2026 年 software 反转的赢家，靠的是什么

Thanks, Yuka, and thank you all for joining us to go over a very strong start to 2026. Let me begin with this quarter's business drivers. I'm very pleased to say that our teams executed very well and delivered revenue growth of 32% year-over-year accelerating from 29% last quarter, and 25% in the year-ago quarter. We showed broad based acceleration of revenue growth across cohorts including both our AI, and non-AI customers. Our AI native customers' cohort continued to grow and diversify rapidly both in the number of customers we serve, and the scale of those customers. 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.

Thanks, Olivier. This was a very strong quarter for Datadog. 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. 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. 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. RPO duration increased year-over-year as the mix of multi-year deals increased in Q1.

Yeah. 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.

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. I wouldn't, give that much weight to the very specific word, we're deliberate but not all that deliberate. And similarly, both David and I have a raspy voice today, but there's no hidden meaning. But I will remind everybody that we did not change. So if the question also, I think you asked is did we change or is this a different methodology of both the overall and the large customer, then the guidance, the last quarter or the previous. The answer is no. It is the same methodology, and that we've had, so no change. But this has been what we've always been doing.

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.

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. 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.

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. 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.

Well, look, if we provide more value to, we'll get more. As I was saying earlier in the call, like, we sell more software by helping customers make more money or save money or both. And I think if we can automate more and let them do more, we'll provide more value. That's as simple as that. I think the future observability is not just observing, it's fixing, it's not waking up people in the middle of the night because something broke, but fixing it for them, it's not letting people do a damage control on a security incident because an attacker is in, it's preventing the attacker from getting into start with by auto-remediating issues. And we're very, very busy building all of that. 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.

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. 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. So, in a weighted, and you see that in our gross retention, the vast, vast majority of our customer base, given the gross retentions we're talking about, have stayed with Datadog. And the fact that the net retention has been increasing means that they have been putting more of their workloads on Datadog rather than less.

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.

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.

Yeah, Saket, I think it's a pretty wild time. I was asked in the end of 2025 – in November 2025, when I thought that non-human traffic would pass human traffic and we pulled all the data, we ran all the numbers and we were pretty confident that it was going to be the second half of 2027. I was asked the same question again in March of 2026 and we did the same exercise and were surprised to see that it had moved up to be – that would cross in the first half of 2027. So, I was quite surprised when in May of this year, our team came to me and said, you won't believe it, but non-human traffic is now passed human traffic online. And to give you a sense of how this trend is playing out and with the big caveat that I have called it wrong at every point along the way, if the current trends continue, we think in five years, non-human traffic will be as much as 1,000 times as much as human traffic. In other words, humans will be a rounding error on the Internet, not because human traffic goes down, but that's just how fast we're seeing non-human traffic grow. And so, that's resulting in a number of things. The first is that for some of that non-human traffic, it's malicious. And that could be malicious like it's hackers or bad guys. It could also be it's malicious from the perspective of a particular customer's business model, where it's traffic that is maybe an AI company trying to take the content from a media company that relies on advertising. And in those cases, we block that traffic and we don't charge the customers anything more for blocking up that traffic, because we think that that's the right thing for us to be doing and delivering. And that's part of being a security company. At the same time, though, there are some people who want that traffic. And so, we're doing everything we can, not only to serve that, but to make it as efficient as possible to serve it, because if we're going to have 1,000 times as much traffic online, we've got to get a lot more efficient and companies like Cloudflare are critical to be able to support that for customers, whether they're large and small. The thing that I think is most interesting, though, is that, increasingly, as we talk to the AI companies and over 80% of the major AI companies are Cloudflare customers and rely on us as we have a great relationship with them and as we talk to others, it's clear that as agents are accessing all of these sites and the volume that they're accessing them on, the sort of give to get that you have with human traffic is different. And so, if you look at some of the things that we've announced this week during Agents Week, things like cloudflare.pay, that's us setting the foundation to be able to say how do we charge agents some – again, what will be a very, very small fee, fractions of a penny for every request that goes through, but for the request that pass through that traffic, because somebody has to pay for the bandwidth, somebody has to pay for the server, somebody has to pay for the people doing the work to create the content. And I think that the business model of the Internet for the last 27 years has been largely defined by advertising and really defined by Google. I think the business model of the next 27 years of the Internet is going to be very different and there's no company in a better position to define what it looks like than Cloudflare.

Sure. So, I will start out not fully answering your question and then Thomas can continue to not fully answer your question. I think we haven't broken out by product area what our revenue or revenue growth rate is, but I can say that the Workers platform continues to be something that really drives an enormous amount of both new customer adoption and real revenue growth. And I think for a while, we were saying that we were playing for adoption, we weren't playing for revenue. I think that has tipped and we are now very much, again, not trying to maximize every dollar by any means, but that Workers has become a meaningful contributor to revenue. I think it shows up in the anecdotes that I gave about customer wins. More and more of them are signing pool of funds deals because of the fact that Workers is part of it and that their developer teams are saying, hey, yeah, sign with Cloudflare, use their Zero Trust, use their reverse proxy services, but also make sure that we've got something available where we can use their Workers platform. And so, that tends to drive that and drive a lot of consumption through that as well. I think that – so, for sure – and again, I think you see this in the paying customer count that the PLG, kind of product-led growth, is definitely going great. But you're also seeing, if you just look at like – you do this across earnings calls, you add up, we give basically the same number of customer win examples every time. If you just added up the dollar amount of those contracts and figure out what the average is, those have been ticking up and up and up. And I think that that is still very much a function of a great enterprise sales team where people want – if you're going to sign $10 million plus contracts, no one is putting it on a credit card. Those are deals that humans negotiate with each other and there needs to be a level of trust and relationship and a knowledge that if something goes wrong, someone's there to fix it. And that's what great enterprise sales leaders do. What I think our enterprise sales team is doing is getting more technical. And again, that – well, even those people who are great salespeople, but they weren't particularly technical, I think we're giving them the tools to become much more technical. And I think over time, I mean, what I've challenged our team and the sales team, and I think that I will – that this challenge will be met fairly quickly, is who's going to be the first salesperson who is in a meeting and while sitting in that meeting, delivers a feature before they even leave the meeting and it gets shipped to production. We're building the system to be able to do that. And I think our sales team has leaned in to say, yeah, it's really important to still go build relationships, but we need to be technical. We need to be able to understand and talk to the technical leaders that are on the team. And at the end of the day, the best product is still what's going to win.

