2026.10.09 | GP | Internet | DoorDash: Muse, AI & B2B Growth

Primary:

DASH

Associated:

UBER

DHER:GR

CART

Viewpoint:

Former Executive

Moderator: Raghav Nayar (RN)

Nipun Kant (NK), Former Director and GM of B2B SaaS - DoorDash

SUMMARY

Overview

The discussion examined DoorDash’s approach to AI agents, third-party connectors, text-based ordering, corporate catering, DashOS, and its Costco partnership. The expert argued that DoorDash is moving cautiously because external agents could expand demand while weakening its direct customer relationship and advertising business.

DoorDash has invested heavily in AI and appears technically capable of supporting agent-driven ordering. Its main strategic challenge is deciding who controls checkout, payments, customer data, advertising inventory, and the overall purchasing experience.

Key Insights

AI connectors create an innovator’s dilemma. They could materially broaden DoorDash’s addressable market but may also allow external platforms to become the primary customer interface. The expert views DoorDash’s delivery network and service quality as highly defensible. Advertising is considered more vulnerable because AI platforms may sell restaurant placement directly. Text ordering should increase order frequency by reducing friction and adding SMS as a high-engagement marketing channel. However, food discovery remains visual, supporting continued use of the DoorDash app. DashPass could retain or increase its relevance as connectors attract less frequent users who may generate favorable subscription economics. DoorDash appears well prepared for demand spikes due to technical stress testing and its ability to adjust courier supply and incentives. DashOS may be too complex for mainstream restaurants. Its strongest fit is sophisticated, higher-end full-service restaurants with advanced loyalty and guest-management needs. The expert was strongly positive on Costco, citing brand strength, large baskets, broader selection, and potential customer acquisition benefits.

Competitive Landscape

10/11/2026 at 11:33 AM UTC

Positioning

Advantages

Risks or weaknesses

Leading delivery platform expanding into AI access and restaurant software

Strong delivery network, consumer demand, selection, and logistics expertise

Connector disintermediation, advertising pressure, slow catering execution

More aggressive connector adopter

Enterprise integrations and early AI partnerships

Weaker consumer loyalty and higher reported Costco markup

Company or product

DoorDash

Instacart

Uber Eats

Delivery and catering competitor

ezCater

Established corporate catering platform

Potential catering expansion through an acquisition

Discovery surface and specialized restaurant tooling

Weaker delivery position according to the expert

Clunky software, restaurant dissatisfaction, and limited scale

Meta Muse and other AI agents

Potential commerce gateways

Massive reach and end-to- end transaction control

Could pressure delivery platforms on data, payments, and advertising

Toast and restaurant software providers

Quantitative Insights

Merchant technology competitors

Established restaurant workflows

DashOS could increase DoorDash’s competitive leverage

Metric

Detail

AI investment

Tens of millions of dollars

Strategy development

Approximately 2 years

DoorDash monthly active users

Roughly 36 million

Potential agent reach

Illustrative scenarios of 2 billion to 3 billion users

SMS delivery rate

Estimated at 99% to nearly 100%

Costco markup

DoorDash around 10%; Instacart reportedly around 30%

Example promotion

20% discount

ezCater fee

Approximately 75 dollars per month

10/11/2026 at 11:33 AM UTC

RN: Good morning. My name is Raghav Nayar with the Consumer Insights Group at Guidepoint Global. This morning we are discussing DoorDash's connector strategy. We plan to address the potential Meta/Muse relationship, what restaurant delivery could look like with third-party agents, the programs in beta testing, and DashOS. As a reminder, if you have any questions for the advisor during the call, please email ask@guidepoint.com and we will try to get your question answered. I am very glad to welcome back Nipun Kant, former director and GM of B2B SaaS at DoorDash. Nipun, thank you for joining us this morning. I just want to make sure that you will not speak about any material, non- public information, security prices, your investment holdings, or provide any investment recommendations during the teleconference.

NK: Sounds good and good to talk with you again.

