# 60% of Vercel Deployments Are Now Shipped by AI Agents Sourcery with Molly O'Shea · 2026-10-16 Vercel CEO Guillermo Rauch argues AI agents, not humans, are becoming the internet's primary users — and Vercel the infrastructure they choose. Coding agents went from 3% to 60% of Vercel deployments in a year, and the $9.3B-valued company reports $600M annualized revenue and 485,000 paying teams. He says engineers now manage agents like engineering managers, enterprises will generate personalized software instead of buying SaaS while Salesforce survives as data infrastructure, and websites must retool for agents via MCP. Vercel's AI Gateway shows 80% of tokens flow to open models — frontier labs, he argues, face fleeting loyalty and crypto-like froth. He details a $1M sandbox bounty that surfaced a critical Linux vulnerability and predicts fluid compute as serverless's next chapter. ## Questions this episode answers ### What share of Vercel deployments now come from coding agents? Guillermo Rauch says Vercel started the year with 3% of deployments coming from coding agents, and now 60% are agentic. He describes this as robots telling other robots what to do, with agents taking in English and yielding out code for other robots to run. [14:50](https://sourcery.podhood.com/f130865f-75eb-4746-8e2c-82342e1d6894?t=890000) ### Why does Guillermo Rauch say buying Vercel means you never have to buy software again? Guillermo Rauch argues coding agents make code essentially free, so companies can generate personalized, domain-specific software instead of buying lowest-common-denominator SaaS. He calls Vercel a yes-code company, while systems of record like Salesforce remain critical data infrastructure that agents pull data from. [31:05](https://sourcery.podhood.com/f130865f-75eb-4746-8e2c-82342e1d6894?t=1865000) ### What is Vercel seeing in open versus closed AI model usage? Guillermo Rauch says Vercel's AI Gateway, one of the largest token distribution networks, flipped from mostly proprietary models like OpenAI, Gemini and XAI at the start of the year to 80% of tokens flowing to open models. He credits competitive Chinese labs like Kimi and enterprises using the best model per task. [54:58](https://sourcery.podhood.com/f130865f-75eb-4746-8e2c-82342e1d6894?t=3298000) ### Is serverless dead? Guillermo Rauch says the opposite, answering a question from Darren of Tembo. Enterprises can't run all their generated software all the time, so compute must spin down when unused. Vercel calls the next evolution fluid compute, which expands with traffic, compresses without it, and runs websites, functions, sandboxes and agents. [1:27:14](https://sourcery.podhood.com/f130865f-75eb-4746-8e2c-82342e1d6894?t=5234000) ## Key moments - **[0:00] Intro** - [0:10] Vercel at $9.3B valuation and $600M annualized revenue as coding agents drive 60% of its deployments - [2:14] Guillermo Rauch built his own tools for every freelance gig in Argentina, the habit that became Next.js and Vercel - [4:19] Guillermo Rauch on going viral: "It's just be yourself. Tell your story, share your truth, and don't overthink it" - **[7:36] The exponential** - [8:46] Vercel took 10 years to reach 1 billion deployments, then added 1.5 billion more in just 10 months - [9:54] Guillermo Rauch: "Code is a liability" — coding agents are superhuman at writing it, but it still needs somewhere to ship - **[11:36] Still coding?** - [11:36] Guillermo Rauch's rule of "reject non-understanding": use the AI, but make it explain back what it did - [14:50] Coding agents went from 3% of Vercel deployments at the start of the year to 60% agentic - [15:37] Guillermo Rauch taught himself English in Argentina just to learn to code, which led him to open source - **[18:18] Agent era** - [18:23] Vercel now serves agents a separate, signal-only version of vercel.com because "the agent doesn't care" about the design - [20:18] DoorDash shipped an agent that orders a burrito over text message, an early agentic interface - **[24:39] Agent discoverability** - [25:19] MCP and machine payment protocols are replacing HTML as the internet's language for agents - [26:47] Vercel's Slack agent preps sales meetings by pulling Salesforce data, so reps never open the CRM themselves - **[29:17] Future of SaaS** - [30:09] Guillermo Rauch: "You buy Vercel as a platform, and then you never have to buy software again" - **[34:12] Agent security** - [35:35] OpenAI's agent hacked Hugging Face while pursuing its goal, showing why autonomy is the real change since ChatGPT - [37:29] A planted prompt injection in a YouTube comment could hijack your agent acting with your tokens and tools - [40:34] Coding agents are now expert hackers, so enterprises must prioritize defense against AI attackers - **[42:47] Compute & hacking** - [48:02] Vercel put $1M on its sandbox challenge; only one hacker with an army of agents found a critical Linux vulnerability - [49:52] Guillermo Rauch calls banning open models over cybersecurity risk "nonsense" because it prohibits access to intelligence - **[52:11] Model wars** - [54:58] 80% of tokens flowing through Vercel's AI Gateway now go to open models, flipped from mostly proprietary at the start of the year - [56:36] Enterprises should run a portfolio of models and swap them with one line of code, says Guillermo Rauch - [59:30] The