Yeah, the trajectory of the business is clearly up and to the right for both revenue as well as operating income. But as we said at Investor Day, their business model is also going to evolve. It's moving away from a purely ratable SaaS model towards a much more diversified mix of pool of funds, consumption-based structures and also what we call T-shirt sizes. As the business accelerates, that also leads to more and more customers burning through their T-shirt sizes faster, pool of funds are getting consumed faster and getting renewed. So we are really stacking these things on top of each other. And that leads to, you would call it a higher standard deviation across the individual customers. So the momentum over a multiple quarter is clearly pointing up into the right. You don't – you not only see this in revenue, you also see this in RPO and CRPO, but moving – how we move from quarter-to-quarter becomes a little bit more difficult to forecast. That's why we continue to be prudent in terms of how we set the guidance, but the trajectory is clear – is clearly up into the right.

Sure, Patrick. So I think not all revenue is created equal. And if you're selling what is just commodity compute, if you're basically letting a AI company use your balance sheet and your credit rating in order to buy servers that are the same as everybody else's servers, then like that's just not attractive business for us. And so I think that we're very disciplined. It's not that we don't get asked. But I don't think that's particularly interesting for us. I will say that we have a lot of interaction with the largest AI providers that are out there. And they're – again, they're all customers of ours. And so I think that it's – that you could imagine in ways where – that we would find something that where we could actually be a bigger and bigger player in providing the services to power more and more of the serious AI players that are out there in more meaningful and deep ways. But ultimately, like, we're not going to do that just selling commodity hardware because that, again, is bad business for us. I think the other thing that's different about us is that we've just always focused on how we can squeeze as much out of every CapEx dollar as possible. And so how is it that we can, not just throw money at a problem, but actually say how do we get more utilization out of that? And so our team is constantly trying to figure out how can you get more inference, how can you get more CPU, how can you get more out of the memory or storage that we have in every box that we have. And I think that that's just a very different attitude than you see elsewhere. And so, again, we know what the typical kind of inference loads are at the hyperscalers and their GPU utilization. And it's super low. And that's not the hyperscalers fault. It's that what they're selling is just a box and it's up to the customers in order to actually maximize the utilization. And customers often don't have the diversity of traffic or the ability to really schedule things in order to get the highest possible utilization. And so if we can, and in some cases, get 10 times as much utilization out of every CapEx dollar, then we can play these games and continue to grow revenue, but it's really high-quality revenue and it's revenue that doesn't require us to just throw money at it. Again, I think it's worth repeating things that somebody has said before, which is we're just in a very different business than the hyperscalers. The hyperscalers, the traditional first-generation cloud, are in the business of buying a server and then trying to sell it back, lease it back and get 5 turns of revenue off of it. We are in a very different business, where we're selling actually work getting done. And then instead of it being up to the customer to get as much out of the server as possible, we need to do the work to get as much out of the underlying equipment as possible and build the scheduling and efficiency and everything else that's out there. And I think that that is fundamentally the difference in our business and why like I don't think we want to get into the business of renting servers because, over time, it's a commodity business and it's not very attractive.

To your first question about traffic, I think the key here is something that we've always understood, but I'm not sure the market has understood as well, which is not all traffic is created equal. A lot of the times, if you were sort of a traditional CDN customer or company, what you're trying to chase was things that drove lots of bandwidth, so video streaming of live events and things like – and things like that. And again, that was an okay business. I think it had a lot of pressure. It was largely a commodity. But that was kind of the game of being a CDN. We never saw ourselves that way. We thought what we wanted to get in front of was the most essential traffic that was out there, not bandwidth and video streaming, but APIs and applications and seeing that. And I think that in this new world of agentic commerce and agentic transactions, our approach is showing kind of the wisdom and durability of that. So today, literally, we're seeing hundreds of billions of agentic requests per month, and that number is growing exponentially. They are interacting with us, and we are setting what the rails and the guardrails are for that, and that's driving our Act 1 business. And then on the other side, in our Workers platform, we have built a platform that allows you to build agents that are just significantly more efficient than anyone had before. And so across all of the parts of our business, because even in the Zero Trust and SASE space, it turns out that having more fine-grained controls about data is exactly what you need if you have kind of these somewhat new agents running around doing things. You want to make sure that they only have access to the things they should. It's – I wish I could say that we saw all this years ago and built Cloudflare for it. But I think that the reality is that we happen to have built exactly the right set of tools for this moment. And I think that's what's separating us from some of the people that we sometimes get compared to.

Yeah. I mean, I think for the Act 3 products, the developer platform products, I think it's the most obvious which is that we're just seeing more code being created. It needs an efficient place to run. It needs to have a set of primitives that can act as fast and as ephemerally as agents need to act. So if you imagine that you're a company that's building something that's going to plan a vacation for someone, what you really need is the ability to essentially spin up an extremely lightweight sandbox that writes code, that assembles kind of all of the different parts of what that vacation may entail, puts it together, pulls it back and then blows the sandbox up very, very quickly. To do that with any other platform is extremely expensive and slow. And we've made it with some of the things that we've announced during Agents Week and otherwise, we've made that just drop-dead simple. And so, that's driving a ton of use. And again, I think that the way to see that is the fact that we added 1 million developers to our platform just the last quarter, almost as many as we did in all of last year is extraordinary. And if you look at growth across the Workers platform, more than three-quarters of that is from new customers. And the growth rates are pretty extraordinary across that. That's also, again, to some extent, putting pressure on gross margin because those are less optimized for gross margin products. But we are confident that as we can continue to drive GPU utilization and all those things, that we'll be able to get more out of that. I think to the other part of your question, just as more – every time an agent does something, like if you think about it, if you just – if you type something into ChatGPT or any of the things, like, to search, the number of sites that get searched, the amount of traffic that gets generated, if I'm looking for a digital camera as a human, I might visit five websites, if I really care about it. My agent's going to visit 5,000. And so, that's going to just drive significantly more usage which is the biggest driver of kind of our Act 1 revenue. And again, I think unlike some of the pure-play CDNs that are out there, the agents aren't going to go watch reruns at the Super Bowl. They're going to drive things that actually drive real traffic to real e-commerce sites. And that's where Cloudflare and the huge swath of the Internet that sits in behind us is really valuable. And then for Act 2, again, as we talked about already, I think being able to very narrowly define what data an agent has access to and what data they don't, we're just seeing more and more of that usage, especially in the self-service category, which there really isn't another sort of SASE, Zero Trust, self-serve competitor out there with any sort of scale. And so that's with things like OpenClaw driving a lot of usage there. And what we've found time and time again is, as hobbyists or individuals adopt a technology, they inevitably start to bring that technology more and more to work. And that's what we're seeing as we win more of the enterprise accounts across Act 2.