RN: Nipun, please tell us about your background, your work at DoorDash, and when you were with the company.

NK: I spent the past 10 to 12 years mostly in startups as a founder and after that was a GM at DoorDash, primarily responsible for its software offering. So everything software as a service sold into restaurants to help them grow their business.

RN: A question. On the August earnings call, Tony said that DoorDash is testing with agentic partners. This is kind of different from what we've seen from Instacart, who is already integrated with Gemini in May. Could you just talk about DoorDash's connector strategy and what is it?

NK: I can't share confidential information, but what I can say is that the team has spent the better part of 2 years thinking about this. So this is something that they've been extremely intentional about with at the most senior levels of the company. What you might be getting at is the fact that, frankly, they've been a little bit slow to move. There are really 2 big buckets of things here. This is not necessarily not obvious, but 1, obviously DoorDash makes the majority of its operating income on advertising. That's actually increased over time as their ad business has grown dramatically. 2, to actually have enduring market value, both from an advertising standpoint as well as a consumer stickiness standpoint, they need to own that customer connectivity. Connectors can be good to generate demand but also risk disintermediation.

RN: Great overview. And it seems like Instacart has probably been the most forward. Why are they second, third, or last to go to market with connectors? I guess what is holding Tony or the company back from being more forward or being more strategic? And also, how important is it to be first or second?

NK: They are being extremely thoughtful about this. So they are absolutely strategic; you can debate the strategy, but they're definitely trying to have a strategy here. Instacart, I can't speak to their exact strategy, but they're in a completely different market position. Their position is being threatened more and more. I don't exactly know what their advertising revenue looks like vis-a-vis DoorDash, but they certainly have much less consumer stickiness than DoorDash does. You can imagine that might be explaining why they have a slightly different strategy there. To give you a flavor, and to peel back the onion 1 more layer, these are discussions that I'm sure DoorDash has had for multiple years with all the big players. Those are pretty complex discussions. An example, and again, I'm not going to provide too much detail here, but an example here is: what does that integration with ChatGPT look like? Can they complete the entire order within ChatGPT or is the consumer kicked to a DoorDash checkout screen? That's an example. Who owns the payments work stream or monetization stream? What consumer data is shared between the ChatGPT platform and DoorDash? All those things are highly relevant and I think very, very strategic discussions.

RN: Excellent. And then does DoorDash have the AI infrastructure to serve connectors? I guess the question here is really, is there anything structurally holding DoorDash back from moving forward with connectors more quickly?

10/11/2026 at 11:33 AM UTC

NK: They have thrown tens of millions of dollars in an AI labs team, so these people know what they're doing. You can debate whether that's a good use of funds or not. They already have a chat functionality within their app. That's the sort of core AI primitive, if you will, which essentially is wrapping their search capabilities as well as ordering capabilities inside an MCP. That's not really a driver of any delays at this point.

RN: I'm sort of wondering, Instacart has deep enterprise integrations, and I know you did a lot of work on the software side. I'm just wondering if DoorDash has been a little bit slow to scale or a little bit slower to communicate its connector strategy because it wants to push forward on its enterprise solutions with restaurants.

NK: Its enterprise solutions with restaurants. What do you mean by that?

RN: Or not enterprise, or it's a suite of restaurant solutions like, beyond the marketplace and delivery.

NK: I don't think so. I really am pretty confident that this is something that the team has invested a ton of time on. There's a discrete team and work stream, very actively thinking about this. They've obviously done some kind of beta work with ChatGPT, I think maybe even a year ago on the grocery side. It's just classic innovator's dilemma. You're damned if you do; you have to figure out how to plug the hole from an ad spend perspective and you're damned if you don't because these are platforms that are going to start to become key sources of how consumers discover new products and services.

RN: So, just going to the Meta news or Meta's announcement with different partners, what does an official link with Muse look like, Muse and DoorDash?