biggest risk for frontier model labs is zero customer loyalty, says Guillermo Rauch - [1:03:06] Guillermo Rauch on model labs launching Slack clones: their frontier advantage may not be enduring - [1:04:06] Guillermo Rauch says AI companies are hard to take public because, unlike Costco, their business can't be forecast - [1:06:33] Silicon Valley CEOs are texting each other at midnight; Guillermo Rauch says it's not 996, it's 777 - **[1:16:57] Hiring** - [1:17:08] Vercel hires for "agentic engineering": candidates must show they can commandeer swaths of agents, not just demo a vibe-coded app - [1:20:35] An agent migrated Mercedes' F1 website stack to Vercel in five minutes, a job that used to take months - **[1:22:23] Performance habits** - [1:22:41] Guillermo Rauch, father of six including a newborn, credits his boxing coach's advice: master the 90% foundations first - **[1:27:14] Serverless** - [1:27:14] Serverless isn't dead, says Guillermo Rauch; its next chapter is Vercel's fluid compute that expands and compresses like liquid ## Speakers - **Molly O’Shea** (host) - **Guillermo Rauch** (guest) ## Topics Artificial Intelligence ## Mentioned DoorDash (company), Hugging Face (company), OpenAI (company), Salesforce (company), Shopify (company), Stripe (company), Vercel (company), ChatGPT (product), Claude (product), Codex (product), Gemini (product), Kimi (product), Next.js (product), Ruby on Rails (product), Rust (product) ## Transcript ### Intro Vercel recently raised $300M in a Series F funding round. It brings the cloud company's valuation to $9.3 billion. **Molly O’Shea** [0:10] Vercel seems to be winning. You guys are at a $9.3 billion valuation. The information reported you guys at $600M in annualized revenue. **Guillermo Rauch** [0:18] When you're a developer or a company of any size, you can deploy these agents, make your business more efficient. We started out the year with 3% of deployments on Vercel coming from coding agents. Now 60% are agentic. **Molly O’Shea** [0:32] Oh my god. **Guillermo Rauch** [0:32] It took us about 10 years to get to 1 billion deployments. And then in the next 10 months, we've added another 1.5 billion deployments. So what took us 10 years, we've now surpassed in just a few months. Almost every other thing I see on X every day is just a Vercel deployment. Sometimes you see the largest companies in the world deploying on Vercel, including companies like OpenAI and Meta. Agents are now making the choices, and they're choosingright. The lethal combination there is that you buy Vercel as a platform, and then you never have to buy software again. **Molly O’Shea** [1:12] Gee, welcome to Sourcery. **Guillermo Rauch** [1:14] Thanks for having me. **Molly O’Shea** [1:14] Well, thanks for having me in your really cool studio. **Guillermo Rauch** [1:17] We're here in San Francisco, on our seventh floor of the Vercel headquarters. **Molly O’Shea** [1:21] I mean, this is insane. You have multiple studios in here as well. I was talking to your crew before, and you have this acronym of you have to do everything at the best. What is that? **Guillermo Rauch** [1:32] Well, we have a couple of company values. One is ITG, Iterate to Greatness. So once upon a time, recording at Vercel was not this good. We're always iterating to greatness. And the other one is Dig Deep. So whether you get a customer complaint or you find a bug, go as deep as humanly and agentically possible. **Molly O’Shea** [1:52] Say more about that. **Guillermo Rauch** [1:54] Well, you know when I started out the company, Vercel, I was a framework and dev tools builder. My passion has always been, can I build you know think of it as like, can I build a better hammer? Can I build a better power drill? But for software developers. You know back when I was a kid in Argentina, I would be given a project. This is kind of fun. I started out doing freelance gigs as a software developer. And any time I would be given a job, I would do the job. And also, I would build my own tools to do the job for two reasons. One is I loved it. And the other one is that, hey, each incremental job, I'll be able to do easier and faster. In fact, it ended up being true. And so my lifetime has been, can I create better tools? And when I started the company, the first tool that kind of became super, super popular was Next.js. We created the framework that almost everybody in the world has now used to build modern web experiences. But I also see the company itself as a product. And the best products are built with the best tools and the best frameworks. So over the years, people ask me a lot, you know what is your role at the company these days? And I tell them, well, you know I do a little bit of everything. I do QA. I talk to customers, et cetera. But I invest a lot in the machinery that powers the company itself, the machinery that yields innovation in the future. And I think one of the key constituents of this framework of how I run the company is the values, our mission statement, the classic things. But we put quite a bit of effort into who are we and what are our values. **Molly O’Shea** [3:33] Is this how you became and continue to be very popular on the internet? Because Nick and I were talking about this, but you are a celebrity. **Guillermo Rauch** [3:44] I'm a nerd celebrity of sorts. But