Well, we – one of our Founders, Michelle, we make fun of her, because every year, when we have kickoff for the year and talk about the targets for what is to come and how we approach the year, we are only getting started. And now, this is our eighth year as a public company. So the beginning of the year, she said that the eighth is time. And I truly think that what you just said is true. When we think about our life in Act, Act I, Act II, Act III and Workers AI, we'll talk about it, Act IV, content protection, how we monetize it, we still – we think that that opportunity is going to thwart (02:04) what is already behind us. Last year, we were sitting here and agentic AI was still far away. The step-up of the LLMs only came in November and we were primarily concerned about convincing you all that our sales execution was on track, but productivity was coming that we felt comfortable investing in sales capacity, and that was driving the early steps of acceleration and growth in a quite significant way. Now, I think we are what you described in this paradigm shift of what an agentic AI world is going to do to us. And we find ourselves at a point where the network, the architecture, the biggest competitive mode of Cloudflare, the network itself seemed to be designed for what is going to come and support what agentic AI workflows are looking for, super low latency, secure Sandboxes to execute code and so on. So, what you look for are the signs that signal this early – the early innings. And because of the architecture of the network and the benefits, developers are flocking to Cloudflare. So, that is an early sign that I think is signaling this dramatic shift. Developer count was up 2 million developers quarter-over-quarter alone. You'll see that the acceleration of revenue is now happening in what we call variable revenue, right, so, consumption-based revenue. Another big shift, I think, in the agentic world is that we move and disrupt SaaS from a seat-based metric into highly variable consumption-driven revenue. That is where we see acceleration. We're at a point now where pretty much all the large names of AI native companies, whether it's Lovable, Sigma, Anthropic, OpenAI are building on Workers on top of the network. And one of the stats we gave at Investor Day was that we think about 80% of the significant AI enterprise infrastructure is a customer of Cloudflare. So, all the signs you want to see. And there's one trend that I think is signaling what is going to come when it – when we talk about Act IV. There was this vertical that we always were highly underrepresented. It was content creators. Media was always a vertical that was revenue weak at Cloudflare because there was not a lot of value for us just moving bytes from one place to another. And they are now shifting towards Cloudflare to start to protect and make their content that they create scares (05:04) in order to generate leverage. So, I think there are so many structural signs that signal what is coming. And that is, to be very honest, what gave us the confidence at Investor Day to raise our North Star, right? We said for many years now, 40% was our North Star. We shifted that to 50%. We'll exit this year with significant momentum in that direction. And I think there's clearly a world, and that is what we also said beyond 2027, where we will live in a world beyond the Rule of 50, and that will be the result of acceleration of Act III and Act IV.

Well, if you think about – so if the upside comes from variable consumption, it is more agentic AI traffic, it's more requests, it's faster adoption of products across pool of funds. If it's a pool of funds contract, people burn through their commitments way faster. We renew mid-quarter. That is the dynamic around that. But you asked for structural signs of us participating. I think this is one of these structural signs where we see that this flywheel is starting to spin and it's probably going to spin faster and not slower moving forward.

So, today, as you can see, we have four primary models of how we generate revenue. The easy ones, the flat rate, this is where we came from, where the world was easy on the IPO-ed and almost 100% of our revenue was subscription based. There is a larger share of pure usage-based revenue today, still under 10% where customers pay us in the rear. The complexity comes from the two other components of how we interact with our customers. Cap models and pool of funds. And you'll find out has a lot to do with T-shirt sizes and Nike Swooshes. And why is that? So, the majority of our contracts are based today on what we call a cap model. And requests in these contracts are the most significant measure for revenue. So, what does this mean? Think of a T-shirt you buy. A customer buys a T-shirt, commit to a ceiling of utilization of a product (01:51:38), for example, type buys, medium-size T-shirt, 10 million requests per month. And it gives us predictable and smooth revenue as long as the customer is using under this umbrella of what he has committed to. But what happens if a customer outgrows that medium-size T-shirt and has to go out and buy a larger one, let's say, small – from medium to large to $20 million requests per month? The traffic is increasing (01:52:09) more than the customer has planned. He needs to buy up. But however, remember, the customer is not going to buy what he consumes in the moment. He's trying to estimate what his future consumption is going to be. And this transition is letting our revenue grow, but it will lag to the inflection point of usage and makes, a little bit of sympathy from all of you, our job of guiding and forecasting a little bit more tricky.

AS: Still, if you were to compare between Q3, which is ending September, versus the Q2, I would say it's pretty much flat from a growth perspective. So we have not lost any business, but we haven't gained any significant business either. But starting into Q4, I think the momentum has picked up a little bit. The pipeline is looking pretty healthy as well for Cloudflare from a perspective. So that's where we are overall. The Cloudflare business is healthy for us from a year-over-year perspective. It's not concerning at all and a very healthy business.

AS: Okay. That area is definitely picking up the conversations because I think Cloudflare or maybe some other industry vendor released some of the report where the bot traffic is gaining more and more on the internet. So as the bot traffic gains the traffic momentum and tokenomics come into play, if you couple that together, I think the solutions from Cloudflare, which is probably unique by itself because no other vendor that I know of provide the kind of bot visibility and bot traffic scraping solutions out there. So media clients are definitely asking us looking into it. The pipeline that we have strong today is probably some of the mixture of that as well. So how much we have closed, it's difficult to say that the majority of the business that we have closed is coming from the bot scraping solutions from Cloudflare, but clients are asking about that and the momentum is picking up around the bot traffic visibility.