NK: At the most basic level, you can imagine what DoorDash does is it exposes a set of tools via an MCP that lets the Muse agent essentially do anything that a consumer can do inside the DoorDash app. What can a consumer do inside the DoorDash app? They can search for restaurants, they can see prices, they can see menu items, they can build a cart, they can view their past orders, they can see their rewards. Technically, DoorDash could expose any of those capabilities that the user can do inside the app to an agent, and that agent, that's a sort of an MCP tool that you expose to Muse. DoorDash, of course, won't expose every single thing, but strategically, the tension is going to be in a few things. One is, do you let a consumer complete checkout on Muse directly? And if you read Muse's connector guidelines, which they just published this week, they are pretty clear that they want consumers to be able to complete the end-to-end flow within Muse versus being kicked to a third-party checkout screen. That's one piece of it. The second piece of it is data sharing. What data is shared when that user completes a purchase via DoorDash? Does DoorDash get to match their order to an actual profile, or is their account hashed in some nature? I think the biggest interesting thing, and interesting to see how Muse develops, and I think you're getting to this in some of your other questions, is from an advertising perspective, you can imagine that's also actually exposed to the tool. If consumers on the Muse app start asking, "I want a burger," Muse can use the DoorDash MCP to search for DoorDash restaurants that have burgers and also showcase an ad or sponsored listing if, let's say, Burger King is paying DoorDash for that ad. Basically, Muse can then become another conduit for that ad placement. Those are, I think... obviously this is very much new ground for all of us... some of the contours of the discussion.

RN: And as far as the path to purchase goes, what happens then to the path to purchase with our agents or restaurant orders? So you'll have the user, NK: Really, there's a possibility that the user could circumvent the app and never open it.

RN: Oh, that's exactly right. Now, is DoorDash actually concerned about that?

NK: Maybe at some point. It really depends on, if you start from the premise that Google search has 4 billion monthly active or something like that, 5 billion monthly active, let's just assume that Muse gets 2 billion monthly active. I don't

10/11/2026 at 11:33 AM UTC

think that's crazy at all. Then there's a huge risk. Just like Google became the front door of the internet, there's a risk that Muse becomes the front door of all products and services. And it can offer free to consumers and make money on payments. So I think there's a big risk of disintermediation. I think that Muse can make money on payments pretty easily there. So I think there's a huge risk there. I think Tony knows that. Now, obviously, there's also some pros, which is it could massively unlock TAM. You guys can look at the publicly reported data on DoorDash monthly active. Is this going to increase demand for DoorDash? Absolutely. Is it going to threaten DoorDash's enduring value? To some extent. I think the one last thing I'll say is, what creates enduring value for DoorDash? Obviously, the delivery network is extremely hard to replicate. Extremely hard. Amazon tried to do food delivery; they basically gave up. Uber Eats is not doing that well. So that is extremely defensible. But the big question is going to be the durability of the consumer relationship. I'm not necessarily as concerned about that because DoorDash has the delivery network and the best quality delivery. So Muse is probably always going to pick, likely pick DoorDash more often than its rivals. I think there's a big threat on advertising revenue, as you've discussed.

RN: Got it. And just why pick DoorDash over Uber Eats if Muse?

NK: That's actually a good question. In some ways, you could ask Muse, "Compare prices on Uber Eats and DoorDash and find me the best price." No one actually does that. You can do that today if you Google any restaurant and you can compare prices. No one does that. So I'm not too concerned about that, but that is an interesting question. So, okay, if a user doesn't specify that they want to use DoorDash, and let's say maybe Muse has a connector into both Uber Eats and DoorDash. Again, it just gives you platform power. This is classic platform power. It lets you pit DoorDash against Uber Eats. Again, that then I think really is, okay, how does DoorDash have its own surface, which I think you're getting at with text-to-order? I think DoorDash needs to really nail, and look, it has nailed so far, the consumer relationship. There's arguments against this world from happening, by the way, like discovering food is a very visual journey. You want to scroll, you want to see your menu items, you want to be able to do X, Y, Z. So you can imagine a dedicated app for that is valuable. There's going to be a lot of gray area. Does DoorDash expose the actual UI within Muse to let users scroll? It certainly could, technically, through an iframe or other. There's benefits there on the ad side. I'm not sure Muse will like that. So those are some of the contours of this sort of new world, and I'm happy to go into detail.

