u-rddt-googl-amat-memory-more.md
发言人：TMT Breakout (TMTB) 日报，转述 Sohn 投资大会上 Whale Rock (Alex Sacerdote) 和 Maplelane (Leon Shaulov) 的 AI panel 观点。买方观点，TMTB 摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Sohn 大会 AI panel 上 Whale Rock 和 Maplelane 都把 Datadog 列为赢家，称其是 '一切的明确赢家'；同时指出软件已不再是 CIO 首要事项，token 和 AI 才是。属于 DDOG 买方共识/拥挤度的信号。
问题：panel 发言未分清是哪位投资人说的；nature 实为买方观点 (大型对冲基金)，体现仓位共识
A 2026-05-19 · FactSet 会议转录 · JPM TMC 大会：Pomel 谈全面加速、超大厂买单、训练成新市场、Bits 按调查次数收费 16 条
文件：KB/FactSet/2026-05/2026-05-19-corrected-transcript-datadog-inc-ddog-us-jp-morgan-global-technology-media-and-communications-conference-19-may-2026-1-3.md
发言人：出席人：Olivier Pomel，Datadog CEO 兼联合创始人 (主持人开场明示)。主持人为 JPM 分析师，FactSet 标为 Unverified Participant、未署名 (推测为 JPM 软件分析师)。视角局限：CEO 投资者会议发言，倾向强调利好；没有新的量化数据，多为定性说法；转录有明显错字 (如 '23 AI natives'、'five coding'=vibe coding、'authorization'=automation)。
阅读范围：全文 L1-L272 可信度：可作主要证据
Pomel 称 Q1 加速遍及所有客户群，非 AI 客户加速主要来自上云和现代化、多产品渗透，企业自己上 AI 生产仍“很小”。超大厂开始买 Datadog (训练场景)，被他用来说明自建不划算，但他同时说超大厂“善变”、不是商业模式。Bits AI SRE 目前按调查次数收费，长期模式未定。他也预警 AI 支出“合理化”可能比预想来得早。
问题：'23 AI natives, spending more than $1.5 million, spending more than $10 million' 明显是转录错误，数字不可用；'five coding'=vibe coding、'authorization'=automation、'financial models'=foundation models 等转录错误；本场未谈 SBC、回购、单价/折扣；主持人姓名未署名
A 2026-05-19 · TMTB (Substack) · TMTB 摘录 JPM TMT 大会 DDOG CEO Pomel 发言 12 条
文件：KB/Substack/2026-05/2026-05-19-tmtb-jpm-tmt-conf-key-quotes-intc-alab-anet-ddog.md
发言人：TMT Breakout (TMTB) 日报，摘录 Datadog CEO 兼联合创始人 Olivier Pomel 在 JPM TMT 大会 (2026 年 5 月) 的发言原话 (明示)。是管理层一手说法但经 TMTB 挑选、整理成段，二手转述，看不到底稿；看不到分析师问题原文，可能遗漏不利问答。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：可作主要证据
Pomel 称加速遍及所有业务：AI 原生和更有意思的非 AI 部分都在加速，原因是企业为 AI 加快现代化、上云仍早；Gartner 口径 ITOM 份额仅 13.6%；多家 hyperscaler 在训练场景选择 DDOG，做后训练的公司从 5–10 家增到 50–100 家；生产力越高复杂度越高，前 8–10 家 coding 公司都用 DDOG；客户要求 Bits AI 走向自动修复并跨其他系统调查；研发约占收入 30%，未来会更多花在 token/GPU 而非人力。
问题：TMTB 只摘录要点，没有分析师问题原文；'5x' 一句缺上下文 (疑指长期收入空间)
B 2026-05-27 · TMTB (Substack) · TMTB 晨报 (5/27)：BofA 基于 Cognizant 渠道检查上调 DDOG 目标价到 $260 2 条
文件：KB/Substack/2026-05/2026-05-27-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (渠道检查对象为 Cognizant)。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
BofA 把 DDOG 目标价从 $225 上调到 $260，Cognizant 渠道检查显示可观测性、安全需求改善，AI 推高应用复杂度、企业现代化推动对一流监控平台的需求。
A 2026-05-28 · FactSet 会议转录 · Bernstein SDC：Pomel 回应“agent 取代 Datadog”、LLM obs 竞争、BYOC 与降本机制 18 条
文件：KB/FactSet/2026-05/2026-05-28-corrected-transcript-datadog-inc-ddog-us-bernstein-strategic-decisions-conference-28-may-2026-1-30-pm-et.md
发言人：出席人：Olivier Pomel，Datadog CEO 兼联合创始人 (明示)。主持人 Peter Weed，Bernstein 软件分析师 (覆盖 IT 基础设施/开发者软件，自述前 Microsoft 产品经理、McKinsey)。视角局限：CEO 宣讲场合；基本无新数字；部分转录错误 (seed-based=seat-based、CECL=CISO、CDUs=CPUs、financial models=foundation models)。
阅读范围：全文 L1-L345 可信度：可作主要证据
Pomel 正面回应“LLM agent 能否替代 Datadog”：数据量多 5-6 个数量级、需实时，LLM 做不了也太贵；但他也承认 LLM 通过 MCP 工具取数“可能得出类似结论”。他说加速“在数学上主要来自现有产品”，前 20 家 AI 公司中 14 家用的也是传统产品，说明新 AI 产品尚未贡献。他把 BYOC、采样、成本反馈列为客户控制观测支出的三大机制，自我蚕食风险隐含其中。
问题：转录错误：seed-based=seat-based，CECL=CISO，CDUs=CPUs，financial models=foundation models；联邦 1% 是分析师转述 CFO，非本场管理层原话；未谈 SBC、回购、最大客户、单价
C 2026-05-28 · TMTB (Substack) · TMTB 晨报 (5/28)：SNOW 大涨带动 DDOG +4%；此前买方多偏好 DDOG 1 条
文件：KB/Substack/2026-05/2026-05-28-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，作者对 SNOW 财报的评论中顺带提及 DDOG 仓位。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
SNOW 财报大超后 DDOG 同情性 +4%；TMTB 提到此前 SNOW 仓位偏轻，很多人更偏好 DDOG——说明 DDOG 在消费型软件中仓位较重。
C 2026-05-29 · TMTB (Substack) · TMTB 晨报 (5/29)：MDB 点评中称 DDOG/SNOW 的 AI 消费叙事更好 1 条
文件：KB/Substack/2026-05/2026-05-29-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，MDB 财报点评中顺带提及 DDOG。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
TMTB 在 MDB 点评中认为 DDOG 和 SNOW 在 AI 消费领域的叙事更好，AI 收入加速会先出现在可观测性、数据仓库等环节。
B 2026-05-31 · Acecamp · 'AI 观测大厂' (推测 Elastic) 专家：模型厂的观测需求以结构化指标为主、偏向 ClickHouse，Anthropic 签的是 Datadog 的 RUM 部分 7 条
文件：KB/Acecamp/2026-05/2026-05-31-aiguan-ce-da-han-yu-mo-xing-han-he-csphe-zuo-fang-xiang-jing-zheng-ge-ju-he-bi-lei-pan-duan-guan-ce-sou-suo-he-an-quan-s.md
发言人：未具名；标题是 'AI观测大厂'，内容几乎都以 Elastic 内部视角回答 (清楚 Elastic 上月刚出的指标数据结构、Elastic Cloud 收入占比超过 50%、内部 AI 渗透率口径)，推测为 Elastic 的现任或前任员工。局限：Datadog 的竞争对手视角，可能低估 Datadog；Anthropic 合同信息自称来自公开信息。
阅读范围：全文 L1-L97 可信度：需交叉验证
他认为模型厂商运维的是超大 GPU 集群，需求主要是结构化指标和格式单一的日志，所以选 ClickHouse 这类列存，不是 Elastic。Datadog 被用在用户体验监控 (RUM)，Anthropic 和 Datadog 签的是 RUM 部分 (八位数合同)。有自己云的公司 (GCP、AWS、阿里) 不用 Datadog。企业在观测上的预算不如安全多。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；'Anthropic 签的是 RUM 部分' 与 MeritForum 汇总稿 'Anthropic 小范围部署 database monitoring 及 infra monitoring' 矛盾；'两个PCV过亿的客户' 口径不明 (可能指 TCV 超过 1 亿美元)；'AI用力渗透率' 应为 'AI 用例渗透率' (转录)；'opencloud' 等术语有转录错误；身份为推测，属竞争对手视角
B 2026-05-31 · Acecamp · Acecamp 作者 2030 收入拆解：DDOG 2030 年约 175 亿美元，OpenAI 自建是 outlier，但 Anthropic/xAI 可能在 2027–29 年效仿 8 条
文件：KB/Acecamp/2026-05/2026-05-31-agentic-ai-shi-dai-de-jiu-da-ji-jian-wang-an-ruan-jian-long-tou-xia-pian-zi-xia-er-shang-chai-jie-2030-nian-shou-ru-shui.md
发言人：Acecamp 作者撰写的分析文章 (不是专家访谈)，作者未具名。份额数据自称 '绝大多数为笔者估算'，市场规模引自 Gartner/IDC/MarketsandMarkets 等。局限：自上而下估算，DDOG 产品线拆分表是图片，无法读取。
阅读范围：关键词检索 (Datadog/DDOG/observ/观测/监控/Chronosphere/ClickHouse/Grafana/Dynatrace/Splunk/SIEM/日志) + 精读 L1-L40、L110-L160 (含 DDOG 章节 L128-L153)、L300-L330、L360-L384；其他公司章节未逐段读 可信度：仅作局部线索
作者以 Q1 2026 收入 10.06 亿美元 (+32%) 为基准，推到 2030 年 175 亿美元 (年化 34%)。他把 OpenAI 自建看作 '超大规模导致第三方监控不经济' 的 outlier，但预计 Anthropic、xAI 会在 2027–29 年效仿，所以把超大 AI 公司约 20–30 亿美元的需求算作 DDOG 拿不到。他认为 DDOG 的核心客户应是部署 AI agent 的传统企业；安全业务没有 CISO 关系，是短板。
问题：作者分析文章，不是专家纪要；份额为作者估算；Observability 市场规模口径前后不一：L116 是 $80B，L275 是 '$60B 同前述'；当前份额 13.6% 到 2030 年 30% 的假设很激进；DDOG 产品线拆分表是图片 (L134)，无法核对
A 2026-06-03 · FactSet 会议转录 · BofA 全球科技大会：CFO Obstler 谈上云比例、训练客户、Bits/LLM obs、CloudPrem 15 条
文件：KB/FactSet/2026-06/2026-06-04-corrected-transcript-datadog-inc-ddog-us-bank-of-america-global-technology-conference-3-june-2026-3-30-pm-et.md
发言人：出席人：David M. Obstler，Datadog CFO (明示)。主持人 Koji Ikeda，BofA 软件分析师。视角局限：CFO 口径，偏谨慎 (“we generally don't overpromise”)；有转录错误 (如 '$2 billion of Datadog' 应为 $2 million、'we don't charge by fee' 应为 seat)；部分数字记忆不清 (CFO 自称 'I'm not sure I have all of this exactly right')。
阅读范围：全文 L1-L348 可信度：可作主要证据
Obstler 给出上云比例：只有 25-30% 工作负载在云上，70%+ 未上云，用来支撑非 AI 加速的长期性；但他也承认消费型软件会有“优化期”。训练客户 (一家超大厂 + 一家大科技公司) 被定性为背书，收入仍主要来自生产推理。Bits SRE 2,000 客户、10 万+ 调查；LLM obs“开始有收入”但未量化；CloudPrem 目标是对客户存数据位置“无所谓”，隐含自我蚕食。
问题：'$2 billion of Datadog' 应为 $2 million (转录或口误)；'over 650 ... spending over $10 million in 10-plus products' 句子混乱，不可直接引用；'we don't charge by fee' 推测为 seat；CFO 自称数字不一定准确 (L78)；未谈 SBC、回购、最大客户
B 2026-06-04 · Merit 专家电话 (音频转写) · 网络安全软件调研 I：AI 对安全支出的影响，顺带谈 Chronosphere/Datadog 定位 8 条
文件：KB/2026-06-04/260604 - Merit 网络安全软件跟踪调研Ⅰ(英文)(2026-06-0.md