I would roughly call it even, our AI products which is a pretty broad swath at this point. Absolutely, it's CoCo and CoWork, but it's also things like AI functions that make data operations proceed at an impressive scale, or even newer products like the AI Gateway. They contributed approximately half of the acceleration that we are seeing. But there are a lot of other products that are also demonstrating robust growth. And Brian touched on some of them. Whether it's Notebooks or applications written in Streamlit or React that are deployed into Snowflake and, of course, migrations themselves going faster. I have talked pretty much in every single earnings call over the past six quarters about migrations, and that is an area where we continue to get faster and faster. And some of the recent advances, both in models and harnesses, are letting us run long-duration tasks of a scale and complexity that we haven't been able to do before. And the rate at which workloads are coming onto Snowflake is also an important factor. And one anecdotal example that a big network equipment manufacturer is doing a Teradata migration in less than three quarters this year, and this is something that would have taken probably two to three years in any previous time. So, these are some of the things that are contributing to our acceleration and beat.

Yeah. Thanks, Sridhar. I'll unpack that a little on terms of impact. And so, CoCo had the largest driver to the increase in our forecast. As a reminder, when we forecast we only forecast observed behavior. And as Sridhar mentioned, that just happened in the quarter. And so this quarter we had a very unique opportunity to layer CoCo in the model, and that's reflected throughout the remainder of the year. We also saw acceleration in our core business, and that informs our outlook as well. And so there's no change to our guidance philosophy where 3% we view as a really strong beat.

Yeah. Karl, thanks for the question. You're absolutely right. Our AI products have a lower gross margin than our core platform. The one thing that we want to do with our AI products when we launch a new product like CoCo is make sure that we develop a great product, that we get massive adoption, and we've seen really good adoption with CoCo. We're up to roughly about – a little over 7,000 accounts have adopted CoCo. With that said, we're offsetting that and keeping the same product gross margin at 75% for the full year, and lower bandwidth costs, i.e., I talked about the AWS contract, and so we're offsetting it there. So that's how we're able to do that. We're committed to find efficiency to be able to maintain that 75% gross margin.

Within our existing base, customer expansion is healthy, as evidenced by our net revenue retention rate of 126%. This expansion is underpinned by growth in both migrations and AI use cases. In Q2, 48 net new customers surpassed $1 million in trailing 12-month spend. We now have 828 customers spending above the $1 million threshold. Remaining performance obligations grew 30% year-over-year, totaling $9 billion. As a reminder, we continue to see customers favor Q4 renewals. As a result, we expect bookings to be increasingly weighted towards the fourth quarter.

I think, great question. I think we've seen early stage traction on CoCo. I think it converts faster than CoWork. And I think that's a scope difference rather than product quality difference. CoCo is just a coding agent. A lot of people are already used to coding agents for local files, and it's leveraging Snowflake data, which was there. The reason for that is because when the model answers wrong, you immediately go back to the data, right, and the architecture that powers it. And that is when that developer notices immediately, hence CoCo just becomes a comfortable, familiar interface. CoWork, I think, is where I'm personally interested because it's aimed at business users making decisions about output, but that ICP comes in afterwards, right? It directs the whole question. I think business users are really worried about the insight that is to be trusted or not. I think what we're also seeing is early stage defensive posture on costing, right? It's more conservative, and the problem is that ad hoc queries don't behave like scheduled dashboard refresh, right? It is hard to forecast; a finance function that can forecast, it's a warehouse spend, which was easy, right? We could have predicted a few % every quarter, every time period. Now, what we don't know is if someone asks a question, what is the predictable token users associated with it? I think that does stall some decision-making within the organizations. So that's where some of the decisions get stalled because there's more analysis and paralysis that comes with it. At the same time, I think a lot of that is necessary to really model the governance structure so that we can at least from finance to functional technical aspects, be more reliable in the ability to build business cases and the contracts that are associated with those cases accordingly. And then the movement happens. I think right now, fundamentally, every client of mine is worried about compute costs.

Yeah. So, we actually have token-based pricing as part of AIDR. So we already have that today. And what's important from a customer perspective is they do want some certainty on what they're paying. So, we provide a certain number of tokens across an environment, and obviously we scale that to how big the customer is. And then, if they exceed that token usage, they can buy extra token packs. And again, the beauty is it's all consumable via the Falcon Flex licensing. So, it's a model I think that has served us well, where it's not runaway cost for customers; they can define it, they can budget for it. But certainly, as they scale up their AI, if they exceed their budget allowance, they certainly can uplevel the tokens, and it's friction-free to do that with Falcon Flex.

So, on the first point, it is separate and incremental. It's another module and it's priced separately, which is, again, where we're seeing a lot of the growth. So, that is certainly a good fact to make sure that we clear up if there was any confusion about it. When we think about the beauty of the architecture and the platform, it's still using the same agent. And if you want to enable AIDR, again, that's a license entitlement. It's part of our friction-free deployment. Again, why is Falcon Flex so successful? Why have we been successful immediate time-to-value? You don't have to roll out yet another agent. So that's very important. And when we talk about sort of like the ramp of this, it's been incredible over the last couple of quarters, and we've seen almost the results, sort of, parallel some of the frontier labs' growth. As they continue to add more agents and capabilities, you have customers that are deploying more AI. They actually want to deploy AI faster, and they need to scale incredibly fast. So, when you combine what we're building with – again, it's on the same platform, that's the beauty, but we're able to monetize it separately. We think that's a home run, and we're showing the value to customers, and they're willing to pay for it because it's very unique in the environment. And they don't want to deploy yet another agent, right? They're getting it all with CrowdStrike, and that's what they're looking for.

Yeah, I think you have to look at just a crazy inflection point in the adoption of frontier type models, right, and we've all seen the revenue curves with the big players that are out there. So, we're tending to follow the slope of that adoption curve, and the token spend has been just jaw dropping. And when you have that level of adoption, you're going to have incremental funding around security. I think if we look at this realistically, two years ago, there wasn't these big token budgets, right? All of a sudden, there is. And people are finding money and it's incremental and they need to adopt more of it. We have plenty of companies that are tracking token usage not only from the spend standpoint of what are they spending, but they want to know who is spending and are they spending enough or are they going fast enough. So, we've seen it being incremental. And at the end of the day, getting back to something I've said time and time again, if you want to create AI, you need GPUs. If you want to use AI, you need security. And that's what we're finding.

