Stripe's AI Chief: How AI Agents Will Buy, Sell, and Pay

29:02It's really about, okay, how can agents transact without taking on risk and how can businesses remain in control as the merchant of record. And so I would think of like one time use virtual cards were like a useful backstop as we went and built the shared payment token. Shared payment token is a new payments primitive that is wallet agnostic. And by the way, also payment processor agnostic. You can pass it over to Adyen or whoever else you have. It doesn't have to be processed on Stripe. And then think of like LinkWallet as the consumer experience, the consumer wallet for agents that leverages that same shared payments token primitive, but is more fully featured for the consumer.

29:44And it's software, so it's fully programmable. programmable. So just to double click on some of what you just said, you can restrict certain categories of merchants or you can only buy from Amazon or you can only buy in the US or in France. Yes, exactly. Fantastic. Or you can literally set it to say, I need to approve every single transaction. Right. And, you know, when people are making reasonable size transactions, that's fine. And I think we're going to move to a world very quickly sort of at the intersection of agent of commerce and what stable coins enable in terms of sort of microtransactions and then reasoning about agents as the buyers.

30:24Agents are very, you know, microtransactions never actually really made that much sense, even in the context of content, because nobody wanted to put in, even if it was only five cents, no one wanted to put in their credit card to buy a five cents article. Like that was just too much friction. A and B, nobody wanted to process a credit card for five cents because you would have negative margins. But I think when we enter a world where like the human is not doing any work to execute the transaction, you're not typing in any credentials, you're not navigating to any webpages, the agent's doing it for you.

30:55Then microtransactions become viable and you pair that with, okay, and now it's, you know, burning down some stable coin balance. And we can talk what the work we're doing with tempo there, suddenly microtransactions become very viable. And I think we're going to quickly move to a world where, especially in buying inference or tokens or sort of AI sassy products, there's going to be, you know, you are going to be using your link wallet to make a ton of microtransactions and not approve everyone. But yes, if today you want to approve everyone, you can. that's not going to scale because no one's going to want to approve the one cent transaction for the little bit of data or the little bit of research or the token that I needed for this job.

31:40So what's our current thinking in terms of like everything that can go wrong? You know, you were saying like, when you talk about this to your friends, they say, well, what if the agent buys the wrong thing and so on and so forth? I mean, presumably that's the frontier. Everybody's trying to figure it out, but what's the current thinking? Yeah, we're thinking about this in a few dimensions. One is, to your question on, you know, what if the wrong thing gets sent, or the thing's not good, or whatever, businesses have to remain the merchant of record. Like, that's a core design principle for us, which means, like, we see our job as having agentic transactions behave the way human transactions do.

32:22Now, that doesn't mean that agency to behave like humans or be constrained in the same way humans do. But from the perspective of the business, they remain the merchant of record. They are selling, you know, they happen to have that sale facilitated through an agent, but like the business is still the business. Another design principle, which we've touched on a bit in different ways, is always provide appropriate granular programmable controls and guardrails to enable commerce to happen at scale. And that's basically saying, like, no matter how scoped in and in control you want to be or how much you want it to be a free for all, like you shouldn't have to adopt a new tool, move to a new wallet, change, you know, the underlying payment rails.

33:07You shouldn't have to change your product catalog or where you expose it. like basically out of the box, those controls and guardrails should ebb and flow. And I think in the consumer case, that looks like consumers remaining in charge of how Link authorizes agents to buy. And in the business case, you know, I touched on this briefly in the context of shared payment token, but it looks like making sure that the information we have about the goodness of the underlying buyer and the agent operating on that buyer's behalf are passed fully to the business to action intelligently. And so, you know, radar is our fraud protection product.

33:51We've had it for over a decade. It used to be really about transaction fraud. Now it looks at all kinds of fraud and abuse and bot and multiple layers of goodness where there's not just the end customer, but also the agent operating on behalf of the customer. And so those radar scores are actually included out of the box in the shared payment token for businesses to reason about it. But basically, I think business remains merchant of record. Consumer has as fine-grained guardrails and controls as they need. And then in the ecosystem, there is as much symmetric information as we can create. create.