发言人：未具名专家，Accenture (埃森哲) 前任人员 (身份由同场 MeritForum 06-04 标题'埃森哲离职专家'推得；本篇无自我介绍)。以'my portfolio'口吻谈客户，推测为咨询/集成侧的安全业务负责人。视角局限：咨询商视角，只见自己客户组合；主题是 Palo Alto/CyberArk 安全，DDOG 为顺带；Chronosphere 数字为举例性量级。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：需交叉验证
专家认为 AI 工作负载使遥测量暴涨 (传统应用约 100GB/天，AI 应用 4-5TB/天)，可观测性是 AI 的间接赢家；Datadog 核心可观测性地位稳固，未见客户因 AI 从 Datadog 迁往 Palo Alto，但在安全分析/SIEM 上与 Palo Alto 重叠增加，且 AI 带来的平台整合压力是 Datadog 的弱点。
问题：Chronosphere 遥测数字为举例性口径，未说明来源；同一通话的 MeritForum 版本把传统应用翻成'不到几个 GB'、把 Chronosphere 写成'直接的 AI 胜出者'、把 Datadog/Chronosphere 写成'敏感度较低'，与英文原文 (100 gigs；indirect winner；'From observability layer perspective, Datadog, Chronosphere' 列为第二梯队) 不符，以本篇英文为准
C 2026-06-04 · MeritForum · 埃森哲离职专家：AI 推动安全边界转移 (与 Merit 06-04 同一通话的整理版) 1 条
文件：KB/MeritForum/2026-06/2026-06-04-ai-sen-zhe-chi-zhi-zhuan-jia-ren-gong-zhi-neng-tui-dong-an-quan-bian-jie-cong-chuan-tong-de-yong-hu-he-she-bei-zhuan-yi.md
发言人：埃森哲 (Accenture) 离职专家 (标题明示)，具体职位未说明。同 Merit 06-04 音频那篇。视角局限同前：咨询商的客户组合视角。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
与 Merit 06-04 是同一通话，内容重复：可观测性随 AI 遥测增长受益，Datadog 与 Palo Alto 在大客户中共存。译文在若干处与英文原文不一致，以 Merit 06-04 英文为准。
问题：L174 把传统应用遥测写成'不到几个 GB'，英文原文为 100 gigs/天；L98 把 Chronosphere 写成'直接的 AI 胜出者'，原文为 indirect；L300 写 Datadog/Chronosphere'敏感度较低'，原文把它们列在可观测层，不属低敏感度；与 Merit 06-04 重复，不要重复计数
C 2026-06-05 · Merit 专家电话 (音频转写) · 网络安全软件调研 II：Palo Alto 增速与 Chronosphere 是否冲击 Datadog 2 条
文件：KB/2026-06-05/260605 - Merit 网络安全软件跟踪调研Ⅱ(英文)(2026-06-0.md
发言人：未具名专家，Capgemini 前任人员 (由同场 MeritForum 06-05 标题'Capgemini 离职专家'推得)；以'we/our customers'口吻谈 Palo Alto 增长预期，推测为 Palo Alto 渠道/集成伙伴侧负责人。视角局限：伙伴视角只看到自己的客户群；对 Datadog 仅顺带评论。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
专家称未见客户大规模从 Datadog 迁往 Palo Alto/Chronosphere，Chronosphere 更聚焦安全相关可观测性，不与 Datadog/Dynatrace 的 APM/RUM 广泛业务直接竞争；Splunk 按 GB 收费被认为过贵。
问题：Chronosphere 定位为'安全相关可观测性'与市场一般认知 (云原生指标/日志平台、OpenAI 迁移对象) 不一致，可能是专家对产品理解有限
C 2026-06-05 · MeritForum · Capgemini 离职专家：安全定价转向消耗制 (与 Merit 06-05 同一通话的整理版) 1 条
文件：KB/MeritForum/2026-06/2026-06-05-capgemini-chi-zhi-zhuan-jia-an-quan-chan-pin-ding-jie-mo-xing-xun-su-yan-bian-chuan-tong-ji-yu-xi-wei-he-ji-yu-she-bei-d.md
发言人：Capgemini 离职专家 (标题明示)，具体职位未说明。同 Merit 06-05 音频那篇。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
与 Merit 06-05 内容重复：未见大规模客户从 Datadog 迁移，Chronosphere 不是 Datadog 在 APM 和追踪上的直接竞争对手。
问题：与 Merit 06-05 重复，不要重复计数
B 2026-06-06 · Acecamp · '北美 AI 搜索大厂' (推测 Elastic) 专家：Datadog 受益于上云；Chronosphere 靠采样降本和 GPU 方案赢得 OpenAI/Anthropic 6 条
文件：KB/Acecamp/2026-06/2026-06-06-bei-mei-aisou-suo-da-han-chan-pin-clihua-qu-shi-xiao-shou-he-kai-yuan-ce-lue-xiang-2bbian-xian-ce-zhong-aiguan-ce-xu-qiu.md
发言人：未具名；以 Elastic (ES) 内部视角回答 (清楚 ES 组织架构调整、API 文档策略、Jina 整合)，推测为 Elastic 现任或前任员工，带明显中国区视角 (国内定价、阿里/腾讯 ES 服务)。局限：竞争对手视角；对 Datadog 的判断是间接的。
阅读范围：全文 L1-L99 可信度：仅作局部线索
他认为 Datadog 是云上的标准化开箱方案，AI 推动新项目上云，会让 Datadog 的云上观测收入增加。Chronosphere 的优势在于入库前过滤、聚合、采样降低数据量，以及原生 GPU 指标方案，因此得到 Anthropic、OpenAI 认可。模型大厂是 C 端场景，特别在意 RUM，ES 在这块不如 Datadog。另外他说产品都在 API/CLI 化、面向 agent，UI 重要性下降。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；'IUM' 应为 RUM (转录错误)；'Datadog Spark' 指代不明；'Datadog 只是云上的方案，对云下没有产品' 不准确 (Datadog agent 可以部署在本地主机上)，属专家判断偏差；中国市场视角，身份为推测
C 2026-06-09 · TMTB (Substack) · TMTB 晨报 (6/9)：DASH 大会日程；WFS 称 SNOW 与 DDOG 估值差缩小 1 条
文件：KB/Substack/2026-06/2026-06-09-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，日程提示 + 转述 Wells Fargo 对 SNOW 的观点。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
只提到当日 DDOG DASH 大会，以及 Wells Fargo 认为 SNOW 与 DDOG 的估值差在缩小 (DDOG 作为估值参照)。
B 2026-06-10 · ThirdBridge · 大型银行平台工程负责人 (PNC) 谈安全与观测：Dynatrace 更优、日志计价伤大企业、续约杠杆在客户 5 条
文件：KB/ThirdBridge/2026-06/2026-06-10-cybersecurity-sector-update-customer-perspective-demand-trends-in-edr-observability-e-mail.md
发言人：匿名专家 ([Expert])。Third Bridge 主持人介绍为 'former Head of CCoE [Cloud Centre of Enablement] and Platform Engineering at PNC Financial Services'；但专家自述 '我是' PNC Cloud CoE 和平台工程负责人、在该行两年 (L338，现任/前任表述矛盾)。负责容器、中间件、数据平台及平台安全 (本地和云)；上一家机构曾把 Splunk 整合迁移到 Google Chronicle (L124)。视角局限：受严格监管的大型银行买方，偏 '最佳单品 + 多厂商冗余' 策略，不代表一般企业；角色偏平台/安全，不直接管观测账单细节；对 Datadog 的评价较笼统、无数字。
阅读范围：仅读观测/Datadog 相关段落：Key Insights L1-L17、Q1 L26-L38、Q7 L76-L82、Q12-Q15 L112-L145、Q21-Q24 L186-L222、Q26 L230-L235，以及第二份副本中的 Introduction L327-L339；EDR/邮件安全问题 (Q2-Q6、Q8-Q11、Q16-Q20、Q25) 跳过；L236-L564 为全文重复副本 可信度：仅作局部线索
银行买方认为在复杂遗留金融环境中 Dynatrace 优于 Datadog (自动拓扑、可靠性)，Datadog 不会取代 Splunk 做安全 SIEM。对价格面最有用的是：按日志量计价对大企业不利，客户更想要按席位计价；续约谈判杠杆在客户一方，集中在一两家厂商时随量增可压价。但他也说 AI/agent 让安全和观测预算 '显著增长'。
问题：主持人称 'former Head'，专家自述现任，身份时态矛盾；文件中全文重复出现两次 (L1-235 与 L236-564)；专家对 Datadog 的评价笼统，无账单或折扣数字；Key Insights 中 'Dynatrace justifies its premium pricing over Datadog' 是对专家观点的概括
A 2026-06-10 · Acecamp · 某 Datadog 用户的运维负责人：非核心监控迁往 Prometheus/Grafana，Datadog 上的指标和日志量减少、支出不再增加 13 条
文件：KB/Acecamp/2026-06/2026-06-10-datadogguan-ce-jing-du-he-yi-cao-zuo-xing-duan-qi-nan-bei-qu-dai-xiang-kai-yuan-fang-an-de-qian-yi-ce-zhong-fei-he-xin-s.md
发言人：未具名；某使用 Datadog 的公司的运维/监控负责人或资深 SRE (从回答内容推断：讲自家公司怎么规划监控迁移、用 Claude Code 写代码、'向总部申请增加人员')。提到 '国内的开源Agent'、GDPR，推测是外企中国分支或出海企业的技术团队。局限：单一客户样本；公司规模、行业和 Datadog 账单金额都没披露；可能是中国区团队，与北美大企业情况未必一样。
阅读范围：全文 L1-L118 可信度：仅作局部线索
单一客户样本，Datadog 与 Prometheus/Grafana 混用：Datadog 计费复杂、成本高，AI 降低了运维开源方案的门槛 (写 YAML、做 dashboard)，所以非核心监控逐步迁往开源，'整体趋势是向开源方向倾斜'。Datadog 上的指标数、dashboard、日志量都减少了很多，支出没有增加。另一面，他认为 Datadog 的 agent 采集精度、APM、合规背书短期难以替代；token 监控、幻觉监控暂时用不上。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；单一客户样本，公司和账单规模都没有披露；篇内矛盾：L36 说迁移的是 '非核心' 监控 (标题也这样写)，L102 又说 '核心指标等内容其实在减少，很多都迁移到了开源的 Grafana'；很可能是中国区团队 ('国内的开源Agent')，代表性有限
B 2026-06-10 · TMTB (Substack) · TMTB 晨报 (6/10)：Wedbush DASH 后上调目标价到 $260；BofA 列 DDOG 为下半年首选 3 条
文件：KB/Substack/2026-06/2026-06-10-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Wedbush (DASH 2026 后) 与 BofA (基础设施软件 'Fab Five') 报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Wedbush 在 DASH 2026 (100+ 新品) 后上调目标价到 $260，称 DDOG 正变成全栈 AI 运维平台，提到 Bits AI、Agent Console、AI Guard 等客户采用强。BofA 把 DDOG 列为 2H26 首选基础设施软件之一。
问题：'强客户采用' 未量化，DASH 刚发布即称强采用，与 Funda 渠道 '多在 POC' 的说法不一致
B 2026-06-11 · TMTB (Substack) · TMTB 晨报 (6/11)：BNP DASH 后客户调研偏正面 2 条
文件：KB/Substack/2026-06/2026-06-11-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 BNP Paribas DASH 2026 后客户调研摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
BNP 在 DASH 与客户交流后更有信心，客户对 Bits AI、AI Guard、数据库/网络监控兴趣强，多位客户视 AI 为可观测性支出的催化剂而非替代。
问题：'兴趣' 不等于付费
C 2026-06-15 · FactSet 会议转录 · 2026 年度股东大会：程序性会议，简单多数投票提案未通过 1 条
文件：KB/FactSet/2026-06/2026-06-15-raw-transcript-datadog-inc-ddog-us-annual-general-meeting-15-june-2026-2-30-pm-et.md
发言人：主持：Olivier Pomel，CEO (自我介绍明示；FactSet 标签把首段归于 Operator)。在场：CFO David Obstler、总法律顾问 Kerry Acocella、IR 负责人 Yuka Broderick、Deloitte 审计。股东提案人 John Chevedden。视角局限：纯程序会议，无经营问答 (无股东提问)。
阅读范围：全文 L1-L73 可信度：仅作局部线索
四位一类董事连任、高管薪酬咨询投票和审计师通过；股东简单多数投票提案未通过。提案人称内部人持股 7% 却控制 70% 投票权。没有股东提问，和 thesis 基本无关。
问题：原始 (未校对) 转录；7%/70% 来自股东提案人，非公司披露；无股东问答