Yeah. I think, Brian, the numbers back it up. Look, you start with the record Q3 pipeline. So, that prompted us to raise our full-year guidance by an additional 630 basis points. That means we moved our FY 2027 net new ARR growth rate from 27.7% at the end of last quarter to 34% this quarter. When compared to our initial guide for the year, we raised the growth rate 1,150 basis points or $116 million in just two quarters. That talks to the durability of the things that George talked.

So for those that aren't familiar with Flex, it's the most elegant go to market, I think, that's ever been invented, and I wish I could take credit for it. It was the customers coming to us and saying, hey, you don't have 25 modules, you have 26. How do we get access to all of it? So we kind of work through with our customers what would make the most sense for them and we wanted to make it really easy. I mean, as I said earlier, George is from New Jersey. He's just how do we make this so easy for customers to purchase. So we come with this license where basically, you know, you negotiate once and then you've got a deal for a period of time and then you're able to add more but not renegotiate terms. It's already done. It's just adding another PO to an existing contract where the blocker is me, the CFO. But if the business unit owner goes to the CFO and says, Hey, I need more for CrowdStrike, but I'm going to give you back X for something else and the X is going to be more than what we're asking for CrowdStrike, well, you don't need to be a CFO to do the math to say, okay, let's go do that. In the entire life of software and security for sure no one really was able to connect, grab on to a license where you have three winners at the same time. You have the economic buyer, you have the vendor, you know, us and procurement. I've never seen all three of those win at the same time. So the economic buyer wins because they're able to consolidate, get the best outcomes at a cheaper cost. We win because we're locking in our customers for longer, bigger contracts and being bigger supporters and bigger kind of advocates and trusted advisors and procurement wins because they're after the biggest discount. The biggest discounts are our Flex contracts. They get paid based on additional discounts they're able to achieve. And guess what? Flex allows for exactly that. So you have three winners and then you saw some of the stats like we are now have over 2,900 customers on Flex. You mentioned the 40% uplift when we move somebody from a non-Flex to a Flex, that's up from 34% the year before.

CrowdStrike had a unique trajectory in terms of like discounting practices because for a long time, they were much tighter on not discounting much. Certainly, they do discount, but not discounting too much compared to the competition. Then the big outage in 2024, they were able to retain 97% of customers is the public figure that they released. That was successful. A big way that they did that was by offering significant discounts to customers to say, "Hey, stay with us." Now in 2026 is when a lot of those customers that received a heavy discount for one year or two, some of them have already come off that, but some of them may have received discounts that lasted a little bit longer. Now, we're at a period where CrowdStrike has returned to its normal pricing methods. I see them needing to do a little bit more discounting than in the past, in part, because just the competition is very high. Palo has kept pace. While I speak negatively about Microsoft in terms of efficacy, which is, I think, very true, Microsoft does compete heavily with CrowdStrike and does win a lot of deals still because they're priced much cheaper.

That's a great question. So identity, cloud, and endpoint, they're not going anywhere. You can't just shift the spending and neglect those items. So in my opinion this has to be new budget that companies have to find, create, etc. Because this is a new vector of attack, this isn't something we've dealt with in the past.

I think this is an interesting question, and I think we need more time to really understand it because the frontier models price by tokens. It is somewhat accepted that that's how AI is priced, but I don't think that's going to last. I think that's going to shift as no one really knows what their AI token usage is going to be this month, this quarter, this year. I think everyone's trying to just figure it out and the easiest way to get started on that is the token-based pricing that we see today, but I think that's going to change over time. Again, kind of how we spoke about the SIEM piece. Customers aren't happy about pure ingestion pricing. They're not going to be happy with tokenized pricing here where it limits their use and their spikes and unpredictability. So I think that's going to change over the next year quite dramatically.

So, Ryan, first is that AI natives what we are finding and I shared the example of somebody like ElevenLabs at .local in London a few weeks ago. They were using first party database for operational data. They were using another software for search, and basically most of those product lines were really choking as ElevenLabs was growing significantly, right. They are now at a $500 million ARR. So when asked the team, technically the engineer who made that decision saw that the growth of the company as in that AI native company ElevenLabs was being held up by the data layer and us having search, vector search, and operational data in a single platform. They made the decision to move to MongoDB not too long ago. And two things they said that really resonated with me, Ryan. Number one, they are like, gee, we should have done this a lot sooner, otherwise we would have not to deal with all these outages and other things they dealt with the previous platform. And number two, now choosing MongoDB, even though they have scaled significantly on their ARR as an AI native company gives them peace of mind. I'm hearing them from other AI native companies who also chose maybe a Postgres or something, and Postgres completely choked on the performance. So that just gives me a lot of confidence that if AI native company where AI is the business or agentic layer is the business, and they feel that they can scale with MongoDB, when that moves over to the enterprises, whether banks, healthcare and other firms, they will also realize the same thing a little bit later. And as Mike shared and I shared earlier, the contribution is there. We are seeing very encouraging signs right now, but a lot of growth was still driven by core enterprise workloads, which I would argue are also getting ready for AI work (00:41:06).

So, as we talked about, we feel very good about the Atlas business. And what we look at is this was the fifth straight quarter of approximately 29% year-over-year growth, very consistent. We've increased the full-year guidance by 300 basis points from the previous guide, and that is also buttressed by a record net new $127 million net new Atlas dollars, as well as the increase in the net ARR expansion rate, and now Atlas is almost a $2.3 billion run rate. So as we look forward, we continue to expect really good growth from our larger enterprise customers, especially in the US. We started to see some benefit from AI even though it's small, but we are excited about the momentum, and we do expect consumption to continue to be consistent with what we've seen during the first half of the year.

Yeah. Thank you. Great answer, CJ. I just want to underline what he said, is our guidance philosophy, Alex, has not changed in terms of how we guided the rest of the year. We'll always be prudent more than a quarter out, and that's what's reflected in the guidance. I also want to address your comment about Q4 and just be super clear on the call. Hey, there were no large bundled deals in the quarter. There was none of that Q4 dynamic this quarter.