34:31I mean, we now look across just about 2 % of global GDP. In the context of AI, we look at a very, very large share of that GDP because basically all AI buyers and all AI sellers are on Lync and on Stripe. And so, you know, that allows us to actually understand when something is going sideways. And we see very much our job to protect the ecosystem by providing that information up front. It's really fascinating as one unpacks this entire line of thinking around authentic commerce because that's what the technology can do, that's what you can do as a responsible player in the ecosystem. But ultimately, a lot of this will have to go to court one way or another.

35:18I mean, the legal system will need to adapt and evolve because if something goes terribly wrong in a large commerce transaction, who's at fault? Could it be the model provider underneath because like the agent went haywire? I guess all of this is going to take time to work this way through the ecosystem. Yeah, not just in purchasing, right? Like, you know, who is at fault for not just bad purchasing behavior, but bad behavior of all types when there's an agent involved. So I do think the landscape is changing there. When it comes to sort of payments in particular, you mentioned a bit ago that back in the day, people didn't really trust just like putting their credit card on the internet.

36:01I think it's interesting to think like with things like shared payment tokens, agent never sees the underlying credentials. Each transaction is scored in real time by Stripe Radar. Merchants go and handle the transactions. Credentials like never travel through untrusted services the way they can when a human types a credit card number into a random website. Not Stripe, but other random websites. And so anyway, I think it's also interesting to ask, is there sort of meaningful payments, trust, and safety upside here? And I could imagine, yes. So to play it back, agent payments would eventually be safer than human just typing cards in terms of fraud, in terms of mistake, presumably as well, right?

36:49Like if you type in the wrong number or something like that. Yes. done right in the limit, I believe it should be much safer. Today, I think the first order effect is it's just more convenient. Usually, it's not a fraudster on my credit card. It's like me fat fingering my own CBC. It's annoying and I'm frustrated. But right now, I think most of what we're getting is convenience. But in the limit, yeah, you could totally imagine, okay, now nobody's passing these random payment credentials over the internet. And by the way, people also aren't really necessarily reasoning about transactions coming in through 10 different payment credentials they have, right?

37:30They really have a wallet. They see what happens in the wallet. And that wallet is used in a very tokenized, secure way across sellers. So another really interesting topic I wanted to cover and that you mentioned briefly is Vibe deployment. What does that mean? It's what happens after you build the thing. And it's actually not talked about enough, but like the coding part, a lot of people talked about AI for coding. You know, when we were talking a year ago, that makes sense because it wasn't solved. Now that's basically solved. And we actually see that it's solved in our data. These are semi-random facts, but I find them interesting.

38:08So like agent traffic to Stripes documentation grew more than 10x since we talked a year ago. Agent traffic is now about 40 % of all our docs traffic. And meaning that the agents are trying to figure it out and they go to the Stripe technical documentation to understand... Meaning the coders now are almost as much agents as they are developers. And in some segments, they are basically all agents and not at all developers. And another example actually is our CLI, our command line interface, which historically was like a pretty niche tool used by like a pretty small group of developers. It now, it's just like exploded.

38:49And we were like, what is happening? And it's now 70 % of its API resource requests are from agents. So like you could think of like the majority of entities that are using our CLI today like aren't people. So anyway, I think those are just two like fun Stripe anecdotes, but there's lots of them across the ecosystem you can look at that tells you like vibe coding is real, right? Just your own lived experience is the same, right? That part worked. But then what? So like an agent writes you like a complete working application in 20 minutes. Fabulous. Like we love it. Okay. But you've still got this pretty big friction before that app is actually live, right?

39:31You got to go like, I don't know. I mean, it depends what you're doing, but you got to probably create an account with your database provider, then your auth provider, then your hosting service, and you're like bouncing around. You got all these dashboards. You're like, I don't know how you do it, but copy pasting stuff by hand and managing credentials and API keys and whatever else. And every one of those services has its own onboarding flow. So like a little different flavor than the payments flow we talked about, but like it's a pretty inefficient fill out a bunch of steps. and every one of them, just like the payments flows, was designed for like you and me as humans sitting down and clicking through this like weird setup wizard thing.

40:07And I don't think it really bothered any of us that much because we weren't doing it that much because the hard part was the coding part. But now that like the app can be built in coded in 20 minutes, like, okay, the long pole is deploying the thing. And so anyway, vibe coding was easy. Vibe deployment has become like more of the binding constraint. And so we actually, we recently, like in the last couple months, launched Stripe projects for this. But basically it's like agents should be able to well, sign up for and configure and integrate all of the services they need to deploy an app. and they should be able to do that like right from the command line.