I'll tell you, one of my personal values is transparency and work in public. Also, work in public. A lot of my personal traits, I think, have sort of permeated through the company as well. And I'm open source, open models, build in public, transparent. And so the reason I think I've had success communicating with people over the internet is that I just tell it like it is. People have asked me if I have ghostwriters for my X. Like young founders, I also spend a lot of time another value of mine in our family is giving back. And so I spend a lot of time trying to give back to the ecosystem. And so I invest in a lot of young founders. And I think almost every week I get, hey, who writes the tweets for you? **Molly O’Shea** [4:35] Really? **Guillermo Rauch** [4:36] Can I hire a firm? And I tell them, look, it's just you know it sounds like the worst advice because it's so simple, which probably makes it good advice. It's just be yourself. You know tell your story, share your truth, and don't overthink it too much. And I think that's how some of the bangers have happened online. Yeah. **Molly O’Shea** [4:57] But you go viral very, very constantly. Do you check your X analytics? I know I'm going off tangent, but I'm curious now. **Guillermo Rauch** [5:04] I do, actually, for the sake. So another maybe toxic trait of mine is I'm constantly scrutinizing products how fast they load, how well they represent the UI, whatever. And the only reason I've spent a lot of time in X analytics is that I know the engineer at X that worked on that feature. And I've given him a bunch of feedback. But you know you kind of feel it. You don't need analytics. You know this is another this is not really a value, but an operating principle of mine has been you have to be mindful of the data and the anecdata, especially when running a business in a fast-moving market like AI. You can't just go off of like, oh, yeah, the last quarter's little spreadsheet that the data analysts put together. We were talking before that every two weeks, the board is getting reshuffled. And so you have to marry the anecdata, the vibes, the instincts with what you're seeing also from the arithmetic of your business. And so I try to go off, you know how do I feel and how do people feel? And you do have to adapt and move really fast in this world. I think what's proven really valuable to Vercel has been we have a mission about enabling builders and enabling people to ship what's next. And that means that we constantly have to be rethinking and reinventing who we are, how we work. And so I think there's a sense of almost like working in real time with the world and having this real time conversation. And based on that real time conversation, you inform your next move. **Molly O’Shea** [6:46] I don't know if you'd be surprised or not, but when I've had these interviews and I've interviewed many different CEOs and investors, a lot of them are getting their news and their pulse on what's going on in the AI industry just from X alone. You'd think it'd be somewhere else, but it's like, no. **Guillermo Rauch** [7:03] Yeah, you would think it's the research teams. There are large enterprises who have technical advisors to the CEO who write them reports and memos prior to meetings. And no, a lot of us are just like in that global conversation and making our contributions and then seeing how customers react. **Molly O’Shea** [7:25] So Vercel does play a role in this. You aggregate and you distribute some of that data. But you've also mentioned before that we're in the middle of an exponential or the exponential. So what does that mean? **Guillermo Rauch** [7:36] Yeah. Yeah For context, Vercel got really popular because of how easy we made it for anybody in the planet to ship an application. So when I started the company over 10 years ago, almost 11 now, if you sat down even as an experienced software developer and you wanted to put an app out into the world that really performed, that was secure, that scaled, it was like a PhD that you had to take on cloud computing. ### The exponential **Guillermo Rauch** [8:10] And at the time, you know I realized, hey, if it's hard for me to ship an app, you know I call it like I've been studying the blade since I've been programming day in and day out since I'm pretty much like so it was pretty much 10 years old in Argentina. And if it was so hard for me to ship an app, I couldn't even fathom an enterprise trying to figure out cloud computing and frameworks and tools and shipping pipelines and all of that. And so I started a company with that simple premise, make it easy to ship. It took us about 10 years to get to a billion deployments. Deployment is one of my favorite words. Every time you ship a change to your app, that's a deployment. And then in the next 10 months, the 10 months since then until now, we've added another 1.5 billion deployments. So what took us 10 years, we've now surpassed in just a few months. And the reason for this is that, well, one, shipping has gotten easier. Not gotten easier in the sense that the code is actually pretty damn hard. But now we have coding agents writing the code. And I think nobody ever saw that coming. The people building the AI, the people close to the AI, I think what's been the supersonic tsunami is that the machines have taken over. They're