RN: Go. And I'll just stick with this because you made a very interesting point about if DoorDash opens with Muse, they could see exponential customer growth. And I'm wondering, is DoorDash ready for something like that to happen from an infrastructure perspective?

NK: I think so. DoorDash, what, 36 million monthly active? That's not that much if you think about it. Think about how many consumers there are in the US alone. They have done... how do you match? Basically, the key infrastructure is 2 things. 1 is technically, and 2 is data network. And they have actually been in meetings. They've done a ton of stress testing on the technical side because there's such crazy spikes, Super Bowl, Mother's Day, those are crazy spikes in demand for DoorDash. So they've definitely done a lot of testing there and they've invested heavily. The Dasher network, they're pretty used to scaling up and down Dasher acquisition and incentives to basically clear the market. That's literally their core capability, to match demand of food with Dashers to fulfill that demand.

RN: Got it. Great. We'll go to Apple's Messages app and text-to-order. I'm wondering if you've tested the messages text-to-order capability and what do you think about it?

NK: I have not tested it, but I think really the key debate here is, is this a defensive play or an offensive play? The offensive side is, does this lower friction to order and just unlock more ordering value, with of course the threat of ad dollars, although you can easily do ads within a text-to-order interface. The agent can message you about a sponsored listing. It's not a great consumer experience, but I'm sure they've thought about that. The defensive side is exactly their fear of being disintermediated. If I'm Tony, and I think this makes sense to try this, we want to basically...

10/11/2026 at 11:33 AM UTC

we know this is going to be a thing, the ChatGPT, LLMs of the world. We need to have enduring consumer value in every single channel that people operate in. So let's test connectors, let's test our own essentially agent that's exposed by a CLI or SMS and let's obviously keep doubling down on our own app. That's what I would do if I was Tony, so I think it makes sense. This surface, using your word, is open, or this capability is only available to DoorDash users, which I think is interesting. And I'm wondering if this would, having restaurant ordering available in iMessages or whatever, Apple Messages, would that increase order frequency?

RN: Yeah, I mean, sorry. What? Sorry, go ahead.

NK: What? Sorry, go ahead.

RN: Go ahead.

NK: Or I guess the question is, what would happen to order frequency? What happens to total order volume? There's no question. You can debate the incrementality, but there's no question that this is going to probably increase order frequency. There's a few mechanisms here. DoorDash can pretty easily get people to consent to SMS. How? They can just offer a $5 promo if you let us start texting you. Today I think it's actually user-initiated text, but that's actually very interesting, which is: can DoorDash encourage consumer opt-in to DoorDash text? What that can do is it gives DoorDash another channel to nudge consumers to place orders. Obviously, you can do that through the DoorDash app, but not everyone has push notifications enabled. The beauty of SMS is that 99%, it's going to be 100% delivered. We're pretty immune from push notifications at this point. We typically just dismiss them versus an SMS, you typically read every SMS you get, at least today. So there's a few mechanisms to grow order frequency and order volume. 1 is user-initiated, which is, in theory, you're going to lower the friction to place an order. You're on your couch. Now, you can debate that. How much is that? Because, again, food is visual. I don't think even DoorDash knows, but they at least need to try and experiment. The second piece is even if food is visual, you can start with a text and then the text can also give you a link to review your options, and that link automatically drops you in the DoorDash app. So it just becomes another conduit to go to the app. Then 3 is this agent-initiated, which is essentially a glorified push notification but in SMS form, which has the potential to grow frequency for sure. Does that make sense? I'm happy to provide more detail.

RN: I guess if you could go a little bit more into the agent-initiated part.