Absolutely. So, what I have seen, and this is across many conversations, Karl, is – I'm just going to first look at the bucket of enterprise, okay. Enterprise is as in whether you want to say Global 2000 or Fortune 500 or Fortune 100. When I look at those, MongoDB was always – almost always a database platform that was used for customer-facing workload. Whether it's insurance claims, healthcare policies, whether it's credit card transactions and getting the fast data for the end users, MongoDB always shines when it is a massive workload that is customer-facing. What I'm seeing is initially, say you are a wealth manager at a bank and there are lots and lots of knowledge base articles that you want to vectorize, use our embeddings and then use as a chatbot for folks that are doing wealth management and talking to clients real time, that is one very specific example where a particular large bank is using MongoDB. There are also other examples where because there are lots and lots of documents, employee-facing use cases where knowledge base articles so that employees can leverage, do a search because now search is fully integrated into the operational data and documents get loaded, and then embeddings make the vectorization better, that will be another large enterprise example where we are seeing use cases. But the clarity that I got was that it was almost always, hey, we want to use MongoDB where the scale matters on the agents that we are trying to create for our customer-facing activities, whatever the customer-facing activities are. We are not seeing early traction with, hey, I created a Copilot kind of a thing that appeals to a couple of hundred employees. We are not seeing MongoDB being used because they are like, hey, this is too big, MongoDB, of course, gives us scale and all this other functionality. So that's number one. And, Karl, the other thing I would say is when you look at AI natives, including I'm going to put Frontier Labs in there, but you look at the example that I shared which was on EU or whether it was Fireflies, which are agents in production, but when you look at these agents in production, we have used ElevenLabs in the past and others, these are millions of agents in production that are doing something that is customer-facing, and they would say, we want to use MongoDB for scale, performance and, of course, run anywhere, and that's where you're using it. So these are the two vectors that I'm seeing. In enterprises, our agents going into production, where it makes sense on Atlas, okay, let's do that. And on EA that I touched on in my prepared remarks, make no mistake, as they – these regulated industries are trying to get their operational data AI-ready, was also the reason EA growth was driven across the industry including tech where customers says, hey, I'm building an AI agent for my tech, whatever technology platform uses MongoDB or technology company, and we saw the growth there. So that would be my overall summary on where we are seeing, and I'm going to have Mike comment anything additional if you need it.

All right, I'm going to give you a second to stare at this chart. On the left – your left is the number of customer count in the AI native cohort that I just defined for you. On the right is the revenue for Atlas and the percent growth. As we sit here today, AI natives make up less than 5% of the Atlas revenue. I'm just going to cut the question off. When you think of that, think of 3% to 5% as we sit here today. Okay? But look at that incredible growth, 65% (sic) [67%] (03:21:05), 129%, 165%. We talk a lot about this which is it's still small, but it's growing very, very quickly and this is why we're comfortable with this continuing to drive growth in the future and it's off of a small base, but it's super powerful.

Sure. So, as we look at Atlas, and we've said it multiple times in the last several quarters, we continue to feel very good about the Atlas business. Technically, going from 3% to 5%, the 3% was a big thing, Alex. It allowed us to say, hey, what do we really see call it in the near term? And we've seen all the growth that has been driven by the move-up market, which has been significant, that as we look at the geo expansion, the new products, as well as the platform and that is going to over time hopefully start to monetize. Again, good usage, but it needs to monetize more as we look at that. And then of course, you have the AI natives. So underneath the covers, I mean, look at the AI natives are 3% to 5%. It means the core by definition as a percentage continues to call it slow down but the dollars continue to grow. And we look at that over the next three years and we feel very good about growth in both of those and we really haven't baked in a lot of upside for enterprise AI. It's the thing we talk to you, folks, the most about in the question earlier on when is that going to pick up. And at some point, keep in mind AMP is great. Dev did a great job of through that. We also want sales to be able to sell it. That's going to help both self-manage in Atlas, but we do expect growth there. So, the confidence is in the go-to-market engine, the growth we see in our larger customers, the new product and continued growth in AI natives.

If I can just add one other thing that I think is working is also gives me confidence that I have the benefit of stepping back and now stepping back in is that when we launched 8.0, there was some concern both in the investor community is that going to cannibalize some of your business, right, because, if you're providing 30%, 40%, 50% faster machine, this is going to need to use less gas. Well, it turned out that they wanted to drive that car more because they put more workloads on it because they saw so much benefit. And the fact that 80% of our fleet is already at 8.0 faster, two times faster than any other prior release gives us confidence that that same phenomenon will happen with 9.0 because 9.0 is that much more performant and that's a pretty material data point. (04:03:38) whether Atlas Infinite where you feel like is there cannibalization risk concern is completely net (04:03:57) to growth? No. And let me address that. Ever since I've been at MongoDB, MongoDB has been viewed as a great technology. No one has ever questioned the caliber the technology built. Some people have questioned are you – some people used the term Ferrari or turbo sports car. It's great for certain workloads for the mainstream meat and potatoes. Sometimes there's internal workloads that don't need all the sizzle. Maybe MongoDB is not the right profile for. By increasing the price performance, we make it just that much easier for people to put more and more workloads on MongoDB. So, it helps our expansion to larger accounts because there's more workloads we can track, it helps our upsells and cross-sells in our existing customers and obviously, we can go after these new AI workloads as well.

Yeah. So thanks for the question. First, I would say I think what CJ did well was he really elevated the company to be able to sell at the C-level. He has a pretty big rolodex coming from ServiceNow where he knew a lot of senior decision makers. And historically, with infrastructure software, you have to sell both at the end user level because you have to get people buy in because no one's going to try and shove infrastructure software down and user's throats unlike application software. And then you have to still sell high to kind of get the economic buyers bought into whatever the compelling reason is. And I think he, especially in some key accounts where he had deep relationships, he helped us I would say accelerate deals. But what I'll also tell you is that and I've seen this myself, I show up at a C-level, we shake hands and a deal happens, it wasn't because of that meeting. It's all the ground work that had happened before that meeting to set the stage for us to kind of consummate the deal. And so, I don't want to suggest that that was a one man show. It was really the go-to-market team who really helped galvanize the momentum in those accounts. But clearly, CJ added some value on key accounts where he helped push the deals over the line.

I think the main thing is that is the growing… It's funny. I don't think it's specific to AI. I think it's just more around Postgres gaining so much popularity. I'll give you an example, if you are vibe coding on Lovable, which is a very popular vibe coding platform and you need a database, Lovable will default you to using Supabase, which is hosted Postgres. Any vibe coder is automatically putting their data in Postgres if they're using Lovable. I think unless MongoDB starts getting into those kinds of partnerships, that will be a challenge for them to grow.