B 2026-06-15 · TMTB (Substack) · TMTB 晨报 (6/15)：Truist 升级 DDOG 到买入，目标价 $300 3 条
文件：KB/Substack/2026-06/2026-06-15-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Truist 升级报告摘要 (基于客户与伙伴交流)。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Truist 升级到 Buy、目标价 $300，认为企业 AI 采用的紧迫性压过 AI 优化担忧，非 AI 原生企业部署 agent 需要更多遥测；对与前沿 AI 实验室关系的持久性更有信心。
A 2026-06-16 · Merit 专家电话 (音频转写) · Merit Datadog 产品进展调研 I：AI 原生客户消费 +35–50%、传统客户 +10–20%，部分日志迁出与定价压力 18 条
文件：KB/2026-06-16/260616 - Merit Datadog产品进展调研I(2026-06-16.md
发言人：未具名专家。音频原文没有专家介绍 (开头只有合规声明，L471)；身份只能从同源的 MeritForum 纪要标题推出："北美软件渠道商离职专家" (M1 L2)，即北美软件渠道商/经销商的前员工，任职年份和具体岗位均未披露。说话用 "from what I have observed"、"customer conversations"，像是从渠道/客户侧观察，不是 Datadog 内部人。视角局限：只看到自己(前)所在渠道覆盖的客户，样本量和客户结构不明；离职后信息可能滞后；回答高度结构化、偏通用框架 (类似咨询话术)，具体客户案例几乎没有；数字多为区间且未说明口径。
阅读范围：全文 L1-L471 (L13-247 为 AI 摘要，L249-465 为中文整理转译，L471 为中英交传原始转写单行，quote 以 L471 英文原话为主) 可信度：需交叉验证
专家称其观察到的客户 Datadog 消费 Q1 同比约 mid-to-high teens，Q2 温和加速到 high teens/low 20s；AI 原生客户 +35% 到 50%以上，传统客户 +10–20%，这就是 '35–50%' 的出处，属专家个人观察、无样本量。同时明确讲了反方：FinOps/日志优化/低价值遥测迁到更便宜平台在拖慢增长，客户比 2–3 年前更价格敏感，部分把日志、基础设施监控迁到 Elastic/OpenSearch/Prometheus，云厂商在挤压低端和商品化的基础设施监控/日志；Datadog 用批量折扣、企业协议、打包和成本控制工具 (采样、过滤、保留策略) 应对。
问题：音频原文没有专家自我介绍，身份只来自 MeritForum 标题 (北美软件渠道商离职专家)，无法确认具体公司/年份；'35–50%' 是专家个人观察的 AI 原生客户 Datadog 消费同比 (问题问的是 Q2)，没有样本数、客户数或加权口径；不是公司披露；且是 '客户消费' 不是 Datadog 营收；专家给的整体客户 Q1 mid-high teens / Q2 high teens-low 20s 远低于 Datadog 公司报告的 Q1 +32%、Q2 +36% 营收增速，说明专家样本偏传统客户或口径不同；转写错字：Data Log / Datado / Terado / Verdog / the Road's / that a dog / the dog = Datadog；BIA = Bits AI；'B or open search' 中 B 应为 Elastic；'positive' 应为 position；'RTA' = RCA；中文整理版 L253 与 L257 对 Q1 数字表述不一 ('十几的中高段' vs '15% 至 20%')；L295 把 hyperscale AI model providers 误译为超大规模云服务提供商；回答多为通用框架式论述，可能是准备好的话术，具体客户案例几乎为零；与 MeritForum 2026-06-16 纪要 (M1) 同源同场；但 M1 把 Q1 写成 '二十个百分点低位'、Q2 写成 '增速放缓'，与本音频原话相反，以本音频为准
A 2026-06-16 · MeritForum · MeritForum 文字纪要：北美软件渠道商离职专家谈 Datadog AI 原生客户 +35–50% 消费增长 (与 Merit 调研 I 同源) 2 条
文件：KB/MeritForum/2026-06/2026-06-16-bei-mei-ruan-jian-qu-dao-shang-chi-zhi-zhuan-jia-datadogzai-aiyuan-sheng-ke-hu-zhong-shi-xian-35-50-xiao-fei-zeng-chang.md
发言人：标题写明 '北美软件渠道商离职专家' (L2)，作者 Lantian；内容与 Merit Datadog 产品进展调研 I (2026-06-16 音频) 逐段对应，判断为同一专家同一场的英文转写再翻译版。局限同音频 I：只看到渠道客户，无样本量，回答偏框架化。
阅读范围：全文 L1-L228 可信度：仅作局部线索
与音频 I 同源，内容基本一致 (AI 原生 +35–50%、传统 +10–20%、AI 可观测性采用 70–90% vs 20–40%、部分迁移、折扣)。但本纪要把 Q1/Q2 增速写成 'Q1 二十个百分点低位、Q2 增速放缓'，与音频原话 'Q1 mid-high teens、Q2 medium acceleration' 相反，属整理错误。为避免重复，只抽与音频不同的句子。
问题：与 Merit Datadog产品进展调研I (2026-06-16 音频，/mnt/user-data/uploads/Knowledge Base/2026-06-16/260616 - Merit Datadog产品进展调研I(2026-06-16.md) 同源同场，其余内容不重复抽取；L46 '第一季度的消费同比增长在二十个百分点低位。随后第二季度增速放缓' 与音频 L471 原话 'mid to high teens' 和 'Q2 is showing um, medium acceleration' 相反，纪要有误；L220 写 '未来七年'，音频原话为 'next several years'；L185 '五十万美元' 音频中没有；均为整理加工；标题把 '35–50%' 作为卖点，但这只是专家对 AI 原生客户的观察区间，无样本量
C 2026-06-16 · ThirdBridge · ServiceNow 前联盟渠道高管：AI Control Tower 和 Datadog 的分工 1 条
文件：KB/ThirdBridge/2026-06/2026-06-16-servicenow-its-anthropic-partnership-former-senior-executive-alliances-channels-at-servicenow-inc.md
发言人：ServiceNow 前 Senior Executive, Alliances & Channels (标题明示)，有约 25 年软件行业经验；访谈实际日期 2026-04-17 (Client-led)。视角局限：ServiceNow 一方，自然强调 ServiceNow 的价值；不熟悉 Datadog 产品细节。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
客户追问：Datadog 是专业观测厂商、OpenAI 也在用，为什么还需要 ServiceNow 做 AI 治理。专家称 Datadog 只做'部分'，不监控跨系统的工作流，也没有端到端的 Agent 风险面板。说明 Agent 治理层有别的厂商在争。
问题：ServiceNow 视角，有利益偏向
A 2026-06-17 · Merit 专家电话 (音频转写) · Merit Datadog 产品进展调研 II：SI 视角客户支出 Q1 +33%/Q2 约 +35%，AI 遥测弹性约 1.1，Bits AI 自主修复短期难落地 17 条
文件：KB/2026-06-17/260617 - Merit Datadog产品进展调研II(2026-06-1.md
发言人：未具名专家 (访谈中提问者叫过 'Dexter'，不确定是专家还是译员)。音频没有正式介绍；从原话推断其所在公司是系统集成商/渠道伙伴：'we work with enterprise customer mainly. We don't work as much with AI native companies, they don't they tend not to need uh system integrator uh support' (L314)，以及 '客户谈判时有时找我们帮忙' (L314)。MeritForum 同源纪要标题称 '北美软件渠道商离职专家' (M2 L2)，但原话用现在时 'we'，在职/离职有矛盾。视角局限：客户基本是传统大企业，几乎不接触 AI 原生公司；看到的是其客户在 Datadog 上的支出，不是 Datadog 全局；作为渠道方可能偏向推销 Datadog 生态。
阅读范围：全文 L1-L314 (L13-182 AI 摘要，L184-308 中文整理，L314 中英交传原始转写单行) 可信度：需交叉验证
SI 视角：其企业客户在 Datadog 上的支出 Q1 同比 +33%，Q2 预计约 +35% 小幅加速，主要来自老客户扩张 (净留存)；AI 拉动可观测性的弹性约 1.1 倍。护城河是生态和切换成本：客户每次都没能完全迁走，但在把高量低价值日志分流到 Elastic/Grafana；Datadog 给一些折扣但定价权很高。Bits AI 有兴趣但 '还看不到数字'，自主修复因合规短期难落地；Agent Console 更直接但现在不是大增长驱动。
问题：专家身份推断自原话 (系统集成商)，MeritForum 标题称 '渠道商离职专家'，但原话用现在时 'we'，在职/离职说法矛盾；33%/35% 是该 SI 客户在 Datadog 上的支出增速，与 Datadog 公司 Q1 +32% 营收增速非常接近，可能受公开数字锚定；Q2 是 6/17 时的预期；转写错字：'Dynamics' 应为 Dynatrace (M2 写作 Dynatrace 和 AppDynamics)；'cloud code' = Claude Code；'Dyna Trace' = Dynatrace；'tearing' = tiering；'data bricks' 一处应为 Datadog (专家自己更正)；中文整理版 L188 等与原话一致；但交传口译 (L314) 加了 '而且这个data dog的价格确实也就是逐年增长'，原话中没有，属于译员添加；Q2 第二个问题专家没听懂，回答偏题 (口译自己也说明)；与 MeritForum 2026-06-17 纪要 (M2) 同源同场
A 2026-06-17 · MeritForum · MeritForum 文字纪要：北美软件渠道商专家称客户 Datadog 支出 Q1 +33%、Q2 +35% (与 Merit 调研 II 同源) 1 条
文件：KB/MeritForum/2026-06/2026-06-17-bei-mei-ruan-jian-qu-dao-shang-chi-zhi-zhuan-jia-datadogyu-ji-2026nian-q1he-q2ke-hu-zhi-chu-zeng-chang-fen-bie-da-33-he.md
发言人：标题写明 '北美软件渠道商离职专家' (L2)，作者 Lantian；内容与 Merit Datadog 产品进展调研 II (2026-06-17 音频) 逐段对应，判断为同一专家同一场。专家自述 '我们主要与企业客户合作……后者通常不需要系统集成商的支持' (L53)，实为系统集成商视角；'离职' 与原话现在时矛盾。
阅读范围：全文 L1-L183 可信度：仅作局部线索
与音频 II 同源。只抽音频原始转写里没有的内容：专家补充认为 token 成本长期下降，'大规模 token 支出' 的故事 5–10 年后未必存在 (对 AI 用量拉动可观测性是反方)；以及 Dash 2026 提到与 Azure 的集成。
问题：与 Merit Datadog产品进展调研II (2026-06-17 音频，/mnt/user-data/uploads/Knowledge Base/2026-06-17/260617 - Merit Datadog产品进展调研II(2026-06-1.md) 同源同场，重复内容不再抽取；L47 写 '第一季度尚未完全结束'，音频原话是 'Q2 is not completely finished yet'，纪要有误；L154 'Trace' 应为 Dynatrace；L160 '替代 Databricks' 原话是专家口误后自己更正为 Datadog；'Cloud Code'/'Code' = Claude Code；L179-181 的 token 成本长期下降论述在音频原始转写 (L314) 末尾没有，可能是音频转写截断，也可能是纪要整理加入，需谨慎
B 2026-06-17 · TMTB (Substack) · TMTB 晨报 (6/17)：Citi 上调目标价到 $270，回应 tokenization 和实验室集中度担忧 3 条
文件：KB/Substack/2026-06/2026-06-17-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Citi (DASH 后与客户检查) 报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Citi 目标价上调到 $270，客户对 AI 可观测性反馈正面，Bits AI 和云成本管理兴趣强，取代 Splunk/New Relic 的机会稳健，开源竞争压力有限；承认投资者对 tokenization、出口管制、实验室客户集中度的担忧，但认为被高估。
问题：'tokenization' 担忧的具体含义未解释
B 2026-06-18 · Merit 专家电话 (音频转写) · Merit Datadog 产品进展调研 III (交传)：渠道伙伴称今年 Datadog 业务 +25–28%，Bits AI 兴趣大但采用少，可观测性新客户已饱和 12 条
文件：KB/2026-06-18/260618 - Merit Datadog产品进展调研III(交传)(2026.md
发言人：未具名专家，提问者称其 'Jakes'。音频没有身份介绍，MeritForum 中也没找到同源纪要。从原话推断是 Datadog/Dynatrace 的渠道合作伙伴或经销/集成商：说 'Q1 was around 20% for us'、'we're doing a lot of uh proof of concepts'、'go to market strategies in the channel'、'we are doing a lot of migration' (帮客户迁出 Oracle)，同时也卖 Dynatrace ('both DataDog and Dynatrace for us')。视角局限：只看到自己渠道的 Datadog 业务；回答笼统、多处数字前后不一致 (软件涨价 10–12% vs 10–15%，AI 支出口径混乱)，宏观 IT 预算数字 (今年 +35%、明年 +40–50%) 明显偏离常识，可信度偏低。