NK: I actually don't know how it works today with DoorDash text-to-order, but I think this is interesting where, historically, these text message apps, V1 was like, "Here's your 2FA," 2-factor authentication code. That's just transactional. Then you have a few apps like United has this, where you can start texting them, "Hey, I have a problem with my flight." That's essentially what DoorDash is now doing with text-to-order. I think the user can text the DoorDash agent, "I want a burger delivered XYZ." But I guess the third piece is, I think you basically get consumer consent for agent-initiated text messages. That can be one of the key levers to drive DoorDash consumer demand, tapping them on the shoulder. "Hey, here are the new restaurants in your neighborhood, we noticed you haven't ordered in a while, come check out XYZ." "Oh, your favorite restaurant, Bob's Burgers is doing a special." Those are classic consumer engagement tactics that DoorDash does via email and push notification today. SMS is another conduit there. I think all these apps will try to do agent-initiated SMS with the caveat being you can't abuse it, because if you do that, then you're just going to have consumers opt out.

RN: Got it. We'll go to Dash, and I'm wondering if you have, you know, you had talked about the text-to-order, and platforms we'll go into a little bit later, but what happens to DashPass membership program? Does it become more or less relevant?

NK: I think it's interesting. I think it becomes, at the very least, the same relevancy and arguably more. The reason for

10/11/2026 at 11:33 AM UTC

that is, DashPass subs, I don't know what they are these days, but what Tony wants to do is how do you make that like the Amazon Prime, which has whatever, 100 million subscriptions or whatever, or like a cost of membership. For that to happen, you need new users. That is where I think this becomes very interesting, where you're probably going to trade off ad spend in exchange for DashPass subs. The other thing you can do is, maybe Tony gives Muse or ChatGPT access to the economics of DashPass, and in turn then ChatGPT can incentivize DashPass. "Hey, have you noticed you've done 3 orders? Did you know you can save X bucks if you pay for DashPass? Do you want me to set it up for you?" Finally, the value prop of DashPass to the user is obviously savings on a per-order basis. I think that continues to make that value prop just as valuable. It makes it more relevant to the user, but it arguably makes it more profitable to DoorDash, because where does DoorDash make its money on DashPasses? Probably on people who pay for DashPass, but don't order a ton. If you unlock a lot more new users who are probably not necessarily your power users, that probably makes each incremental DashPass sub more profitable.

RN: Great. Well, you know, now we're sort of looking further out into the future, just agentic adoption, and here I'm sticking to restaurant platform for discovery. What do you see happening between GPT, Gemini, Cloud, or Muse as far as restaurant delivery? I know that not much has been said so far.

NK: You mean as in which one of them will win?

RN: Yes.

NK: It's hard to say, obviously. The framework I would have is which platform will most enable users to get stuff done. Because that's the platform that people will think about when they're like, "I want a burger" or whatever. It becomes your default. This is what Muse is trying to do. It becomes your default way of interacting with the world and solving problems. You need the most holistic agent that can do stuff for you. I would argue for the mainstream consumer, that's probably Muse today, but that may look very different in a year. Obviously everyone's going to try to play in this game. The other interesting thing I'll say about Muse is the hardware side becomes interesting because imagine this world where 3 billion people are talking to their personal agent all day, every day. How are they talking to their agents? You can imagine Meta's hardware ecosystem actually becomes the delivery channel for Muse. It makes Meta's hardware strategy make a lot more sense, because no one's going to want to walk around all day with their AirPods in their ear, 12 hours a day.

RN: Yep. And then do you think that platform demand is really incremental or how do you see platform demand evolving over time? DoorDash has said that it's a very, very small or very insignificant in grocery and Instacart has said that it's also very small. I think DoorDash's use case has really been on the grocery side so far.