It was, but our leadership had pushed us to that 15% mark. It's what we're hoping for, but we did not see as much AI work from a MongoDB perspective as what we were hoping for, as well as the other core differentiator that we didn't see as much of a competitive advantage for MongoDB when it came to selecting your vector database layer. And so I think one of the things that we're hoping for is more of that differentiation from MongoDB to say why do we need to use you as our vector DB of choice? And what are you offering in addition from an AI tool perspective that would separate you from the pack? And so unfortunately, although we had a lot of buzz and evaluations of MongoDB, we just didn't have as much pull-through that said, hey, we're definitely going the MongoDB route. That is the thing that we're seeing with MongoDB. There's a very loyal user base to MongoDB because it's been around for, you know, over 15 years, and that loyal MongoDB base is the one that will say yes, we're doing it. But if you don't have that, then it's a very crowded field and then it's a pick of choice of hey, pick a flavor and we're moving forward with that. That's what I'm seeing in the field from a MongoDB perspective.

Can I take that? Yeah, I'm going to have to turn it over to Miguel, but the first thing I'll tell you is, we offer our customers a wide variety of capabilities, products, and functionality. And you mentioned Agentforce is one. Coworker, if you haven't seen Coworker, it's our fastest growing AI product. We have another superfast growing product, Slackbot. We expect Claudeforce to be growing superfast. Agentforce also is growing superfast. We have a lot of them. And what we're trying to do is let customers have very flexible capability so that they can choose all of them and grow with them as they want. And we have these new bundles that you've already heard about. And these new bundles let customers do that. See, customers want to buy and want to price in different ways. This is something I've learned really aggressively recently. Some are still buying by user and by agent, and that's important for them. Some of them want it by consumption, and that's important to them. Some of them are just basic usage customers and they want to pay that way. Some of them even want to pay by outcome, and that outcome could be by a transaction outcome or a business outcome. Because customers want that kind of diversity in pricing, we've created a high level of flexibility. And that, I think, has really expanded our ability to sign very large transactions with our customers.

Yeah. And I think that that's what's really key. I think your question is so good, Gabriela. I'll tell you why it's so good, because I think that if you would ask us a year ago, we'd be like, well, we don't really know. But I think now what we see is, wow, all this new action on the network, call it agentic action, new user interfaces, new applications. It's just a lot more activity on our service. I mean, the amount of action at Salesforce on our service, the value that we're adding, the actual value that we can give to our customers is immense. So have we fully optimized that yet? No way. We're still trapped in some ways in old, per user pricing models. But the opportunity to build much more aggressive pricing to really represent the value that we're offering our customers, I think is enormous. And Miguel is really at the very beginning of what I think will be one of the most exciting journeys he's ever been on.

**The bear case** is that this was a narrative-improving quarter rather than a fundamental step-change. Reported revenue grew 10.8%, but approximately $456M came from Informatica, leaving organic growth around 6.4–7%. F3Q organic growth is still estimated at only 6–8%, while most of the FY27 guide increase came from Contentful and Fin. Even mgmt’s stated $100M organic raise falls to roughly $65M after accounting for the quarterly revenue upside. Sales growth slowed to 9%, Service remained at 5%, Marketing still declined and Integration and Analytics fell 6%. Total RPO growth of 10.7% was below Street, and Data 360, Platform and Other decelerated from 23% to 20%. The central bear argument is therefore that CRM is buying enough growth to preserve an 11–12% headline while the mature core remains in the mid-single digits.

Here's a rapid fire round for you. Are we worried about seat compression? Not at all. Our addressable user base is growing, and 50%, 5-0, of our net new business is already non-seat based. We keep seat-based pricing because customers prefer it for predictability, particularly now, where a lot of pricing out there has been less than predictable. Are customers going to build their own? I've yet to meet a customer who would even consider it. The best tech leaders know. It will cost 5x to 10x to build an agent, versus run one on ServiceNow.

Yeah. Hey, Tal. So we've been very clear in terms of keeping our pricing very customer-friendly and thinking about how they want to adopt our products and where they see value. So we're never trying to kind of do strange things, which is very hard to manage or hard to understand. So, hybrid pricing structure, which we introduced, where it's a combination of license and usage, directly applies to giving the customer predictability and flexibility. And when you use more, that means you are getting good outcomes, otherwise you will not use more. So, outcome is already tied in the underlying assumption of usage, right? So, trying to kind of define another metric, which is very hard to understand or measure, and then have to go back and forth in terms of contracts and everything else, in enterprise software doesn't work and we've seen that for years, and it's played out so well for us now because the way we have addressed our pricing requests from customers seems to resonate with everyone. Right? So, that's really how we're thinking about it. And the question you asked about tokens and things like that, it's really based on the outcome in terms of what you get, the result, the full solution, versus just counting each of the individual pieces of LLMs or things like that. So, customers are seeing the outcome from the usage, but they're also now being very predictable in terms of what they will pay us. And that has worked quite well.

The gap between management’s AI numbers and real usage in our channel feedback persists: checks show 2Q Now Assist real usage flat to slightly worse vs. 1Q, and customers’ AI-driven consumption at only 6-12% of new ACV (management’s broad definition puts it at 40-50%); the AI ACV expansion is driven more by pricing bundling and SKU restructuring, and metrics like the 9x growth in agentic AI customers come off very low bases.

Yeah, I'll start. Thank you for the question, Kirk. So first of all, you're right in that it's early days. The difference on these agents versus traditional applications, they iterate and become better so much faster. So, I'm very optimistic that we're going to see ramping up usage of these agents faster than we ever saw of our apps. And, we're also new to the flex credit consumption model. And, that's a delayed gratitude, delayed gratification model that is frankly newer to Workday. Again, very optimistic. We're seeing great early signs. But I think that's one of the reasons for cautious optimism going into next year, because we are aggressively moving towards a hybrid model between subscription and this consumption model. I don't actually see it as crowding out. I actually see it as customers, new customers choosing Workday because of our agentic strategy, existing customers, they have an AI budget, and now we have products that actually fit in that AI budget, which is a big win for us. But again, a lot of these new AI products are consumption based, so we won't see the impact from the revenue until months or a year down the road.