阅读范围：全文 L1-L316 (L13-174 AI 摘要，L176-310 中文整理，L316 中英交传原始转写单行) 可信度：仅作局部线索
渠道伙伴称自己的 Datadog 业务今年 +25–28%，Q1 约 20%、Q2 加速，来自新老客户、平台整合和多云。Bits AI 和 Agent Console 兴趣大、POC 多但 '实际采用不多'，产品对企业还不成熟、渠道 GTM 没准备好，认为明年才是 AI 起飞年。他认为 80–85% 企业已有可观测性，新客户已饱和，增长要靠 agent 流量；agent 定价可能是订阅+高级附加而非按 agent 计。
问题：无身份介绍，身份全靠推断；MeritForum 本批中没有同源文字纪要；'25–28%' 与 'Q1 约 20%' 是该渠道自己的 Datadog 业务 ('for us')，不是 Datadog 全公司；且年增 25–28% 与 Q1 20% 的关系没有解释；AI 摘要表格把年份写成 '当前状态 (2024年)'、'预计2025年'，与 2026-06 的访谈时间不符，属 AI 摘要错误；软件涨价：原话 L316 '10 to 12%'，整理版 L262 写 '10% 到 15%'，篇内不一致；AI 支出 '增长 15–20%' 与 '占软件预算 8–10%→15–20%' 口径混用；IT 预算今年 +35%、明年 +40–50% 的说法与常识偏离很大，宏观数字不宜引用；转写错字：'hypervisor solutions' 应为 hyperscaler；'matrix' = metrics；'Anthropics'；口译把 'net-net new' 理解为饱和；Anthropic 能力判断 ('不能采取行动') 时点为 2026-06，可能很快过时
A 2026-06-18 · MeritForum · SHI 离职专家：客户在 Datadog 上的支出今年增 25–28%，Q1 约 20%、Q2 加速，增长集中在中端市场 14 条
文件：KB/MeritForum/2026-06/2026-06-18-shichi-zhi-zhuan-jia-2026nian-aixiang-guan-zhi-chu-yu-ji-da-fu-zeng-chang-ruan-jian-xiao-fei-ming-nian-jiang-ti-sheng-25.md
发言人：标题写的是 'SHI 离职专家' (SHI International，大型 IT 产品转售/渠道商)，但正文全程用 '我们' 的现在时 ('我们尚未部署该功能'、'我们正在与许多客户合作')，任职是否仍在、离职多久都不清楚。具体职位没有披露。局限：转售商视角，主要看到的是中端市场；增速口径是 '客户在 Datadog 上的平均支出增长率'，不是 DDOG 收入；回答多数是定性的，部分数字 (IT 预算 +40–50% 等) 偏夸张。
阅读范围：全文 L1-L193 可信度：仅作局部线索
转售商口径：2026 年客户在 Datadog 上的支出增 25–28%，Q1 约 20%、Q2 加速，新老客户都有贡献，增长集中在中端市场和 DevOps。驱动是 AI agent 带来的复杂性、日志和指标增加、AI/安全日志，以及多云和工具整合。Bits AI 兴趣大但采用率不高，成本是障碍；他认为 AI 实验室自建观测还很遥远；agent 至少带来 20% 增长 (他的估计)。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；标题说 '离职'，正文却用 '我们' 的现在时，身份时效不清；增速口径是 '客户在 Datadog 上的平均支出增长率'，含义不清 (是渠道销售额还是客户账单)；'Q1 约 20%' 与 '今年 25–28%' 在算术上需要 Q2 远高于 30%，可能口径不一；'Agent Council' 不是已知的 Datadog 产品名，疑为转录错误；IT 预算 '明年增长多达 40–50%'、'同比 +35%' 等数字明显偏高，专家宏观数字可信度低
B 2026-06-19 · ThirdBridge · Elastic 前战略伙伴总监：Elastic 很少在可观测性上赢 Datadog，主要从 Splunk 抢客户 6 条
文件：KB/ThirdBridge/2026-06/2026-06-19-elastic-stretching-the-competitive-advantage-from-core-search-to-ai-retrieval.md
发言人：Christine Malloy：Third Bridge 介绍为 Elastic 前 Director of National Strategic Partners (L34)；页头写的是 Perforce Software 现任 Partner Ecosystem Manager；自述曾在 New Relic 工作数年。访谈日期 2026-06-10 (Analyst-led，主持人 Luke Fragnito)。视角局限：伙伴/联盟视角；离开 Elastic 的时间未说明；曾在 New Relic 工作可能影响她对 Datadog 的看法。
阅读范围：全文浏览 L1-268，重点读可观测性/竞争段落 (L19-30, L44-54, L76-100, L118-132, L230-234) 可信度：需交叉验证
专家称在全栈可观测性上，Elastic 很少从 Datadog/Dynatrace 手里抢到客户，替换主要发生在 Splunk (成本驱动)；New Relic、Datadog 不把 Elastic 当竞争对手，Datadog 胜在易用和开发者认同。Elastic 赢的时候靠价格性能比、数据留存和自定义搜索。这说明 Elastic 不是 Datadog 的主要通缩来源。
问题：专家身份前后不一致 (页头写 Perforce 现任，介绍里写 Elastic 前总监)；曾在 New Relic 工作可能偏向 Datadog；是 Elastic 视角，并非 Datadog 客户
A 2026-06-22 · Merit 专家电话 (音频转写) · Merit Datadog 产品进展调研 IV (交传)：前员工称 Q1/Q2 收入 +30–35%，OpenAI 占比低到高个位数、Anthropic 低到中个位数，Bits AI 环比 +50–60% 22 条
文件：KB/2026-06-22/260622 - Merit Datadog产品进展调研IV(交传)(2026-.md
发言人：未具名专家，提问者称其 'Jason'。音频无正式介绍；同源 MeritForum 纪要标题为 'Datadog离职专家' (M3 L2)，且 M3 L169 专家自述 '我来自印度'。原话以 Datadog 内部人口吻说话：'for a company of our scale' 'a lot of our uh uh uh new business calls' 'When we started doing competitive analysis' (M3)，并对 OpenAI/Anthropic 合同、续约、收入占比有具体说法，像是做销售/GTM 或客户相关岗位的前员工；具体职位、离职时间未披露。视角局限：离职后的信息可能过时；收入占比等是口头区间，可能是推测或记忆；涉及大客户合同的说法有 MNPI 风险，需谨慎使用；作为前员工可能偏向看多 Datadog (几乎所有风险都说 '不担心')。
阅读范围：全文 L1-L323 (L13-179 AI 摘要，L181-317 中文整理，L323 中英交传原始转写单行；原始转写在最后一个问题 '是否面临新进入者威胁' 处截断，答案见 M3) 可信度：需交叉验证
前员工称 Q1、Q2 收入同比都在 30–35%，Q2 很轻微加速，驱动是 AI 用例、Bits AI (GA 6–7 个月) 和安全大单。AI 原生客户占总收入 high single digits，但对增长贡献显著且上升；OpenAI 占比 '低到高个位数'，即使完全流失也能被其他 AI 原生客户抵消；其他模型公司过去 5–6 个月签了多年承诺合同，Anthropic 去年成为客户、已续约，占比低到中个位数，是 Datadog 史上最大单之一。Bits AI 占收入低到中个位数，环比 +50–60%。客户以换供应商/自建为谈判筹码，但说要关掉实例就撤回。
问题：收入占比全是口头区间，篇内自相矛盾：AI 原生 cohort 总共 'high single digits'，但 OpenAI 'low to high single digits' + Anthropic 'low to medium single digits' 加起来可能超过整个 cohort；口译/中文整理把 OpenAI 'low to high single digits' 译成 '5%–9%'、'不到10%'，把 Bits AI 'low to medium single digits' 译成 '5%左右'，均为译员加工，以英文原话为准；Bits AI 占收入 '低到中个位数' (按约 10 亿美元/季，即 1000–5000 万美元/季) 与专家C (L316) '采用很少'、专家B '还看不到数字' 冲突，且公司未披露，疑似高估或口径为 '涉及 Bits AI 的合同'；AI 原生占比 high single digits 需与公司披露的 AI-native cohort 占比核对 (公司口径可能更高)；涉及具体客户合同、续约、占比，属潜在 MNPI；信息来自离职员工，可能过时；'Fable' 模型 (L323) 为专家所称 Anthropic 最新模型名，未核实；转写错字：Dino trace/Dyna Trace = Dynatrace；Central One = SentinelOne；'card functionality' 应为 cart；'data has uh signed' = Datadog；原始转写在最后一个问题处截断，Grafana/Honeycomb 等答案只在 M3 纪要；与 MeritForum 2026-06-22 纪要 (M3) 同源同场
A 2026-06-22 · MeritForum · MeritForum 文字纪要：Datadog 离职专家谈 Q1/Q2 收入 +30–35%、Bits AI 环比约 60% (与 Merit 调研 IV 同源) 3 条
文件：KB/MeritForum/2026-06/2026-06-22-datadogchi-zhi-zhuan-jia-datadogzai-2026nian-q1he-q2shou-ru-tong-bi-zeng-chang-30-35-aiyu-an-quan-chan-pin-tui-dong-zeng.md
发言人：标题写明 'Datadog离职专家' (L2)，作者 Lantian；L169 自述 '我来自印度'；内容与 Merit Datadog 产品进展调研 IV (2026-06-22 音频) 逐段对应，判断为同一专家同一场。局限同音频 IV：前员工，口头区间，可能 MNPI，偏看多。
阅读范围：全文 L1-L177 可信度：仅作局部线索
与音频 IV 同源。额外内容在 L165-175 (音频转写被截断部分)：新进入者里 Grafana (印度初创常用 Grafana+Prometheus+OTel，并转向付费版) 和 Honeycomb (APM) 较突出，但收入都不到 Datadog 的 10%，没看到大量份额被抢；银行是 Datadog 要争取的大市场。另外纪要把 Anthropic 写成 '属于较小的客户之一'，与音频 '史上最大单之一' 不一致。
问题：与 Merit Datadog产品进展调研IV(交传) (2026-06-22 音频，/mnt/user-data/uploads/Knowledge Base/2026-06-22/260622 - Merit Datadog产品进展调研IV(交传)(2026-.md) 同源同场，重复内容不再抽取；L165-175 的 Grafana/Honeycomb 回答只在本纪要有 (音频转写截断)；L145 写 Anthropic '比较小——属于较小的客户之一'，音频原话是 'relatively smaller actually, but a big big uh one of the biggest deals that data has uh signed historically'，纪要漏译了 '史上最大单之一'；L99 写 '包括更多传统公司' 的加速增长抵消 OpenAI，音频原话是 AI 原生 cohort 增长更快，纪要有误；转写/翻译错误：L88 'Todal' 应为 Toto；L104 把 MCP 注释为 '管理控制平面' (应为 Model Context Protocol)；L151 'Accenture' 应为 SentinelOne，'数百亿美元' 原话为 tens of millions；L153 'Agent Council' 应为 Agent Console；L125 '还未提交 IPO' 与原话 'filed for IPOs this year' 相反；标题 'Bits AI 季度增长约60%' 取自原话 '50, 60%' 区间上限
B 2026-07-01 · TMTB (Substack) · TMTB 晨报 (7/1)：BofA 评 DDOG 收购 Adaptive ML；BTIG 称 Chronosphere 年化收入 $300M+ 3 条
文件：KB/Substack/2026-07/2026-07-01-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 BofA (DDOG) 与 BTIG/BNP (PANW，涉及 Chronosphere) 报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
BofA 认为收购 Adaptive ML 增加强化学习能力，利于 Bits SRE/Code/Security Analyst；BTIG 称 PANW 旗下 Chronosphere 年化收入超 $300M，有望超 $1B ARR，是 DDOG 竞品规模的参照。
C 2026-07-06 · Acecamp · Cloudflare 内部专家：agentic 开发让 MR 数翻倍；Cloudflare 自带的观测只是附赠，业务日志仍推到 Splunk/Datadog 2 条
文件：KB/Acecamp/2026-07/2026-07-06-ai-infrada-han-agentic-loopluo-di-chang-jing-harnessjia-gou-he-shang-ye-hua-yu-qi-aiguan-ce-neng-li-he-yu-san-fang-he-zu.md
发言人：未具名；用 Cloudflare 内部视角回答 ('Cloudflare内部'、'我们能看到merge request的数量')，推测为 Cloudflare 现任员工，职位不详。局限：对 Datadog 只是顺带一提。
阅读范围：全文 L1-L126 可信度：仅作局部线索
Cloudflare 内部用 agent 写代码、做 review 后，MR 数几乎翻倍，单个安全 review 场景每天要处理几十亿 token。Cloudflare 自带的观测只是附赠功能，主要用于合规；业务日志的全量存储和分析一般仍用 Splunk 或 Datadog，Cloudflare 通过 log push 把日志推过去。
问题：整理稿非逐字：MeritForum/Acecamp 为机构整理/翻译稿，quote 是整理后的文字而非专家逐字口述；身份为推测