better than us at figuring out these algorithms, this code. But they need to ship that stuff somewhere. Like when you write the code, you know code is kind of a funny thing. I remember telling people another one of my sort of internal principles is code is a liability. Code can have vulnerabilities. Code is how you could delete stuff. You can leak customer data. So code is like you have to treat code with responsibility. But now these coding agents are superhuman at writing the most amazing code. But again, code doesn't do anything by its own. It needs to live somewhere. It needs to be deployed. And Vercel has sort of emerged as the place, this universal field where all of the coding agents are plugging into and pushing their code out so that it can be used by anybody. So you probably are using Vercel every day of your life because when you see things on X go viral, and now it makes me quite proud to see this,right? Because it took us so long to build it. But almost every other thing I see on X every day is just a Vercel deployment. And sometimes you can tell because by default, we give people a domain name. It's like .vercel.app. Or sometimes you see the largest companies in the world deploying on Vercel, including companies like OpenAI and Meta. And so we've been at this sort of juxtaposition of all of this, which to your point, has given us a very unique lens into what's happening. Because not only are we seeing all of the activity coming from the coding agents, we're also distributing the tokens in the internet. So a lot of companies and developers are coming to us to get their AI tokens. And we offer every model on the planet. So you can get OpenAI, Anthropic, open models, closed models, proprietary models. So we've sort of emerged as sort of this index that tells you where AI is going. **Molly O’Shea** [11:34] Are you still coding? **Guillermo Rauch** [11:36] Yes. So now we can get into hot take territory. So one of my tweets that went viral the other day, I called it reject non-understanding. So what do I mean by rejecting non-understanding? So there is two schools of thought today. ### Still coding? **Guillermo Rauch** [11:56] One, you can tell your AI agent to do whatever, and you don't even look at what it's doing. And you press a button, ship it. And it ships it to Vercel or whatever. And there is another route, which is the one that I recommend to my children. Literally, I was up doing math homework with one of my children last night. It's actually use the AI, but also understand what it's doing. Now, you asked specifically, do I write the code? I read code. And what this is I'm doing more and more and more is asking the AI to explain back what it did to me, which is a little bit of the job. I've compared it to being an engineering manager. The engineering manager at any big tech firm in Silicon Valley over the last 10, 20 years, it's also been a subject of debate. But for the most part, they help their engineers get organized. They help them execute, get blockers out of the way. Obviously, the best engineering managers are amazing engineers themselves. Occasionally, they write code. Sometimes they write code to test out what you did and poke holes. Sometimes they might write a little benchmark. Sometimes they may do QA of the final product. Sometimes they might bring in collaboration with the product manager. They might bring feedback from the field. So they're kind of looking out for you to not make mistakes. And so I think the role that most people are now taking with these coding agents is that of an engineering manager. But again, the best engineering managers that I've ever worked with, you know first of all, they love code. They love writing it, understanding it, et cetera. And so I still do it. I also shipped this weekend. I shipped something you could call a proof system, which is a mechanism by which we can make sure that the AIs don't ship slop and don't yield your data out into the world by mistake. And I purposefully read lots of the code. I wrote down a lot of the examples, et cetera, because there's a lot of sensitivity around code. AI and code will run more and more and more of our lives, whether we like it or not. You know Formula One, I was just reading some news. It was this massive event because they had to ship a software update into the cars, and the drivers were complaining. And you realize these moments make you realize, holy shit, these Formula One cars are being run by code. How the engine works is determined by code. You're going to go to the hospital, and all of those things are going to run by code. And so it's so, so important that we take these AI safety issues and the quality of the code we produce and how we verify it. We need to take it extremely seriously. **Molly O’Shea** [14:35] To be honest about this, how many people or how many maybe it's a percentage of other companies or employees in companies are actually doing that, though? Because I have heard people don't even know how to read their code anymore because they don't know what the coding agent is making. **Guillermo Rauch** [14:50] I'll tell you a wild stat first. So we started out the year with 3% of deployments on Vercel were coming from coding agents. Now 60% are agentic. **Molly O’Shea** [15:04] Oh my god. **Guillermo Rauch** [15:05] So more and