40:50And it's not just Stripe Services, right? It was like a whole ecosystem of like, you get like Vercel and Superbase and Cloudflare and Twilio, like whatever you need. Clerk, I think we announced like 16 more partners a week or two ago now. And anyway, there's a lot of enthusiasm from the ecosystem because it turns out that everyone kind of has the same problem. Developers all have the same problem, which is now like the long pole is like the deployment and businesses have the same problem, which is like their onboarding was designed for humans and now they need it to work for ages. And to bring it home, why does Stripe care?

41:27Okay, so I think there's a lot of things that Stripe cares about that on paper, you'd be like, why does Stripe care? The honest answer is we care because it was becoming the bottleneck. Like the barrier to building is gone, but the barrier to deploying is like a real friction. And if you just zoom back, like, okay, separate from being a payments company. Our mission is to increase the GDP of the internet. A big part of that is how do we get more companies off the ground and how do we help them get kind of their first dollar faster and scale from there. And so if a person with an idea can create an app but can't deploy it, they can't sell.

42:00And so removing deployment friction directly obviously like expands the internet economy. And you can think of it primarily as just orchestration. All we're doing is orchestration. So if someone else wanted to and could and did orchestration, we'd be cool with that. But we were looking at the developers trying to get the thing live and we were looking at the businesses trying to get the thing used by developers and we're like, okay, I think we can make this market a little bit smoother. Yeah, and to play back orchestration because you have all those partners so ultimately you do hosting, observability, email queues, secrets, all the things, but that's provided by different vendors and you provide the glue to make sure that people can deploy their agents safely and efficiently.

42:48Yes, exactly. Exactly. Very cool. All right. So that's Vibe deployment. Tokens as money is also a fascinating topic. It's like taking it from the top. What does it mean to monetize tokens from a Stripe perspective? It's a big question because what does it mean to monetize tokens from a world perspective is kind of like, okay, actually, how's the whole next generation of B2B and some B2C going to monetize? Look, like the last kind of decade plus of SaaS had like pretty beautiful and simple to monetize economics, right? And in particular with SaaS, like you build a product once and then you get one more customer and it costs you basically nothing to serve them.

43:28Marginal costs are near zero. And that's why SaaS margins are really good. And that's why, you know, the fixed fee subscriptions or seat-based licenses work really well in SaaS. AI, and I say AI generally because, like, you could literally be selling LLMs, but you could also be selling, you know, some product that's a wrapper on top of LLMs or a product that's heavily powered by LLMs and requires a lot of tokens. It breaks that model because, obviously, every prompt and every API call and every task has a real marginal cost all of a sudden, which it didn't have in SaaS. The inference isn't free.

44:01And so you now have all these businesses where how your customers use your product directly determines whether you make or lose money. And that's a very different game. And, you know, from our vantage point, like part of what that boils down to is the need for usage-based billing as really critical for AI companies. Like you need to be able to meter what customers are actually consuming in real time and then charge them in a way, obviously, that aligns with your underlying costs. And do you see from your perspective, pretty much all the players in the economy from a vendor startup standpoint use usage based billing at this stage?

44:46Or is there still a mix between per seat and per usage? I see very few scaling or scaled AI companies that are still exclusively subscriptions or seed based. And I believe from my conversations with them that that is for the simple reason that the economics there don't make sense because you have some people that are using a ton and some people that are not using very much. And these people cost you a ton and these people cost you not very much. And it is very hard to separate the sheep from the goats and price them appropriately without a usage-based meter. That said, many of those businesses have a usage-based offering, but it is complementary to what you can think of as like a fixed fee subscription.

45:32So like a great example is Lovable. When they launched, they had like a simple subscription through Stripe Billing, makes total sense. They were early, they're moving fast. They wanted to monetize quickly. Subscriptions are also very familiar and easy for consumers. You know, if you think about a Lovable, Like some of their target users are like not very technical. Like how are they going to feel about like, oh, how should I reason about a token or a credit? Right. OK, so they started with subscriptions. But then as the company grew and some of their costs grew, their billing needs evolved and they needed to charge at least somewhat based on actual token consumption.

46:08And so what they did is a hybrid billing model, which actually we're seeing a large number of businesses, especially businesses that have a B2C component do, where they have usage-based billing on top of their subscriptions. So customers hit some threshold. And by the way, many people have freemium thresholds. But then above that, they often have like a fixed fee,$25 a month or$100 a month, up to some number of credits. And then above that, you have usage-based billing. So once customers hit the threshold, lovable, in this case, charges precisely based on the number of tokens consumed above that.