B 2026-07-06 · TMTB (Substack) · TMTB 晨报 (7/6)：Bernstein 上调长期预测但下调评级至 Market Perform 3 条
文件：KB/Substack/2026-07/2026-07-06-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Bernstein 降级报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：需交叉验证
Bernstein 上调长期预测但因估值和近端预期过高降级到 Market Perform：预计 2Q 强，但 3Q/4Q 起 ex-AI 需求 (约占收入 85%) 走平、AI 实验室增长正常化，4Q 增速放缓约 500bp 到约 29%，低于买方 30%+ 和峰值 high-30s 到 40%+ 的预期。
B 2026-07-09 · FactSet 会议转录 · RAISE Summit 巴黎：Pomel 谈用量制不受 agent 冲击、Adaptive ML、传统企业连续 4 季加速 8 条
文件：KB/FactSet/2026-07/2026-07-09-corrected-transcript-datadog-inc-ddog-us-raise-summit-9-july-2026-3-40-am-et.md
发言人：出席人：Olivier Pomel，Datadog CEO 兼联合创始人 (明示)。主持人未署名 (FactSet 标 Unverified Participant)，为 RAISE 创业大会主持人，非卖方分析师，问题偏泛。视角局限：面向创业者听众，后半段谈个人经历；数字少且多为大约数。
阅读范围：全文 L1-L172 可信度：需交叉验证
Pomel 称收入在 $4-5B 之间、增长约 30%，传统企业连续四个季度加速，云原生的加速来自 coding agent 和自建 agent。被问“AI 带来更多数据，会不会提价/压力”时，他用“我们是用量制”回避了价格问题。Adaptive ML 收购后不再卖其产品，转为内部训练专用模型 (称快 100 倍、便宜 10 倍)。他还说客户 AI 成本暴涨、优化空间很大，这个说法两面都可读。
问题：主持人未署名，问题较泛；'40% of all cases can now be completely fixed autonomously' 是主持人引述，未说明来源，管理层未确认；未谈最大客户、SBC、回购
A 2026-07-09 · GuidePoint · F500 渠道伙伴 (Scicom) 7 月渠道检查：其 Datadog 业务 Q1 +31%、Q2 +33%，下半年目标 37–38% 14 条
文件：KB/GuidePoint/2026-07/2026-07-09-observability-datadog-and-dynatrace-july-2026-check-with-f500-partner-on-ai-impact.md
发言人：Sid K Roy (SKR)，Scicom Infrastructure Services, Inc. 执行副总裁 (运营与客户支持) 兼 CTO (页眉明示，Viewpoint: Industry Consultant)；自述 23 年服务 F500 企业客户，覆盖观测、ITSM、数据安全、AI (L179-184)。从 '我们的 Dynatrace/Datadog/Splunk 增长目标' 的说法看，Scicom 是这些厂商的经销/服务伙伴 (推测，依据：用 'our pipeline, renewals, competitor replacements' 描述销售)；本篇是系列回访 ('since our last call')。视角局限：增速数字是其公司自己的 Datadog 业务 (转售/服务) 的增速，不是 Datadog 公司收入；只覆盖其 F500 客户群；作为卖方渠道有看多偏向；其业务 60–70% 集中在 Q3/Q4，下半年数字是目标而非结果。
阅读范围：全文 L1-L623 可信度：需交叉验证
渠道伙伴口径下，传统企业 (F500) 的 Datadog 业务在加速：Q1 +31%、Q2 +33%，全年目标 35%、下半年瞄准 37–38% 甚至接近 40%；驱动是应用现代化/上云、日志 (含 SIEM 安全日志)、On-Call 替换 PagerDuty/xMatters，以及客户开始单独购买 LLM Observability 容量。支持 E01 (非 AI 企业加速) 和 E03 (AI 增强应用生成更多 trace/日志、MCP 服务器监控)。反方是他提醒 Databricks/Snowflake/ClickHouse 等数据平台会因存储和计算成本快速上升而威胁 Datadog。
问题：所有增速是渠道伙伴自己的 Datadog 业务增速 (转售/服务)，不等于 Datadog 公司收入，口径需注意；H2 37–38% 是销售目标，且其业务 60–70% 集中在下半年；L451-452 'they acquired BindPlane with Datadog' 转录混乱 (BindPlane 是 Dynatrace 收购)；GuidePoint 摘要表把 Dynatrace 'Annual Growth Target 40-45%' 和 Datadog '35%' 写成公司层面，容易误读
B 2026-07-16 · TMTB (Substack) · TMTB 晨报 (7/16)：BNP 经销商调查前瞻支出分数骤降；UBS 客户检查支持 30%+ 但定价阻力是主要摩擦 4 条
文件：KB/Substack/2026-07/2026-07-16-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 BNP Paribas 季度经销商调查与 UBS 客户检查 (7 位客户 + 3 位伙伴) 摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
BNP 经销商调查：DDOG 2Q 销售高于计划几个点，但 DDOG 前瞻支出分数从 5.4 降到 2.0，需求可能集中在 2Q。UBS 7 位客户+3 位伙伴：观测性仍是优先支出、DDOG 靠供应商整合拿份额，但定价阻力是主要摩擦；维持 2026 年 30%+ 增速判断，估值约 20x 收入、75x FCF。
C 2026-07-24 · GuidePoint · Cognition 系列访谈：前 Datadog 人才负责人的顺带回忆 0 条
文件：KB/GuidePoint/2026-07/2026-07-24-vibe-coding-series-cognition-product-strategy-gtm-and-competitive-positioning-vs-cursor-claude-code-github-copilot-and-a.md
发言人：Kira Busman，Cognition 前顾问，前 Index Ventures 全球人才网络负责人；更早在 Datadog 负责全球人才招聘 (员工从约 300 人到 5,500 多人，经历 IPO) (L183 自述)。视角局限：招聘/人才视角，谈 Datadog 只是回忆，与当前业务无关。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
访谈主题是 Cognition；Datadog 只出现在专家履历和一个早年案例里 (拿下 Mercedes-Benz 后 Audi 跟进)。对 thesis 没有实质信息。
问题：与 DDOG thesis 基本无关
B 2026-07-28 · TMTB (Substack) · TMTB 晨报 (7/28)：BTIG 目标价 $289、Morgan Stanley 目标价 $300，渠道检查支持 2Q 中 30s 增长 5 条
文件：KB/Substack/2026-07/2026-07-28-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 BTIG (伙伴检查) 与 Morgan Stanley (企业检查) 报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
BTIG 最大联系人 2Q 增速约 35% (1Q 32%、2025 约 20%)，AI 推理是最明确驱动，SolarWinds 整合迁移带来额外上行，目标价 $289。MS 目标价 $225→$300，预期 2Q 中 30s 增长 (街 30.5%)、3Q 指引约 32%、FY26 至少 30%；一家大 LLM 客户、两家研究实验室推动消费；承认约 16x CY27 销售不便宜。
问题：MS 预期 3Q 指引约 32%，实际约 29%，说明卖方/买方预期明显高于公司指引
C 2026-07-30 · ThirdBridge · 前 Monte Carlo 战略客户 AE 谈数据/agent 观测：客户把数据观测指标汇到 Datadog，APM 巨头打包是真风险 2 条
文件：KB/ThirdBridge/2026-07/2026-07-30-monte-carlo-betting-the-house-on-agent-observability-dynamics-of-commercial-technical-scalability.md
发言人：Terry Lucy (TL)。Third Bridge 页眉 'Executive, Go-to-market at CodeRabbit Inc'，介绍为 'former Strategic Account Executive at Monte Carlo Data' (明示)；自述在 Monte Carlo 多年，是伦敦办公室区域第二号员工，负责企业/战略客户 (L42)。视角局限：数据观测 (非 APM) 销售视角，对 Datadog 的看法来自竞争一侧；主题是 Monte Carlo，涉及 Datadog 的内容只有几段。
阅读范围：仅读 Datadog / agent observability 竞争相关段落：L1-L60 (页眉、要点、介绍、公司定位)、L124-L132 (agent observability 押注与增长)、L176-L206 (竞争格局、Metaplane/Datadog、APM 巨头打包、并购) 可信度：仅作局部线索
从数据观测一侧看，Datadog 通过收购 Metaplane 把数据观测打包进平台；客户常把 Monte Carlo 的指标抽出来放到 Datadog 统一看，若 Datadog 能力相当，很多公司会标准化到 Datadog。这是 Datadog 平台交叉销售 (E01/E05) 的旁证，但对 agent 观测的变现没有直接信息。
问题：专家来自竞争对手一侧，对 Datadog 的评价没有数字
B 2026-07-30 · TMTB (Substack) · TMTB 晨报 (7/30)：BofA 依据 AWS 收入分析认为 DDOG 预期保守 3 条
文件：KB/Substack/2026-07/2026-07-30-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 BofA 报告摘要 (AWS 收入分析)。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
BofA 把 DDOG 列为 2Q26 首选，AWS 收入分析显示预期保守、可能从 1Q 的 32% 再加速；两家 hyperscaler AI 研究大单和 5 家 $10M+ AI 客户分散客户基础，降低续约波动担忧；OpenAI 事件强化 agent 可观测性需求。
问题：AWS 收入分析的具体数据和方法未给出；'OpenAI incident' 未说明是什么事件
C 2026-07-31 · Merit 专家电话 (音频转写) · 模型路由调研：AI 网关与 Datadog 集成或带动观测采购 1 条
文件：KB/2026-07-31/260731 - Merit 模型路由进展调研I(2026-07-31 13_0.md
发言人：未具名专家，LiteLLM 前任人员 (由同场 MeritForum 07-31 标题'LiteLLM 离职专家'推得)。视角局限：AI 网关/路由厂商视角，对可观测性采购为推测性判断，没有具体账单数据。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
专家认为 AI 网关已成标准实践 (治理、可观测性、成本跟踪)，许多网关可向 Datadog 导出遥测/日志/链路，企业用网关后会增加此类采购；未给数字。
问题：没有量化；AI 网关本身也带可观测性功能，也可能替代一部分 LLM 观测需求，专家没有讨论
C 2026-07-31 · MeritForum · LiteLLM 离职专家：AI 网关标准化 (与 Merit 07-31 同一通话的整理版) 1 条
文件：KB/MeritForum/2026-07/2026-07-31-litellm-chi-zhi-zhuan-jia-qi-ye-zhong-ai-de-cheng-shou-kai-fa-zhe-cai-na-lu-84-api-kai-zhi-88-ji-zhong-yu-mei-guo-zhuan.md
发言人：LiteLLM 离职专家 (标题明示)，具体职位未说明。同 Merit 07-31 那篇。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
与 Merit 07-31 内容重复：AI 网关与 Datadog 的遥测/日志/追踪导出集成，会带动观测采购。
问题：与 Merit 07-31 重复，不要重复计数
A 2026-08-03 · Funda (Substack) · Funda 2Q26 财报前渠道调研：渠道动能健康、Bits AI 仍早期、OpenAI 优化被 Anthropic 吸收 27 条
文件：KB/Substack/2026-08/2026-08-03-preview-ddog-2q26-healthy-channel-momentum-bits-ai-still-early-openai-optimization-absorbed-by-anthropic.md
发言人：FundaAI 季度渠道调研摘要，6 位匿名专家：北美渠道专家 A (同 1Q)、全球渠道/GSI 专家 B (传统企业客户为主)、北美 F500 渠道专家 C、独立可观测性顾问 D、北美企业客户 E (信息服务平台的平台工程自动化负责人，年 DDOG 支出约 $5m，企业许可约 $23/主机/月)、北美企业客户 F (SaaS 基础设施公司 DevOps 工程师，70% Grafana / 30% DDOG)。身份为 Funda 明示。二手转述，看不到底稿。局限：渠道伙伴有卖 DDOG 的利益偏向，且只看到自己客户；客户 E/F 是单一样本；OpenAI -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，Anthropic 部分 Funda 自承 'limited visibility'。
阅读范围：全文 L1-L335 可信度：需交叉验证