NK: Yeah, they haven't really gotten into restaurants. Users don't think about ChatGPT when they want to go get a burger. The tension is going to be: how much do you believe that users need to go to the DoorDash app to figure out what they want? Obviously, that's a huge use case. That's why DoorDash optimizes for selection, because they want every restaurant available so that it encourages every user to go to the app. That's going to be the counterargument to any of these things, which is: you still need a visual surface for discovery in food in a way that you don't in other surfaces. Now, will it still be incremental? Absolutely. The reason why is lack of friction. You're going to get a bunch more people to use food ordering, even if they don't have a DoorDash app installed, they're not Muse, it's not a habitual thing. It's probably good for unlocking new demands of new users of DoorDash. The interesting strategy for DoorDash is going to be: how do you convert those users to use the DoorDash app over time? That's what I would try to do if I was Tony.

RN: And I know this is a point of contention. You had brought this up at the start of the conversation, ads. When the customer begins their journey, if they begin their customer journey either on Apple Messages or GPT or Gemini in the future for restaurant delivery, how could DoorDash monetize that interaction?

10/11/2026 at 11:33 AM UTC

NK: On Apple Messages, it controls the entire surface. There are 2 mechanisms there. You could provide a text message ad. It's probably a pretty bad consumer experience, but you could. The other thing you can do on Apple Messages, or really any of these apps, especially Apple Messages, is you can route the user with a magic link to go open the DoorDash app. So it essentially becomes a conduit to go use the DoorDash app. That's pretty straightforward. Then you're just monetizing ads with the app. Now, GPT or Gemini is not going to want to do that. Why would they want to let the user go to a third party? Why would they let their agent recommend that? That's going to be the tension. If you agree that's probably not going to happen where they're not going to really want to kick users to the third-party surface, they're going to want to control end-to-end, and they're going to have the leverage to do so because they have billions of users using GPT and Gemini. Then there are really 2 ways to monetize that. DoorDash has to do is... you have some consumer real estate within the app surface of Gemini and GPT. The question is who monetizes that. DoorDash has these advertisers today, and they could say, "Hey, GPT, let us show this ad when we return the response," and we'll give you some economics on it. And then also what GPT is going to do is they're going to go to the restaurant directly and say, "Hey, Burger King, you should pay us directly and we'll give you some placement." RN: We'll go to corporate ordering. Let's see. The question is, how incremental is corporate ordering to restaurant GOV or DoorDash? Whatever you could share with us about corporate ordering, and then what are your thoughts on this corporate press release update, or it was just an update on MCP connector with workplaces?

NK: I think this is interesting. Catering is sort of the Holy Grail, obviously for restaurants, because it's very valuable. Every restaurant wants catering, but then of course, also for DoorDash. DoorDash has probably been slow out of the gate. They've had a big team working on catering, marketplace, third-party catering for a long time. I don't know why they've been a little bit slower to move there. The MCP makes sense. I don't know how much demand they're going to get. They're clearly in this mode of like, they have to experiment with these new form factors and see what happens. There's probably less... what percent of corporate ordering goes to DoorDash today? It's probably pretty low. In that way, if you can make this lower friction, then it's an interesting play. I don't know how well it's going to do, but it certainly is interesting.

RN: And I'm just wondering, I don't know if you saw this news item, but Uber plans to acquire ezCater, which has number 1 share in food tech work with its catering. And I'm just wondering if that's... Can DoorDash actually compete in this space, given where...

NK: For sure. ezCater has 2 things. Number 1 is consumer discovery. It has its own, essentially, consumer surface. Number 2, it's got a ton of tooling to help restaurants manage catering orders. That's non-trivial because you have to integrate with all these POS systems. You need to let users edit orders and how does that flow to the POS? You have to figure out payments, like collect some payments now, some payments later. Who owns the payment stream? Is it the POS company or is it ezCater? There's a lot there. That being said, most restaurants hate ezCater. I'm not surprised why Uber wants to buy this because it needs more consumer demand and it needs more selection. ezCater kind of gives them both of those 2 things. DoorDash has plenty of consumer demand. It probably, I think it's definitely further than ezCater from the standpoint of core restaurant tooling. There's a big work stream to enrich the POS integration so that catering works better on DoorDash than it used to. ezCater spent many years doing that. DoorDash can get there. I'm not concerned about ezCater. It's a small company. Most restaurants hate it, and DoorDash has the consumer demand. It's just how fast can DoorDash move to actually get adoption within restaurants and it just needs to build up all the tooling there.