With that in mind and ahead of our upcoming Financial Analyst Day, I'd like to provide some early context for how we're thinking about next year. Our current target for FY 2028 subscription revenue growth is consistent with our expected second half FY 2027 growth rate of approximately 11%. We also see potential upside across new products, including Sana Enterprise, Workday Extend with Data Cloud, and our AI agents, all of which are seeing great early demand. In addition, we expect our non-GAAP operating margin to expand by at least 2 percentage points next year.

UBS says WDAY reaffirmed FY28 targets of ~11% revenue growth and ~200bps OM expansion but offered no FY29+ outlook, prudent given AI transition and execution history. AI solutions are expected to drive ~70% of incremental ARR in FY28 (>50% FY27), yet revenue still decelerates from ~14% FY26 to ~11% FY27 as core HCM/FINS and app spending remain soft; UBS also flags limited FINS commentary. Neutral, PT $210 from $220.

Let me be direct about what happened. April got off to a slow start, and the quarter we expected did not fully materialize. Two factors drove the headwinds. The first was deliberate. As we discussed last quarter, customers adopting AI want proof of value before they commit, and predictability in what it costs. So in April, we leaned into those trends and made changes across product, pricing and go-to-market. On the product side, we introduced trials so customers can turn on agents and AIO in their own environment with their data for their workflows. The goal was simple, let customers experience real outcomes before they buy. We knew this would create some near-term headwinds by extending the buying process, but we believe it's the right long-term trade-off. Lowering the barrier to adoption and building customer confidence and outcomes will ultimately accelerate AI adoption. So far, this approach has worked best with upmarket customers, where our teams and partners provide the support needed to make the trials successful. Our focus is to scale that experience so every customer can adopt AI with the same confidence. On pricing, predictability has become a defining theme in AI adoption. Businesses have been hit with unpredictable token costs, and they want pricing that is transparent and tied to value. In response, we introduced outcome based pricing for several of our HubSpot agents, lowered entry price points and are providing customers clear visibility and control over usage and spend, including the ability to set thresholds that fit their budget. The second factor was a shift in the demand environment in Q2 with increased budget sensitivity. Businesses want greater confidence that their investments will position them for the AI platform shift. As a result, purchase decisions are facing greater scrutiny, buying committees are larger and more deals require C-suite and board approval, leading to longer sales cycles. For existing customers, unpredictable AI costs across the broader landscape are impacting budgets, leading to budget optimization and downgrade pressure.

Our Q3 and full year guidance reflects our expectation that the headwinds we saw in Q2 will persist throughout the remainder of the year. We now expect quarterly net additions to be approximately 5,000 to 6000, and ASRPC growth to be in the low to mid single-digits in constant currency. For the full year of 2026, we expect net revenue retention to be roughly flat year-over-year. We expect customer dollar retention will remain strong and stable in the high-80s, and we continue to expect net upgrade rates to be pressured in the second half as a result of customer budget optimization. We now expect net new ARR growth to be below constant currency revenue growth for the fiscal year of 2026.

A disappointing print against elevated buy-side expectations. Revenue technically beat the Street, but growth came in slightly below the whisper bar, billings missed, 3Q guidance was cut below expectations, and the full-year constant-currency outlook was reduced. Positioning was mixed into earnings, but the combination of softer demand signals and lower estimates is likely enough to push more investors off the name and could weigh on app sw tomorrow after recent rally.

Adobe's AI innovation has driven an impressive 3x year-over-year increase in AI-first ARR to greater than $500 million. We believe now is the time to aggressively acquire the next generation of Adobe loyalists. The strategic shift to acquire more freemium customers through Acrobat and Firefly lowers our second half ARR growth expectations from individual subscribers. We believe these changes do make Adobe even stronger.

* **AI.** AI-first ARR exceeded $650M, more than +150% y/y and approximately +30% sequentially against the prior disclosed threshold. That is real monetization, but still only approximately 2.4% of total ARR, and the definition includes Semrush AI products. Firefly app and credit-pack ARR grew 40% q/q, with video usage and broader credit consumption supporting expansion. Importantly, that metric excludes Firefly Enterprise and cannot be multiplied against the prior, broader $300M Firefly ARR disclosure. Acrobat AI Assistant MAU doubled q/q, but usage growth is not equivalent to growth in paid subscriptions.

**The broader bear thesis** is not that creative activity disappears. It is that AI makes useful creative output easier to produce while reducing the premium users will pay for specialized tools, established workflows and professional seats. Growing content volume therefore need not translate into growing software spend at historical pricing. Competition is also moving beyond isolated generation tools into more complete workflows. Affinity, for example, offers professional design, photo editing and page layout free for individuals, supports common creative file formats and integrates with Canva, with optional AI features tied to premium plans. That does not establish equivalent capability or prove customer losses, but it raises the burden on ADBE to demonstrate why users should convert to paid products rather than remain free or use alternatives.

Thank you, Roy. The clearest evidence that our strategy is working is what we're seeing in AI adoption. AI ARR doubled from Q1 to Q2 and now represents 17% of net new ARR added in the quarter. This is a meaningful signal, not because the absolute number is large yet, but because the rate of change tells us customers are actively choosing our AI capabilities, not simply inheriting them. Equally encouraging is the early reception to our new seat and credit pricing model launched in May. The adoption has been strong out of the gate and customers are engaging deeply enough with our AI product to invest beyond the default package. Pricing models only work when customers see value worth paying for. We are beginning to see that.

Yeah. Maybe just to add what Roy said. Before the changes in the product and the pricing, the only way customers could expand was to add more people and more seats. This is the first time that we added the new agent that we see customers expand not only on the seats for humans, but on AI consumption. And for us, this represent a new vector of growth that didn't exist before, which is part of why we believe so much in our strategy and why we believe so much this is a fundamental change in how the company will be able to grow going forward.

Yeah. Hi, Howard. So, as I mentioned earlier with regards to the philosophy, it hasn't changed. Few things we took into account I said earlier that we're observing the near-term cost of the 20% workforce and this is something that we had to take into account because there is going to be a short-term impact. And we did not want to layer an aggressive top line rate on top of the execution risk that hasn't fully played out yet. We said that at the beginning of the year also with regards to the top of funnel, that we took it into account. And we prefer to under promise and over deliver through a transition. This is a big transition for us than raise revenue guidance now and with some uncertainties we still have throughout the year.