渠道普遍偏多：专家 A 的 book 2Q 跟踪 33–35%，2H 目标从 26–28.5% 上调到 32–37%；增量主要来自存量客户多产品扩张，AI 让应用数从 400 增到约 650。但也有清晰反证：客户 E 的 LLM 可观测性已输给便宜约 20% 的 BrainTrust，DDOG 定价不灵活使其评估 Grafana/Prometheus；顾问 D 称年花 $20m 的客户环境扩 3 倍也不会花 $60m，未来约 80% 数据会被采样。Funda 估 OpenAI 2Q 用量环比降 10–20%，由 Anthropic 抵消，2Q AI 原生收入约 $170m (约 15%)；预期 2Q +35%、3Q 指引约 30%，但提示股价自上次财报涨 83%、仓位拥挤。
问题：专家 A 1Q 增速从上季的 '~21%' 改为 '~26%'，同一信源前后不一致；OpenAI 2Q 用量 -10–20% q/q 与 AI 原生收入约 $170m 是 Funda 推算，方法未披露；Anthropic 部分 Funda 明说 'limited visibility on the specific numbers'；Bits AI 定价：1Q preview 为 $500/月含 20 次调查 (约 $25/次)，本篇为每次 7–8.5 credits (约 $7–8.5)，若口径一致则单次价格降约 65–70%，但两者计费单位 (完成调查 vs SRE 调查) 可能不同，需核实；渠道伙伴有卖 DDOG 的利益，乐观偏向；专家 A 自承 AI 占 10–15% 'on the optimistic side'；客户 E 的 '30% 增速' 与 '无 AI→可观测性换算公式' 并存，AI 部分 15pp 是主观拆分
B 2026-08-03 · TMTB (Substack) · TMTB 晨报 (8/3)：转述 FundaAI 2Q 预览 + Wells Fargo 客户调查 2 条
文件：KB/Substack/2026-08/2026-08-03-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 FundaAI 预览 (同 8/3 Funda 原文) 和 Wells Fargo 2Q 客户调查 (约 $95M 年 DDOG 支出) 摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
TMTB 摘要 FundaAI 的 2Q 预览 (35% 增长、AI 原生约 $170m、3Q 指引约 30%)。Wells Fargo 2Q 调查覆盖约 $95M 年 DDOG 支出，多数大客户预计 2026 增加支出并继续把份额转向 DDOG，但比 1Q 略软 (受访者偏小客户)。
问题：FundaAI 段落与 Funda 8/3 原文重复，不应重复计数
A 2026-08-06 · FactSet 业绩会转录 · Datadog Q2 2026 业绩会：收入 +36%、非 AI 到 high-20s，最大 AI 客户九位数续约但 Q3 起减量，FY26 指引 +30% 41 条
文件：KB/FactSet/2026-08/2026-08-06-corrected-transcript-datadog-inc-ddog-us-q2-2026-earnings-call-6-august-2026-8-00-am-et.md
发言人：Olivier Pomel (联合创始人兼 CEO，现任)、David Obstler (CFO，现任)、Yuka Broderick (IR 高级副总裁)；身份由开场白明示。卖方：Sanjit Singh (MS)、Raimo Lenschow (Barclays)、Gabriela Borges (GS)、Mike Cikos (Needham)、Alex Zukin (Wolfe)、Ritu Kashyap (KeyBanc，代 Eric Heath)、Koji Ikeda (BofA)、Samik Chatterjee (JPM)、Howard Ma (Guggenheim)、Andrew Sherman (TD Cowen)、Brad Reback (Stifel)、Ittai Kidron (Oppenheimer)、Andrew DeGasperi (BNP)。视角局限：公司口径偏正面；最大客户细节拒绝披露；转录无发言人标注
阅读范围：全文 L1-L472 可信度：可作主要证据
Q2 收入 $1.12B (+36%)、环比 +$115M 创纪录，非 AI 客户增速 18%→high-20s，新客户贡献 30%，剔除最大客户的其余业务连续 5 季加速，支持 '广泛加速'；AI 客户 750 家、>$1M 31 家、>$10M 8 家 (口径扩大)。反方是最大 AI 客户九位数续约但 Q3 起减量，管理层拒绝说明是单价还是用量；Q3 指引环比仅约 +$15–25M；Infinite Cardinality '不额外收费'、BYOC/日志联邦 (含 ClickHouse)、客户转向 AI 成本优化、毛利率 79.6% 下行，都指向单价压力；NRR 停在 low-120s。
问题：转录无发言人标注，speaker 按上下文推定；L250 附近有时间戳残留 '(00:31:39)'，L210 'known AI' 应为 'non-AI' 的转录错误；AI 客户组口径在 Q2 扩大 (新增 hyperscaler 内部 AI 实验室)，>$1M 从 22→31、>$10M 从 5→8 的增长部分可能来自口径变化，不能全当有机增长；最大客户减量幅度、续约期限、单价/承诺水平均未披露；'nine figure renewal' 是合同总额还是年化不明；Q3 指引环比仅 +$15–25M (我按 Q2 实际 $1.12B 与指引推算)，FY 指引中值 $4.46B 减 Q1-Q3 约 $3.27B 推算 Q4 约 $1.19B——这些是推算，非公司披露；'剔除最大客户增速 pretty much the same' 与 Q2 收入 +36% 的组合说明最大客户同比贡献已接近整体增速，但公司未给最大客户收入占比；分析师没有直接问 Chronosphere、OTel、ClickHouse、自建；ClickHouse 只出现在 '日志联邦查询' 产品描述里 (L74)，管理层未讨论竞争；SBC、回购、GAAP 指标未在电话会讨论；仅能从稀释股数指引 (369M→378M) 侧面看到稀释；MCP 只给倍数 (环比 4x、较 Q4'25 22x)，Bits AI 采用和收入未量化，AI credits 定价细节未披露
A 2026-08-06 · Funda (Substack) · Funda 2Q26 财报点评：基本面没变，但预期过高 11 条
文件：KB/Substack/2026-08/2026-08-06-review-ddog-26q2-unchanged-fundamentals-met-with-too-high-expectations.md
发言人：FundaAI 自有财报点评，混合公司披露 (收入、AI 原生客户数、OpenAI 续约) 与 Funda 推算 (OpenAI 2Q 支出 -20% QoQ、Chronosphere 便宜 30–40%、非 OpenAI AI 原生 3Q 超过 OpenAI)。二手转述，看不到底稿。
阅读范围：全文 L1-L42 可信度：可作主要证据
2Q 收入 +35.6% 符合买方 35–36%，但 beat $46m 小于上季 $50m；3Q 指引上限 29.5%，市场想要 30% 以上。Funda 认为不及预期主要来自 OpenAI 优化 (估 2Q 支出环比降 20%)，管理层确认与 OpenAI 续签了 9 位数合同但 3Q 用量继续环比下降；Funda 认为 Chronosphere 比 DDOG 便宜 30–40%，在 OpenAI 内部抢份额。非 AI 增速升到 high-20s，AI 原生 >$1M 客户 31 家、>$10M 8 家；毛利率跌破 80%。
问题：'3Q 非 OpenAI AI 原生超过 OpenAI' 是在假设 5pp beat、非 AI 加速、OpenAI -20% 下的推算，非事实；OpenAI -20% QoQ 估算方法未披露；9 位数续约合同的期限、最低承诺金额未披露
B 2026-08-06 · TMTB (Substack) · TMTB 晨报 (8/6)：DDOG 2Q 后 -16~-18%，3Q 指引低于买方约 31% 的 bogey 5 条
文件：KB/Substack/2026-08/2026-08-06-tmtb-morning-wrap-sandisk-sndk-dash-hubs-app-ddog-u-recaps-the-usual.md
发言人：TMT Breakout (TMTB) 日报，汇总 DDOG 2Q26 公布数字、街上与买方预期，并评论股价反应。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
2Q 收入 $1.121B (+36%) 符合买方 35–36%，OPM 22.9%，FCF $279M；FY26 指引 $4.45–4.47B (+30%)。问题在 3Q 指引 $1.135–1.145B (+29%)，高于街 $1.106B 但低于买方约 31%，并暗示 4Q 更急的减速；仓位拥挤，股价跌 16–18%，不少人质疑跌 20% 是否过度。
B 2026-08-09 · Funda (Substack) · Funda 周报 (8/9)：DDOG 没做错什么，但预期满、仓位挤所以跌；新增两条 DDOG 客户访谈摘要 4 条
文件：KB/Substack/2026-08/2026-08-09-weekly-rsi-is-the-real-ai-variable-fcc-optics-curbs-app-u-divergence-pltr-crushes-net-beat-and-raise-vs-ddog.md
发言人：FundaAI 周报：DDOG 段落为 2Q review 摘要 + Funda 当周两次专家访谈的一句话摘要 (某全球云通信平台的工程高管——疑似 Twilio，因标题标 TWLO；某全球金融信息服务商经理)。二手转述，看不到底稿，访谈全文在 Funda 平台未附。
阅读范围：全文 L1-L335 (DDOG 相关 L23-25、L170-210、L281-293) 可信度：仅作局部线索
Funda 把 DDOG 和 NET 对照：DDOG 增长 35.6%、非 AI 加速到 high-20s、续签 9 位数 OpenAI 合同，'什么都没做错'，但因预期太满和仓位拥挤而下跌；基本面和事件交易是两回事。新增访谈摘要提到 LLM 可观测性流向 Braintrust、价格压力推动客户评估替代品、Grafana 的使用惯性妨碍 DDOG 整合。
问题：两条访谈只有一句话摘要，可能与 8/3 preview 的客户 E/F 是同一批人 (描述高度相似：金融信息服务 + BrainTrust；Grafana 并行)，不应重复计数
C 2026-08-10 · TMTB (Substack) · TMTB 收盘 (8/10)：DDOG 财报后反弹 +11% 1 条
文件：KB/Substack/2026-08/2026-08-10-tmtb-eod-wrap.md
发言人：TMT Breakout (TMTB) 日报，市场评论。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
财报后一周 DDOG 单日 +11%，TMTB 归因为软件投资者开始逢低买入有加速的名字。
B 2026-08-10 · TMTB (Substack) · TMTB 晨报 (8/10)：Barclays 称非 AI 加速广泛、3Q 指引按最大客户合同最低额 3 条
文件：KB/Substack/2026-08/2026-08-10-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Barclays 与 DDOG CEO/CFO 财报后 fireside 会面的报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：需交叉验证
Barclays：CEO/CFO 会面确认加速不只来自 AI 原生，非 AI 增速从一年前约 18% 升到 high-20s，因 AI 推动上云、基础设施现代化和新应用开发；3Q 指引刻意假设最大 AI 客户只贡献合同最低额。GPU 监控、agent 可观测性、Bits AI、MCP、AI 安全仍在变现早期。
A 2026-08-12 · FactSet 会议转录 · Canaccord 成长大会：CFO 讲合同承诺 vs 用量、指引打折方法、Bits 按 token 定价 18 条
文件：KB/FactSet/2026-08/2026-08-13-corrected-transcript-datadog-inc-ddog-us-canaccord-genuity-growth-conference-12-august-2026-3-30-pm-et.md
发言人：出席人：David M. Obstler，Datadog CFO (明示)。主持人 Kingsley Crane，Canaccord Genuity 科技分析师。视角局限：CFO 口径；部分数字为转述 (如 'over 30 of them having a million' 口径不清)；Q2 刚发布后一周。
阅读范围：全文 L1-L223 可信度：可作主要证据
这场 CFO 讲得最直白：合同是 take-or-pay，期限 1-3 年，有最低承诺；量大和期限长都有折扣 (volume-based pricing and term pricing)，所以客户有动力扩大承诺。指引方法是把长期观察到的增长率/NRR 打折。非 AI 业务连续 4-5 季加速到 high-20s，环比增 $115M。Bits 定价从按调查扩到按 token，CFO 回避了毛利问题。CFO 称收费按工作负载，不在乎是人还是 agent 在看。
问题：'over 30 of them having a million' 口径不明 (>$1M 还是 >$10M)；'seed model' 应为 seat model；Bits 毛利问题被回避；未谈 SBC、回购
B 2026-08-12 · Merit 专家电话 (音频转写) · Snowflake Q2 调研：Snowflake 可观测性难与 Datadog/Dynatrace 竞争 5 条
文件：KB/2026-08-12/260812 - Merit Snowflake Q2业绩调研 III(2026.md
发言人：未具名专家，从'我们这边的客户''我们围绕 Snowflake 的可观测性平台开展了大量市场推广培训'推测为 Snowflake 内部 GTM 人员或紧密合作伙伴 (未明示)。视角局限：站在 Snowflake 一方；只看自己客户，未必了解 Datadog 内部数据。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：需交叉验证
专家认为 Snowflake 今年虽为可观测性平台做了 GTM 准备，但很难与 Datadog/Dynatrace 竞争 (两者增速'肯定超过 20%'，产品颗粒度更细)，目前未见客户从 Datadog/Dynatrace 迁往 Snowflake；机会要等 AI 工作负载规模化后。
问题：'Datadog 主要服务中端和小企业'与 Datadog 实际大客户结构不符，专家对 Datadog 认知可能有偏差；'增速超过 20%'为外部印象，非数据；英文引文在 L665 长行内，'Dynat' 与 'race' 被转录切开
C 2026-08-13 · Merit 专家电话 (音频转写) · AI infra Q2 调研：研究员提问称 OpenAI 因成本高自建替代 Datadog 1 条