RN: Why do you say that restaurants hate ezCater?

NK: I think it's just all the same problems of software 2.0, which is it's clunky, it's hard to use. I think they charge $75 a month. They're trying to make money on staff too. Those are some decent reasons, but that being said, it kind of just works and it's clunky, sure, but you need a lot of this fine-grained functionality for the restaurant owner to manage

10/11/2026 at 11:33 AM UTC

their orders and access control where the store associate can see X, but they can't see feature Y. There's a lot of tooling like that that DoorDash hasn't really had to build for its core marketplace.

RN: Got it.

NK: I don't know, I frankly also just think that the execution for DoorDash has been mixed there, if I'm being very honest with you on the catering side. I think they've probably been slow out of the gate.

RN: Good. We'll go to CLI beta. So this is not new news, but back in July, there was a press release that dd-cli was open to beta test to developers in the US and Canada. Why do you think Dash launched this back in July? And I'm not sure who would be... who are they really targeting here?

NK: I think, again, it kind of goes back to this fundamental thing of our entire worldview is that humans are the users of software. What Tony knows is that what's going to happen over time is that agents will be the majority of software users. Again, if you start from the premise that Muse will have 3 billion users and that everyone in the world will have their own personal and business agent that can do anything for them, then it kind of makes sense in that framework. They have to base... all these software companies are basically opening up CLIs and MCP. Even Salesforce is probably the prototypical example. I think that the value of a user-facing interface is being revalued in real time. Again, there are arguments against that where food is a visual thing in a way that just a chat conversation maybe doesn't fully provide that value. That's a bit long-winded, but hopefully that makes sense. There's a few interesting clients here, and again, this is a bit speculative, but there are 2 interesting data points I can share. I don't think this is confidential information, but for example, I know... well, I don't want to share details, but you can imagine, just hypothetically, there's a startup trying to build a business insurance brokerage, let's say. They want to... and a big part of their value prop is they want to be able to give gifts to these corporates to be able to get them to sign up for insurance. This is on the backend side, but they can basically use the DoorDash CLI to automatically send a nice bottle of wine, for example, to a client based on certain triggers. It just becomes another marketing moment. Just like you can dynamically send an email, you can now dynamically send them a bottle of wine. Another example would be, imagine there's an agent for a mortgage loan officer. This agent could do anything mortgage loan officers can do. They can correspond to their clients and research what the rates are, but they can also help the mortgage loan officer network with realtors. The way to network with realtors is maybe the agent finds 10 realtors and then offers to send a nice bottle of wine to them with a nice card. These are hypothetical examples, but those examples are interesting on the non-restaurant side, like gifts is a classic example. I don't exactly know the use case for food other than it's kind of a meme that you can order food through your CLI.

RN: That's interesting. I actually would make a lot of sense during the holiday times when you have corporate sending gifts to each other. Okay, so just on Dash OS, how does Dash OS build on the existing merchant services that are already in place?

NK: What I think DoorDash has been fairly public about, DoorDash has now built a pretty comprehensive software ecosystem to help restaurants grow. Loyalty is a big part of that. For any restaurant of scale, they have to be thinking about loyalty to some extent. With SevenRooms, they're really trying to go into FSRs with table management, guest management, and so forth. This is basically in part, and I can comment on what Tony's strategy is, it's in part offensive, which is they want to dominate the software stack. They're going to make money on the reservations marketplace, et cetera. It's also defensive because it sort of commoditizes, it lets them commoditize their complement, which is the software layer. It probably gives them more leverage against Toast and others because now they have a more comprehensive software ecosystem. There's some interesting BD dynamics there as well.

RN: And then who would be the ideal merchant profile for this?