46:40And I think that alignment is good because you get a bunch of people in the door comfortable with subscriptions. But then at any volumes that matter, your revenue is scaling very directly with how customers are using the product and correspondingly with how much it is costing you to provide that product. So customers pay only for what they get. And you make sure you monetize for the underlying cost that you're going to have to bear. And 11 Labs is another example. They went through the exact same thing, literally, started with subscriptions. They recently moved to this sort of pay-go plan. And this is just a pattern that we see playing out pretty much everywhere in the AI space right now.

47:15Does billing change much once you start charging agents? So I think so for a couple of reasons. One is agents can consume at machine speed. and so even if you're like um even if you're doing a usage-based thing but you're like charging at the end of the month by the time you've gotten to the end of the month an agent can have spent a bit a human can too to some extent with agents working for that but like especially an agent can have spent like an egregious amount and so actually um i think in the world of agents what we're going to see more and more of is real-time metering, like what have you consumed, and real-time billing, which is actually what we have co-built between Metronome, so real-time metering, usage-based billing for very complex models, and Tempo blockchain, where agents are consuming tokens in real-time and paying down for the cost of those tokens.

48:21And that's important because agents can by machine speed, it's viable because of a bunch of the infrastructure between tempo and metronome. And agents are happy to just like pay as they go in a very literal way. Plus, businesses need them to be paying as they go so that they don't rack up a bunch of spend and then go dark. Yeah, it's fascinating also in terms of downstream consequences for what it means for like finance and accounting and like systems that need also to move at that speed. Totally. I mean, we have a revenue recognition accounting product and it's mostly needed by, I mean, we provide it to a bunch of people who, a bunch of businesses who have traditional subscriptions, but where there's the most acute pain is actually like traditional accounting in spreadsheets, like does not work when you have this just like proliferation of rows because these are like microtransactions are truly happening.

49:19I also think it's changing what it means to be an accountant. I will avoid naming the company, but I was talking to the sort of number two in accounting at a pretty successful AI company. And a couple things were interesting. One is they were like a hybrid accountant engineer, which they needed to be because of the scale of the data they were dealing with. And two, they weren't doing like accounting in the traditional you and me sense of like just close the books. They have to close the books for sure. But their job was also to find like weird anomalous stuff happening. And I was talking to them because they were like identifying some fraud patterns and wanted help with them.

49:57But anyway, just just very interesting. Like, OK, what is actually what does it mean to do accounting at like one of these AI companies? Definitely needs new tools. Definitely needs new systems. Probably a different different skill set. And then like what you're accountable for really isn't just closing the books. It's like looking across the whole RevRec stack and saying, what does this tell me about the health of our business? What does this tell me about fraud and abuse in our business? And what does this tell me about breakages in our product? Yeah, all of which becomes a massive data problem that needs to be treated in real time, right?

50:28Hence the rise of the click houses of the world. Like whoever can process massive amounts of data in real time becomes the core part of the required infrastructure. Okay. You wrote somewhere about token theft, and I wanted to make sure that we cover that. So what is that? What is token theft as a new form of fraud, I guess? Yeah, it's a new form of fraud. It's, I think, one of the most under-discussed topics in AI right now, maybe by a lot. Fraudsters have figured out that in AI, you actually don't really need to steal money or credentials. You can just steal tokens. And tokens have real value, right?

51:04You can use them to build things. You can resell them on marketplaces. You can wrap a new product on top without paying a cent and go and sell that new product. So it's kind of like resell, but it doesn't look like you're selling the subscription to the thing. It looks like you're selling this other product that you've come up with on your own, but really in the back end, it's completely powered by someone else's thing. And for AI companies, maybe this is implied by our earlier conversation on SaaS, but the risk is existential, right? If someone stole a little bit of your SaaS, like didn't matter because it didn't really cost you anything on the margin.

51:38When someone steals your tokens, if your fraud rate is high enough, the economics of your product actually break really fast. So anyway, because we work with basically all the AI companies, we have this interesting front row seat into what's actually happening and the fraud patterns playing out. And right now there's three, and I don't think these will be the three forever because we've already largely burned down these three over the last three months. But new ones will pop up. I think these three are just instructive for reasoning about sort of the breadth of the space. One is multi-account abuse.