文件：KB/2026-08-13/260813 - Merit AI infra Q2进展调研(2026-08-1.md
发言人：未具名专家，从'我们已经进行了大量优化''我们还推出了自己的 AI Gateway'推测为 Databricks 现任员工 (产品/解决方案方向，未明示)。视角局限：谈 Databricks；Datadog 只出现在研究员提问里，专家没有回应 Datadog 部分。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
Datadog 仅出现在研究员提问的前提里：'Datadog 成本太高，OpenAI 一直在自建并替代'；专家只回答 Databricks 被 AI 实验室替代的风险，没有证实或评论 Datadog。
问题：Datadog 论断来自提问者而非专家，不可当作专家证据
B 2026-08-17 · TMTB (Substack) · TMTB 晨报 (8/17)：Mizuho NDR 后维持 $300，MCP server 用量强，安全仍在打磨 2 条
文件：KB/Substack/2026-08/2026-08-17-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Mizuho 与 DDOG 管理层 NDR 后报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Mizuho 在 NDR 后维持 Outperform 与 $300 目标价：AI 需求在业务中扩散，MCP server 用量强，Bits AI 扩展到 agentic 用例，新客户贡献健康；但安全仍在打磨，尤其 SIEM 和 Code Security。
A 2026-08-19 · ThirdBridge · 前 Datadog 产品营销总监谈 AI 遥测增长、多产品飞轮、最大 AI 客户依赖和 OTel 14 条
文件：KB/ThirdBridge/2026-08/2026-08-19-datadog-barking-up-the-ai-tree-scaling-telemetry-product-flywheels-enterprise-moats.md
发言人：Eyal Engel (EE)。Third Bridge 页眉写 'VP, Technical Product Marketing at Cyera Ltd'，主持人介绍为 'former Director of Product Marketing at Datadog' (明示)。自述在 Datadog 工作 5 年多，从 APM 产品营销 IC 做到团队负责人，覆盖 APM、软件交付 (CI Visibility/Test Optimization)、云与代码安全，离职前还负责 LLM Observability (现称 Agent Observability) 的产品营销；现任 Cyera (数据与 AI 安全) 产品营销 VP，带约 12 人 (L42)。离职时间未明示 (推测 2025 年前后)。视角局限：产品营销视角，不掌握财务、合同或客户账单数据，多次说 '不知道/只是预测'；作为前员工对公司有正面偏向；文中关键数字 (客户数、NRR、新客户贡献) 均为 Third Bridge 分析师在提问中给出的，不是专家信息。
阅读范围：全文 L1-L186 可信度：需交叉验证
专家整体看多：AI 带来更多复杂度和遥测、10 万美元以上客户还有 4–6 个产品之外的扩展空间、未来 12 个月增速维持约 30%、NRR 回到 125–130%，并强调公司指引一贯保守。但他也承认：对最大 AI 客户 '确实有依赖'；Datadog 在 OTel 上曾被大型金融科技客户逼着支持并重新考虑不同部署的定价；LLM trace 的采集本身没有护城河，壁垒只在平台层关联；AI 预算挤占和客户优化会在短期造成波动。
问题：专家头衔不一致：页眉 'VP, Technical Product Marketing at Cyera'、介绍 'former Director of Product Marketing at Datadog'、自述 'VP of Product Marketing'；Key Insights 中 '58%/52%、125-130% NRR、c30% 增长' 混合了分析师引用的披露数字和专家推测，应分开；专家预测 NRR 125–130% 高于分析师所给当前 '低 120%'，未解释差异；专家说约 35,000 客户，与分析师所给 33,400 不一致；多处 '不知道'，对增速数字明确说 'I don't know about the number itself'
B 2026-08-20 · TMTB (Substack) · TMTB 晨报 (8/20)：BNP 招聘数据——DDOG 招聘加速但完成配额的销售比例最低之一 2 条
文件：KB/Substack/2026-08/2026-08-20-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 BNP Paribas 招聘/销售生产力数据摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
BNP：DDOG 7 月招聘加速，销售岗位空缺比半年前多，是收入的潜在领先指标；但 DDOG 是完成配额销售比例最低的公司之一。
问题：配额完成率低也可能是配额定得高 (Funda 1Q preview 专家 C 称 DDOG 内部目标偏高，80% 完成已算好)，含义模糊
C 2026-08-27 · Merit 专家电话 (音频转写) · 企业 AI 调研 (保险)：AI 部署未带动 Datadog 等采购 1 条
文件：KB/2026-08-27/260827 - Merit 企业AI进展调研(2026-08-27 10_00.md
发言人：未具名专家，现任某保险公司 (金融业) IT 负责人，此前在软件公司工作 (原文 'I worked in software companies and also uh now I'm working in ... the financial sector for an insur...')。视角局限：单一保险企业样本，AI 支出以账号和 Token 为主。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
保险公司专家称部署 AI 后并未专门增加 Datadog、Snowflake、MongoDB 等采购，增量支出主要在 AI 账号、Token 和云存储。这是传统企业 AI 未必带来观测增量的反例。
问题：单一企业样本
C 2026-08-28 · Acecamp · 一个 Token 的旅程 (八)：企业 Agent 的非 Token 成本，顺带谈可观测平台 1 条
文件：KB/Acecamp/2026-08/2026-08-28-yi-ge-token-de-lu-cheng-ba-mo-xing-yi-jing-ping-jie-qi-ye-agent-wei-he-huan-mei-bing-wang-cong-guang-fu-fei-huo-cheng-be.md
发言人：Acecamp 专栏文章，作者未署名 (系列作者)；不是专家访谈，是作者自己的分析。
阅读范围：仅检索关键词段落 (1.5MB，内嵌 base64 图片，只 grep；L36/L39 的 'DDOG' 命中是 base64 噪声) 可信度：仅作局部线索
作者认为 Agent 调用工具更多后，身份、安全和可观测性需求比较确定，Datadog 等可承接运行监控；但利润归谁取决于谁能跨系统看到完整任务，单一告警功能可能被更大的控制平面整合。
问题：作者观点，没有数据
C 2026-08-31 · Merit 专家电话 (音频转写) · Merit 周度路演：Snowflake/Databricks 建 Token 与成本监控工具，想替代 Datadog 2 条
文件：KB/2026-08-31/260831 - Merit 公开路演：科技_制造周度调研重点更新(2026-0.md
发言人：Merit 研究团队 (路演主讲人，未具名)，转述渠道商和 Snowflake 专家的观察并加上自己的判断。视角局限：二手汇总，不是一手专家；主题是 Snowflake/MongoDB。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
Merit 认为 Snowflake 和 Databricks 正转向平台，自建 Token 消耗/成本监控和 Agent 评估工具，想在客户内部替代 MongoDB、Datadog；大客户短期不会迁移，小公司可能逐步被替代。
问题：属研究团队判断，未给替代案例或数字
B 2026-09-02 · TMTB (Substack) · TMTB 晨报 (9/2)：PANW 财报——Chronosphere 收到大型 LLM 客户迁移的 9 位数 ARR，DDOG -3% 5 条
文件：KB/Substack/2026-09/2026-09-02-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，PANW 财报点评，引用 PANW 管理层原话 (CEO/CFO 未具名) 与 TMTB 自己的分析。关于 '大型 LLM 客户' 未点名 (市场普遍认为是 OpenAI)。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：可作主要证据
PANW 管理层称其 4Q 净新增 ARR 含一个大型 LLM 客户从 '现有供应商' 迁到 Chronosphere 带来的 9 位数收益，DDOG 当日 -3%。Chronosphere ARR 从收购时约 $85M 升到 $500M 以上，4Q 有超 $100M 来自这一家，F1Q 还有不到 $100M；另有一个约 $20M ACV 的 AI 推理客户；PANW 说还需约 6 个月才能达到企业级功能对等。
问题：'incumbent vendor' 未点名是 DDOG，市场解读为 DDOG (DDOG 当日 -3%)；9 位数 ARR 是 Chronosphere 的口径 (价格可能比 DDOG 低 30–40%)，不能直接等同 DDOG 流失收入
B 2026-09-05 · Acecamp · 一个 Token 的旅程 (九)：谁在收企业 Agent 的'并网费'，Datadog 被称为'Agent 电表' 8 条
文件：KB/Acecamp/2026-09/2026-09-05-yi-ge-token-de-lu-cheng-jiu-ba-dao-men-kai-shi-shou-fei-shui-zai-shou-qi-ye-agent-de-bing-wang-fei.md
发言人：Acecamp 专栏文章，作者未署名 (系列作者)；引用 Datadog Q2 2026 财报和第三方电话会转录。视角局限：二手分析，数字来自公开披露。
阅读范围：全文浏览 L1-364，重点读 Datadog 段落 (L15-25, L80-115, L140-180, L222-232, L248-335) 可信度：需交叉验证
作者把 Datadog 称为'最像 Agent 电表'的公司：可观测性天然满足持续发生、可计量、任务绕不开三项条件，Q2 收入 11.2 亿美元 (+36%)，31 家 AI 原生客户年化收入超过 100 万美元，其中 8 家超过 1,000 万美元。但 Bits 没有单独披露 ARR，分不清'AI 客户用 Datadog'和'Agent 带来的增量'；按 20.5 倍 EV/Sales 定价，属于'高兑现、高定价'。作者还列出证伪条件：云厂商或模型厂商免费打包观测。
问题：数字转引自公司财报和第三方转录，应以原始披露核对；'Agent 电表'是作者比喻，不是证据
A 2026-09-08 · FactSet 会议转录 · Citi TMT：Pomel 谈非 AI 从 18% 到 high-20s、AI credits 重新打包、OTel 是顺风、价格通缩问题 21 条
文件：KB/FactSet/2026-09/2026-09-08-corrected-transcript-datadog-inc-ddog-us-citi-global-tmt-conference-8-september-2026-11-30-am-et.md
发言人：出席人：Olivier Pomel，Datadog CEO 兼联合创始人 (明示)。主持人 Fatima Boolani，Citi 企业软件研究联席负责人。视角局限：CEO 宣讲；分析师直接问了价格通缩/OTel，CEO 没有正面回答单价。
阅读范围：全文 L1-L481 可信度：可作主要证据
Pomel 说非 AI 客户增速一年内从约 18% 加速到 high-20s。agent 流量和 trace“爆发”，但他明确说这只是“指向未来”，还不是当下的驱动。Bits AI 收入“还看不到”，正从按调查收费改为跨场景的 AI credits，安全 agent 已和 Cloud SIEM 解耦，可用于不在 Datadog 的数据。被问 OpenTelemetry 带来的价格通缩时，他称 OTel 是顺风、毛利率维持良好，没有回答单价。AI 原生约占 15%，远低于疫情时云原生约 40% 的峰值。
问题：平均年化合同 $400K 语境是财富 500 客户还是全体，不清；价格通缩问题被回避；未谈最大客户、SBC、回购
A 2026-09-08 · GuidePoint · 前 Datadog 工程经理 (现初创 CTO 兼 DDOG 客户) 谈 Agent Observability、OTel 与账单 13 条
文件：KB/GuidePoint/2026-09/2026-09-08-datadog-s-observability-roadmap-an-engineering-leadership-view-on-ai-opentelemetry-and-the-competitive-landscape.md
发言人：Joe Cardali (JC)。GuidePoint 页眉标注 'Former Engineering Manager - Datadog'，Viewpoint: Former Executive (明示)。自述在 Datadog 工作 3 年 (约 2022–2025 初)，全程在内部数据平台，负责摄取客户云基础设施数据 (AWS/Azure/GCP) 的 intake 侧；2025-03 起任纽约一家 Series A 初创公司 CTO，现在是 Datadog 客户 (L149-153, L444-445)。视角局限：离职约 1.5 年，自称对新产品的认知已过时；不在 OTel 团队、不在 C-suite 决策层 (L193, L278, L303)；不碰销售/定价数据；作为买方只代表一家小型初创公司；关于客户迁出的案例来自更早 (约 2021 年) 的上一家公司。
阅读范围：全文 L1-L471 可信度：需交叉验证
前工程经理确认 agent 工作流让调用树/MCP 调用更深更密 (E03 支持)，但对 Bits AI 自主修复持保留：客户不愿给生产系统放权；他自己的公司 LLM 观测没选 Datadog。对价格面：OTel 让迁移 '更容易但不是改几个开关'，Datadog 内部曾就是否帮客户降量 (Observability Pipelines/成本提醒) 有长期内部张力，最后因 '大账单不满的客户' 让步；成本是客户迁出 (尤其日志) 的首要原因。