10/11/2026 at 11:33 AM UTC

NK: In the way they pitched it a few weeks ago, they seem to be really going after these FSRs who have loyalty. It's like these higher-end full-service sit-down restaurants where they really care about the guest experience, like a Union Square Cafe. These are restaurants that are pretty sophisticated. They want to be able to know, "Okay, Bob likes this item, but Sally likes this other item." They want to use that knowledge in marketing, in training their staff to welcome Bob with a special. To be very honest, the way they pitched it 2 weeks ago is, I think it's a pretty small sliver of the merchant. I think that's partly a product of the fact that they acquired SevenRooms. SevenRooms played an ultra luxury, the highest-end FSR. If you think about your mainstream, let's say Thai restaurant going down the street, it's obviously a sit-down restaurant, but are they thinking about these things? No, absolutely not. They're not thinking about how to do advanced marketing and this and that. That's the question of DashOS: is it an overfit for most merchant needs?

RN: Interesting. It looks like it's more appropriate for restaurants that have influencers. Okay, we'll go to DashOS and Restaurant GOV. So you had talked about this being at really the top end of the restaurant chain. I'm just wondering, what are they trying to do here? Is this to drive growth frequency? How material is this to Restaurant GOV?

NK: I think it's really to try to drive software demand. The big value prop of DashOS, which I don't think they were that explicit about... I think they were very confusing in their messaging... basically the thing is for the restaurant, the restaurant gets the contact info, which they've never done before. The user can sign up for loyalty on the DoorDash app, and then when they do that, the restaurant gets contact information. So they're exposing contact information for the restaurant for the first time. In my experience, and I led the early development of this, I don't want to share too much, but I think restaurants found this very confusing. It's just a lot for them. Even the press release 2 weeks ago, it's a very convoluted user story, and I'll be very curious to see how it does.

RN: Nipun, thank you. I wanted to ask you 1 more thing. What do you think about DoorDash and Costco, the partnership? And then just this morning I got 20% off on, I'm sure you probably got it too in your promo box.

NK: I think that's going to be fantastic. I mean, finally. Actually, their markup was only 10% on Costco. If you played around with, if you go to the Costco app, as of 2 months ago, I think Instacart was charging a 30% markup. DoorDash has a big focus on reducing what they call inflation, quote unquote inflation, which is they want to reflect in-store prices. Obviously, that's a pretty key part of consumers wanting to go to the DoorDash app. This is just about selection value. They want DoorDash to be the place that consumers think about when they want to get anything. That's going to be the defense, they're going to hope that's a defensibility against Muse because you have this discovery dynamic and it's going to be a lot faster to go to an app to discover your food options versus use Muse. That's what Tony's betting on and Costco is going to be a huge part of that.

RN: Got it. And I guess, okay, so you're very bullish on the DoorDash Costco partnership, but what do you think this does to order growth or basket value, that kind of thing?

NK: Basket values are huge. One value is you're going to see some benefit in new consumer adoption. You're going to see consumers log into DoorDash more often in general because of Costco, because Costco is such a big brand. For Costco's orders by themselves, you're going to see high AOVs that are very, very interesting from an AOV perspective. There's some nuance of how do you price service fees? How much do you pay Dashers? Because it takes a while to shop. There's a lot of nuance to get into on the unit economics. If DoorDash knows anything, it's how to make money on delivery. They're going to look at that stuff from the 1-cent level. I know there's going to be team focus on that.

RN: Excellent. Nipun, thank you so much for joining us this morning. Great call. Always good to have you and we'll connect again on DoorDash in the near future.

10/11/2026 at 11:33 AM UTC

NK: Perfect. Thank you very much. Take care.

RN: Okay. Bye-bye.

DISCLAIMER: The statements or opinions expressed today are those of the Advisor and not Guidepoint, who disclaims all liability for the content provided. The Advisor may not disclose material nonpublic or confidential information or any information that would cause the Advisor to breach any duty or obligations. Guidepoint is not a registered investment advisor and the information provided is not intended to constitute investment advice.

10/11/2026 at 11:33 AM UTC