52:12So this is like bad actors just like hammering your sign up again and again and again and again so that they can get those like new user credits. And the scale of this actually shocked me when I looked at the data. So like more than one in six signups at AI companies are this kind of abuse. And by the way, that's also very confusing for the company. Like, are these good customers? Who are these people? But it's very expensive for the company because like, I've spent a bunch of tokens on these free products for, you know, a relatively small number of people who are spending up a massive number of counts and actually consuming quite a bit of tokens.

52:45Okay. So that's like very, very top of the funnel. Another example that I think is interesting is, is free trial abuse. So these are like fraudsters come in, they create a free trial, they put down a payment method, they drain through all of the credits, but they never have any intent to convert. And free trial abuse has always existed in various forms across the internet, AI or not, but it's more than doubled on Stripe in the last six months. And most of that doubling is coming from AI and AI businesses. And just to give you a sense how lucrative this is, there's whole cottage industries built around it.

53:20So I don't know if you've ever been marketed a free trial card. I was recently marketed one. It's basically like, literally, it's like, oh, you can like spin up free trial cards. They expire in 24 hours. So you'll never have to pay. And like, if you or I had one, we'd probably use it for like some legitimate purpose to try out a service and just not have to go through the pain of canceling. But like literally fraudsters can just like explode these things and spin up a bunch of free trials and spend a bunch of tokens with no intent to convert. And then the third is a little further downstream of this, which is basically like we talked about the usage-based uh billing folks are racking up like thousands of dollars of uh costs and then being billed at the end of the month uh and never paying and so like whatever the dine-in dash that happens in restaurants is like a dine-in dash um but it's but it's for for tokens and unfortunately at that point again this is ai not sass so like the company has borne the costs uh of of those of those tokens and sorry if that's obvious and i guess I would be a terrible fraudster because it's not even super obvious to me.

54:21So if I get tokens for free, what do I actually do with them? So I guess if I get tokens from a general purpose LLM like Cloud or ChatGPT, I could see what I do with it. If I go to a cursor or a lovable or 11 labs, what is it that I actually do? Yes, exactly. OK, so what you do with it very much changes based on like what's the service, like what exactly were these were these tokens meant for. You nail on the head for like, OK, you just like get the tokens from the underlying LM fine for those sort of businesses that are a layer above that. A bunch of resale abuse. So you literally sell it sometimes in other markets for a slightly discounted price, a discounted price on the base price that you should have paid, but you didn't pay.

55:13So there's a dark, like a dark web of marketplaces where you say like use a cursor or lovable for$2, but my cost is zero, therefore I make money. Yeah, exactly. So I just, I give you my login credentials, whatever. There's also, but then there's just like, There's just this like, long tail makes it sound small, but a very domain specific fraud pattern. So for example, people steal tokens to create content that they then use to extract money in all sorts of scammy ways. So a simple example, we see people going in and like mass generating music tracks. And then uploading them to Spotify and Apple Music and then getting fake streamers and then collecting royalties.

55:59or then for like basically all of the rapper businesses there's this whole other layer of rapper on a rapper where like rather than just resell the subscription on places like taobao they'll like literally clone the ai company where you can just like vibe code a website the back end is just like spitting out exactly what you got from the service that you're stealing from and then you like sell the product as yours uh but cheaper you're gonna give it to people that they're creative i mean that sounds almost harder than starting an actual company yeah I don't know. I mean, a bunch. Yeah, they definitely they definitely put a lot of time into it.

56:32Tokens are also very valuable. And, you know, what's been what's been interesting as we sort of started seeing these trends, maybe at this point, six, nine months ago in various flavors, but then they escalated a bunch. You know, talking to AI companies, AI companies, I mean, the large AI companies are all over this. Like, it's like one of the most existential things for their margins. They have been like in the trenches with us, identifying the issues. They have been like, you know, we literally like, we see a problem, we build a model, we like, you know, multi-account abuse. okay, like at the time of login, there's an API, you send us what you know, we send you back a score, you block if they're bad.

57:10At the time of free trial start, same thing. As people are accumulating usage, you send us all the metadata, we send you back whether they're fraudy and you can require a top-up or cut-off service or whatever. Like literally each of those, as we've seen them, we've like order of weeks gotten, you know, generally this isn't even like a full-fledged product, it's like an API. Like you send us some stuff, we send you some stuff. And the adoption has just been like, okay, so the AI companies are all over this. I think what's interesting is there's like, every company is going to become an AI company.