问题：专家离职约 1.5 年，对 Dash 2026 新品、Bits AI 现状基本是外部推测；关于 OTel 定价和内部张力明确说不在相关团队；客户迁出案例是约 2021 年上一家公司的经历，非当期；摘要表格中 '观测支出约为云支出 10%' 是经验法则，非数据
A 2026-09-10 · FactSet 会议转录 · Goldman Communacopia：CFO 谈最大客户续约但用量迁移、指引只计承诺额、Infinite Cardinality 与 Federated Logs 17 条
文件：KB/FactSet/2026-09/2026-09-10-corrected-transcript-datadog-inc-ddog-us-goldman-sachs-communacopia-technology-conference-10-september-2026-12-30-pm-et.md
发言人：出席人：David M. Obstler，Datadog CFO (明示，自称在 Datadog 8 年)。主持人 Gabriela Borges，Goldman Sachs 软件分析师。视角局限：CFO 口径；关于最大客户只给定性描述，不点名。
阅读范围：全文 L1-L248 可信度：可作主要证据
这场最直接谈到最大客户：客户已续约并延长合同，但部分不需要实时的指标/数据库负载挪到了别处，带来波动，所以指引只按承诺额计入。CFO 介绍 Infinite Cardinality (主动把客户转到新 SKU，能提高留存和毛利，但收入影响没说) 和 Federated Logs (日志可留在 ClickHouse/Databricks)。分析师直接问“预算流向 ClickHouse”，CFO 答“用户集中在 Datadog 的 ARPC 更高”，还承认会主动告诉客户“你发了太多数据”。上云比例只有 upper 20s-30%。
问题：最大客户不点名，“volatility”幅度未量化；Infinite Cardinality 对收入的影响被回避；“there's one that's been speaking about it now” 未点名竞争对手；未谈 SBC、回购
C 2026-09-10 · Merit 专家电话 (音频转写) · 企业 AI 调研 IV (LVMH)：Agent 增多带动 Datadog 等基础设施采购 2 条
文件：KB/2026-09-10/260910 - Merit 企业AI进展调研IV(2026-09-10 16_.md
发言人：未具名专家，LVMH 集团内部人员 (原文多次以'集团''我们'谈 LVMH 与 Google Cloud 共建 MAIA，具体职位未明示，推测是数字化/IT 方向)。视角局限：奢侈品集团单一样本；本人承认不清楚 Datadog 等的具体采购金额。
阅读范围：仅检索关键词段落 (grep datadog|DDOG|observability|可观测|观测|监控|Chronosphere|ClickHouse|Grafana|Dynatrace|Splunk|OpenTelemetry|OTel，读命中前后约 15 行) 可信度：仅作局部线索
LVMH 专家称集团会统一采购 Snowflake、Datadog 等 AI infra，Agent 数量增加会让基础设施消耗'指数级'增长；但他说不出具体金额，Datadog 采购混在整体 AI infra 采购里。
问题：'指数级增长'没有数据支撑
C 2026-09-11 · TMTB (Substack) · TMTB 晨报 (9/11)：Wedbush 接手覆盖给 Outperform，目标价 $275 1 条
文件：KB/Substack/2026-09/2026-09-11-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Wedbush 报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Wedbush 接手覆盖、Outperform、目标价 $260→$275，认为 DDOG 是云和 AI 原生负载增长的最佳纯标的。
B 2026-09-15 · TMTB (Substack) · TMTB 晨报 (9/15)：Bernstein 称 AWS 非 AI 负载 4Q 走弱，OpenAI/Anthropic 消耗可缓冲 DDOG 3 条
文件：KB/Substack/2026-09/2026-09-15-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Bernstein (AWS 网站流量数据) 与 Citi (DT) 报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Bernstein 的 AWS 网站指标到 9 月初仍软，3Q ex-AI 增速仍可加速约 50bp，但 4Q 非 AI AWS 增速可能比预期再弱约 100bp；OpenAI/Anthropic 消耗加速可缓冲 AWS 和 DDOG。Citi 称 DT 相对 DDOG 仍有明显估值折价。
A 2026-09-16 · GuidePoint · 前 Datadog 中端市场销售 (也在 Grafana 干过) 谈 AI 需求、Grafana 价差与 AI SRE 创业公司 10 条
文件：KB/GuidePoint/2026-09/2026-09-16-datadog-s-observability-roadmap-a-former-observability-sales-leader-on-ai-demand-product-edge-and-where-durable-growth-s.md
发言人：Michael Conrad (MC)。GuidePoint 页眉 'Former Mid-Market Account Executive - Datadog' (标题称 sales leader，实际是中端市场 AE)；自述在 Datadog 4 年 '经历 IPO' (推测约 2016–2020)，之后在 Grafana Labs 工作，现为 n8n (工作流自动化) 高级企业 AE 和创始 GTM 团队成员 (L126-130)。视角局限：Datadog 经验是 IPO 前后、中端市场客户，距今约 5 年以上；做销售不碰技术和财务数据；有 Grafana 背景，在 Grafana 时把 Datadog 当竞品卖，可能偏向开源叙事；现在卖的是相邻但非观测的产品 (L299)，当期一手观察有限；全文无具体客户账单数字。
阅读范围：全文 L1-L429 可信度：需交叉验证
给出 Grafana Cloud 相对 Datadog 约 20% 的价差 (且 '没以前那么大')，自托管 Grafana Enterprise 节省可能大得多；认为 Datadog 定价 '仍与市场不匹配'，且团队若能自己做 AI SRE agent 可显著削减工具支出。同时认为 Resolve AI 这类纯 AI SRE 公司缺乏数据底座、更可能被收购或淘汰；企业常把低优先级应用/开发测试环境/部分指标和日志分流到开源工具，这是 '量增但不一定流向 Datadog' 的具体机制。
问题：标题称 'sales leader'，实际页眉为前中端市场 AE；Datadog 任职在 IPO 前后，距今较久；对当期价格的判断主要来自 Grafana 任职期间；主持人提问中 '非 AI 客户加速而最大 AI 客户减量' 是主持人转述，非专家信息；L221 'migration fine' 可能是 'migration fee' 的转录误差；摘要页日期为 Sep 15 而文件为 09-16
B 2026-09-16 · TMTB (Substack) · TMTB 晨报 (9/16)：Truist 与 IR 会面——4Q 指引按最大 AI 客户合同最低额，3Q 跟踪高于底线 3 条
文件：KB/Substack/2026-09/2026-09-16-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Truist 与 DDOG IR 会面后的报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：需交叉验证
Truist：IR 重申指引对最大 AI 原生客户保守，4Q 假设按合同最低用量，而 3Q 实际跟踪高于这个底线；增长在产品、客户群、行业上都广泛，非 AI 原生客户因 '为 AI 做准备' 投资进入生产也在加速；安全 ARR 过 $100M，在 >$1M ARR 客户中安全支出占比 <2%，空间大。
A 2026-09-23 · GuidePoint · APM 独立顾问 Bernd Harzog 谈 DDOG/DT：AI 扩大 TAM，大客户为何续约但减量，Chronosphere 只吃走一部分应用 16 条
文件：KB/GuidePoint/2026-09/2026-09-23-ai-observability-datadog-and-dynatrace-the-3q26-and-2026-2027-outlook-plus-ai-impacts.md
发言人：Bernd K Harzog (BKH)，APM Experts CEO (页眉明示，Viewpoint: Industry Consultant)。自述从业 45–50 年，曾任两家风投初创公司 CEO、其中一家卖给 Citrix，此后做一人顾问公司；客户包括大型终端用户 (帮其评估/降低观测支出，如一家年付 Datadog 2,000 万美元的金融机构) 和厂商 (帮其竞争定位，暗示 Datadog 和 Dynatrace 都找过他) (L142-158)。视角局限：顾问视角、多为行业层面判断和轶事；自称只和厂商产品团队合作、不接触销售数据，刻意不掌握未公开业绩 (L716-719)；可能同时收厂商与客户的钱 (利益偏向不明)；转录中说话人标签多处错位 (L254-258、L653-657、L689-691)。
阅读范围：全文 L1-L765 可信度：需交叉验证
顾问认为 AI 从 LLM 监控、agent、vibe coding、自有基础设施四个方向扩大观测 TAM，Datadog 目前的 AI 收益只来自 LLM 监控，agent 控制层 '收入上还没开始'。关于最大客户 '续约但减量'：替代方案几乎总是 OTel + 开源数据库自建，但只有 '很大但应用很少' 的公司 (单一业务、单一应用) 才做得到，正对应 AI 实验室画像；Chronosphere 的成功模式是吃走 '大规模微服务' 那一类应用，其余留在 Datadog。他给大客户降费的第一步是清理闲置监控，说明大账单中有可压缩空间；但他判断预算环境已不再是 '零和'。
问题：说话人标签错位：L254-258 标为 ES 但内容是专家观点；L653-657、L689-691 标为 Bernd 但内容是主持人提问；主持人 '剔除大客户后 28–29%' 的口径未说明来源；L252 'AI will kill us all from Datadog' 疑似转录错误；L760 'Thanks for watching!' 为转录伪影；专家明确称无未公开业绩可见度，全部为行业判断
B 2026-09-25 · TMTB (Substack) · TMTB 晨报 (9/25)：Wolfe 目标价上调到 $340——agent 让 trace 变多，Meta Muse 可能用 DDOG 4 条
文件：KB/Substack/2026-09/2026-09-25-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Wolfe Research 生态检查报告摘要。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Wolfe 目标价上调到 $340：生态检查显示 agent 显著提高可观测性强度，每次 tool call、重试和交接都需要 trace，trace 成本约是 metrics 的 2 倍；AI 编程工具让开发速度提升约 2–3 倍；DDOG 直接参与 Meta Muse 的可能性上升；上行情景 FY25–28 收入 CAGR 约 32%，共识约 25.5%。DDOG 当日 +1.5%。
问题：'trace 约为 metrics 2 倍成本' 口径未明 (每单位? 每主机?)；Meta Muse 参与仅是 '初步交流'
B 2026-10-09 · TMTB (Substack) · TMTB 播客情报 (10/9)：Datadog CPO Yanbing Li 谈 agent 可观测性、机器级 rate limit 和 AI 成本账 6 条
文件：KB/Substack/2026-10/2026-10-09-tmtb-podcast-intel-qualcomms-qcom-ceo-on-ai-phones-semianalysis-on-the-ai-buildout-fcc-chair-on-spacex-spcx.md
发言人：TMT Breakout (TMTB) 日报，Podcast Intel 栏目 (由 Ethmos AI 研究 agent 生成摘要)，总结 Datadog CPO Yanbing Li 在 Stack Overflow Podcast 的访谈 (身份与履历明示)。'THE SIGNAL' 部分含 TMTB/Ethmos 自己的解读 (标注 our read)。二手转述，看不到底稿，看不到逐字稿；AI 生成摘要可能有偏差。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：需交叉验证
Li 称 agent 本质仍是分布式系统，logs/metrics/traces/events 仍是基础，新旧两层需求同时增长；DDOG 自己的数据显示大量 agent 错误来自撞上 API rate limit，客户不断要求 DDOG 放开自家平台的 rate limit (意味 agent 机器级查询 DDOG API)；一位客户用 SRE 调查 agent 几周，平均 3 分钟出根因、无需人工。她也承认 '通过轻量 MCP 访问、DDOG 自家 agent 还是第三方跨厂商 agent 消费遥测' 是一场有意思的争夺，AI 的成本/价值关系尚未理清。
问题：摘要由 Ethmos AI agent 生成，'paid AI-SRE adoption' 是 TMTB 解读，原话未说该客户付费；rate limit 是收费机会还是成本压力，原文标注为 TMTB 自己的推断
B 2026-10-12 · TMTB (Substack) · TMTB 晨报 (10/12)：Jefferies 目标价 $300 但维持 Hold——估值与客户集中度 3 条
文件：KB/Substack/2026-10/2026-10-12-tmtb-morning-wrap.md
发言人：TMT Breakout (TMTB) 日报，转述 Jefferies 报告摘要 (DDOG 与 DT)。二手转述，看不到底稿。
阅读范围：仅检索关键词段落 (Datadog/DDOG/observability/Chronosphere)，多公司日报 可信度：仅作局部线索
Jefferies 目标价 $260→$300 但维持 Hold：核心 ex-AI 原生增速 2Q 加速到 high-20s，750+ AI 原生客户利于 agentic 可观测性；但最大客户预计 3Q 起减量，形成 2027 年集中度/优化压制，约 19x CY27 收入使风险收益不吸引。同日 Jefferies 看多 DT，称其消耗增长 20%+。