57:44And I don't think the industry at large is thinking about this yet, or has really even reasoned that like, actually, most of the fraud that's happening is not traditional like, credential or payment fraud. It's like, what we would historically have called like first party abuse, right? Like resale abuse or account sharing or multi-accounting or free trials, like first party abuse. And I think it wasn't that first party abuse didn't happen before. It's just at least in sort of sassy stuff, first party abuse didn't cost you anything. And so A, it wasn't that useful to get like a little bit of Salesforce for free for, you know, the examples we talked about wouldn't be relevant in, in most SaaS.

58:24Um, and, and B, uh, even if it was useful to figure out a way to like skim off the top, uh, and you could get a little bit of, of profit for it as the fraudster, it didn't actually cost the business that much because their marginal cost was zero. And so, you know, as every business becomes an AI business, I think, uh, we've, we've been, um, in the context of our work on, on radar, really reasoning about our fraud prevention product as moving from transaction to full customer life cycle and moving from traditional fraud to end-to-end abuse. But I don't think the whole industry is there yet. And you and I talked about much of the economic upside of AI, but I think that'll really only be realized if it can happen safely.

59:09So for example, six, nine months ago, I was talking to some of these AI companies. they'd be like, oh, I know I'm going to solve my free trial abuse problem by cutting off free trials. Like I'm going to solve my free trial abuse problem by only having a sales sold motion and only going after enterprises and not having PLG. And like, you know, I hear less of that today. Why? Not because the fraud's totally solved, but because everyone knows they need agents to also be their buyers. And if agents are going to be their buyers, they better have a self-serve motion. They better have a PLG motion. There's no way they want to siphon off that source of growth and only double down on a highly secure sales sold motion.

59:53But it's been interesting to see what's happened with token theft. And I totally agree the fraudsters are creative, but I think that's a manifestation of how valuable the tokens are. And so I actually don't think that creativity is going anywhere. And radar is one real time and two, presumably entirely AI driven as well. It's like AI fighting AI kind of a kind of thing. Yeah, it's real time. It's AI driven. And then actually, like most importantly here for its like differentiation, it's just looks across the Stripe network. So like, there's basically no good AI buyer we haven't seen before. And there are very few bad AI buyers we haven't seen before.

1:00:31And so that combination, you know, it's, yes, the size of the network and this 2 % of global GDP flowing through Stripe, but really when it comes to AI, it's the density of the network. And we talked about Link briefly, but to give you a sense, like, Lovable is a good example. As an AI company, 58 % of Lovable's volume flows through Link. Like, Like, Link is an extremely dense network when it comes to AI. And so, you know, you can sort of extrapolate away from that. But if we know who all the good buyers are and we've seen the bad guys be bad somewhere else, then combine that with Yes.AI and sort of these real-time APIs and you get pretty good fraud defenses.

1:01:15You mentioned Tempo at some point. Should we cover Tempo? How relevant is Tempo to the agentic commerce conversation? So I think there's a couple components of our work with Tempo that I think are really interesting. So when you think about agentic commerce on the business side, we launched the machine payments protocol, or MPP, and we built it with Tempo, and it's an open standard, and the way it works is quite elegant, right? Like an agent requests access to a service, and it can be whatever an API or an MCP server or whatever and then the service responds with a payment request and then the agent pays and there's no kind of account creation or checkout UI or human in the loop or sort of the way you and I would traditionally engage on the internet it's just this very kind of machine readable standardized way for agents to buy from businesses And that is really MPP is really the primary mechanism that we're seeing businesses use for kind of agents, agents as buyer.

1:02:20Um, the other, uh, collaboration with, uh, with Tempo that I am, um, super bullish on is, is more related to, to fraud, uh, because, you know, agents are increasingly becoming the users of AI products and agents can burn through tokens very, very quickly. And so we were talking a little bit about this dichotomy that a business faces where either you can siphon off self-serve and be really safe but grow slowly or open it up, including to agents, but then be putting yourself at risk for quite a bit of abuse and monetary losses. Neither of those is great. What you actually want to do, especially when the agents are the buyers, is you want to track the tokens as they're consumed.

1:03:08And you mentioned the infrastructure. You want to track them in real time at substantial scale. And then as importantly, you don't just want to track them as they're consumed. You actually want to like collect payments as they're consumed. And so we call the streaming payments. And it's what Metronome and Tempo, which is the blockchain optimized for payments that Strype helped co-build, are making possible together. And so just Metronome's job is like track the usage in real time. And Tempo's job is just enable like fast, low cost, high volume micropayments that settle instantly. Obviously, they settle in stables.

1:03:43And so put together, it's like AI companies can charge generally agent buyers as tokens are consumed instead of having to choose between kind of closing off business or getting stiffed on the invoice. So we're very bullish about tempo and stables in general in the agent economy writ large and for the purposes of agent e-commerce. Super interesting. All right. So perhaps as a last topic, you guys have all sorts of interesting stats about the AI economy in general, but in particular, AI startups. I think we covered some of that last time. What have you seen in the last year in terms of AI startups, trends and facts and growth rates?

1:04:32What have you seen? So it's interesting. When we talked last year, we talked about the growth of AI startups and how they looked different than traditional startups. And what I would say today is like, you know, there's definitely a delta between AI startups and non-AI startups. But what's more striking to me is how AI is just changing the startup ecosystem generally. Um, and so, you know, in the vein of vibe coding and vibe deployment and all of that, like, uh, new business registrations are, um, well, so I think in general, the, the business formation story is underappreciated. So new business registrations are up basically around the world, at least for advanced economies.

1:05:13They're up like 40 percent in the Netherlands and 70 percent in Finland and 80 percent in France. And so there's this sort of like surge in dynamism that like, yes, we see and feel in here in the U.S., but it's happening across advanced economies. When we look at this with the stripe lens, the pace of new businesses launching has doubled since we talked last year. And not all those businesses are AI businesses, but many, many of them were made possible because of AI. And they're not just getting started. They're also scaling. So like Atlas is our product for founders to incorporate. And Atlas startups from the 2026 cohort, and it's only June, so it's early in their life cycle.

1:05:56But they're tracking to like five times the revenue of last year's class at the same number of months. Five times. Five times. And some of that is there's more of them. But a bunch of it is they are getting to their first dollar faster. And then they are scaling up more quickly. And getting to their first dollar faster, a lot of that is for sure AI. And then scaling up more quickly. A big part of this is actually how they're going global. And I don't know how much of that is AI or not. But like sort of the old model people had was you get big. And then once you're big, you deserve to go global.

1:06:34And what we're seeing with like, increasingly over the last year is you literally go global from it doesn't literally necessarily literally mean every country, but it's like you're in dozens of countries on day one, like your launch day. And that is how you get big, you get big by being global. And so I guess we're talking about AI. So I could use like emergent labs as an example, right AI platform for, you know, you build and deploy these kind full stack apps. So they were founded in 2024 in the US. 70 % of their revenue comes from international sales. And they do material business in 16 countries, like a substantial share of their revenue comes from many, many countries.

1:07:16So anyway, I think there's this sort of like, yes, there's AI companies. And you know, it's moved from, you know, just being sort of the underlying providers to like a lot of wrapper businesses, proliferation of wrapper businesses across every single vertical. But I think what's kind of more interesting from the tops down macro perspective is just, you used to have to be a developer to be a builder and therefore build a business. Now you kind of have to have an idea plus vibe coding, plus vibe deployment, plus reasonable economic infrastructure. And then, you know, we see this proliferation of new businesses And they aren't, their development is not easily arrested, right?

1:07:56Like they, even with one employee, they start monetizing early, they grow quickly, they expand across a bunch of markets. and I think that's almost certainly helped by all of the operational work that can also be done with AI which is less in sort of the day-to-day core wheelhouse of straight we help with some of that on the accounting or revereck or whatever but you know much of the customer support and other operations are obviously done by other businesses but I think it's an interesting time and there's a lot of discussion of like is AI going to lead to a small number of firms with heavy market share and not a lot of competition.

1:08:35And I think one of the reasons I'm bullish on the AI economy is at least so far, for sure, there are big guys who have things that are highly complementary to AI. We don't need to go through that list. Everyone knows them that are exploding. But there's also an explosion of little guys coming in the market and not just being created, but like reaching customers and growing quickly. And so I think that bodes well for competition and for economic growth. All of this is obviously extraordinarily exciting, But since we talked about token costs and, you know, usage base and all the things, do you worry that like a part of this is a little bit people on the buying side and on the usage side got a little carried away, didn't quite realize the dollar amount that using AI represented and that there might be some kind of backlash against that hyper growth?

1:09:27So, I mean, we've all read the stories of companies who have accidentally gone bananas on token spend because they had like no control over what their employees were doing. I think the companies who have truly gone bananas are by and large the companies with pretty deep pockets, which isn't to say that it's not like going to be a problem in the economy if that spending continues. But there's smart companies that are well run, well managed. They can handle a month or two of poor decisions and runaway costs. And, you know, by and large, I don't exactly know what we're talking about at these companies, but we're talking about like 2%, 3%, 4 % of their headcount costs are going into tokens.

1:10:05And 30 % of that or 40 % of that is inefficient. I think they can like pretty quickly get back to an inefficient frontier and it's not really anything existential. For the little guys, I actually don't think that's happening. Um, you know, many of them are sort of making fixed fee, fixed fee purchases, and they're still on small plans and, and whatever. So I mean, the stories are out there. And by the way, just to be clear, like, I think we have a lot to learn about the efficient frontier of AI use. And there's a component there around, like model routing and what models use like for the job. And there's a component around just like observability, which I think many companies have found themselves to be behind on.

1:10:42And then there's just like, like norms and controls and guardrails. And, you know, I think we all want high ROI usage of LLMs. But we also want employees to know when they're, you know, rack it up really substantive costs. And we want them to be able to tell us whether they anticipate the ROI of those costs is going to be there. So, yeah, I think there'll be like a little bit of recalibration, but I don't think anything existential has happened that's going to annul the sort of long run upside here. All right. So as a last question, when we talk again in a year from now, where do you think we are in terms of authentic commerce, maybe using, you know, L1 to L5?

1:11:30And I want to hold you to the prediction, but like directionally, what do you think realistically is going to happen in the next 12 months? I think the most interesting thing in the next 12 months is, I mean, I think we'll move up and I don't know if it's going to be to four or to five. But I think the more interesting thing is actually that like, we talked about agents as economic actors, mostly in the context of buying. But I think we'll start to see, and I'm not saying this will be proliferated everywhere, but I think we'll start to see agents that are like multifaceted economic actors. They're buying and they're selling and they're provisioning infrastructure and they're running businesses.

1:12:06And they're like doing the thing end to end. And again, I don't know how many of these there will be or how they'll operate or whether it'll be like with each other or like in these like weird niche silos. But I think all of that's just going to continue to demand more purpose built infrastructure, including financial infrastructure versus just the sort of old human centric commerce stack. And so that's kind of where, like, what does it look like when, okay, I'm Vibe coding and I'm Vibe deploying, and then my content is my offerings are default exposed. And actually, we just we just quietly went to public preview on on Stripe Directory, which is just a really easy way for agents to discover providers.

1:13:01And then through Stripe Projects, they can integrate them directly. But like, OK, and then an agent's like also discovering everything and integrating it and buying it and then, you know, creating a service out of the combination of things it has provisioned and integrated and bought and is starting to sell a thing. Anyway, like this, this idea of like an agent as a micro firm, um, I think would probably be the most interesting, uh, thing to see 12 months from now. And again, I don't think the median, uh, firm is going to be, uh, I forget a solopreneur, a, what would it be? A solo agent, uh, or, uh, I don't think that's, I don't think that's a world to live in, but I think 12 months from now, we could totally see some examples of that, that sort of paved the path for, the whole thing end to end.

1:13:52And this always happens with a new technology, right? You take your current processes or market or whatever you have, and you figure out how does the new technology make that 5 % more efficient or 10 % more efficient. It keeps me from having to type in my credit card number, right? But where it actually gets interesting is where we start to reimagine how the system works. And it's not Emily permissioning an agent to buy on on her behalf, it's Emily has an agent who's tasked with running a business. And that includes buying some things and selling some things and making some profits. And that'll be the world that I would like to be talking about 12 months from now.

1:14:31Well, Emily, it's been another amazing conversation. Thank you so much. Really enjoyed it. Thank you. Hi, it's Matt Turk again. Thanks for listening to this episode of the Mad Podcast. If you enjoyed it, we'd be very grateful if you would consider subscribing if you haven't or leaving a positive review or comment on whichever platform you're watching this or listening to this episode from. This really helps us build a podcast and get great guests. Thanks and see you at the next episode.