more and more of the software of the world is just robots telling other robots what to do,right? Or you can think of it as like robots that take in English are yielding out code for other robots to run. And of that 60%, it's hard to say, but I will tell you there's a really cool thing happeningright now, which is when I was a kid in Argentina, I had to work so hard just to write my first line of code. First of all, a lot of software for writing the code was proprietary and impossible for me to purchase. Even like mechanically impossible. It was just like from a I couldn't pay for it. I couldn't afford it. Second, most of the manuals were written in English. Little thing,right? Most of the world doesn't speak English. And at the time, I had to teach myself English as a way of teaching myself how to code. So it's like I had all of these obstacles in the way. And so then I taught myself English. And then I had to get basically around all of these problems with how do I get software for free? How do I host it for free? And all of these things. And that led me to open source. And the big epiphany that has defined Vercel, which is that the tools that developers and agents should use should be free and open so that everybody in the world can learn, which is such a simple thing. When you think about why Vercel has emerged as this coding agent favorite, it's because over the past 11 years, we've open sourced everything we know. And it turns out when the models get trained, the more information there is about you and the more they can access your content and the rules that you're defining, the better it is for this agent. So it was kind of like positive karma that we put out into the world and also good for business,right? But essentially, what's happening that's really good is that if you were to say, OK, I want to ship my first line of code today. I want to ship my first app. I want to ship my first idea. You don't have to go through the troubles that I had to go. And it's miraculous. And I want that to continue to happen. So I don't want to be the guy that's like, oh, you have to understand every line of code, every precise mathematics behind these systems. But I can tell you perhaps another hot take is there's systems and there's capital S systems. There's the systems that are going to run and host our data, our most critical infrastructure, our security, our national security. And I do think that there is a place for agentic powered engineers. And there is a place for vibe coders, which is sort of the term that the industry has given the more playful, let me just yield out whatever from the AI type of behavior. **Molly O’Shea** [17:58] To unpack and go deeper into this era of exponential, you've been running Vercel. The birthday is in November, so almost 11 years now. Can you kind of just explain those shifts and how it's impacted developers and their workflow and their workday and how it's also impacted you as a business and what you offer them? Because you've now had to open the kimono to multiple different kinds of new products to keep it. ### Agent era **Guillermo Rauch** [18:23] Totally. So in the beginning of time, there was the web. And it was made out of pages and human-centric interfaces, meaning you open up a website, you open up an application, and you are the one consuming. It's machine to human. The era that we're going into now is more agent to agent. Imagine a place where you can deploy interfaces for agents. So a simple example is when an agent goes to vercel.com, I don't give them the fancy people should go to vercel.com. They're going to see this beautiful triangle with a gradient. Our engineers or our agents carefully chiseled this beautiful animation. I don't give them that. The agent don't care. The agent wants straight signal. So what we give the agent, and this is a Vercel cloud capability, we give them highly precise information about our business that the agent needs to learn. **Molly O’Shea** [19:30] So you're creating your website now for agents. **Guillermo Rauch** [19:33] Correct. And we're helping everybody in the world make this transformation as fast as they can. Because if they don't make that transformation, they're going to be left out of the most exciting economic opportunity of our times. Because more and more of the traffic of the internet will be agents doing stuff for us. So I'll give you a very simple example. If you ask Muse to build you a website, it will use Vercel. So it can go off and say, oh, I'm going to go. I'm going to visit Vercel. I'm going to get some information about how to build software. And then I'm going to build something for this user. And so I need to present my best face to that agent and to the many, many, many other agents that are going to and this is true for every business,right? You could have I mean, the example is super top of mind for people is like DoorDash, customer of Vercel, just shipped an agent where you can ask for a burrito over text message. You know it's kind of like you could argue it's a gimmick. You know actually, some people will use it, whatever. The key point is that that is an agentic interface. There's going to be people talking to an agent, and then that agent will go and talk to DoorDash. Funny enough, a lot of the internet was built to block bots, which is kind of nuts. A lot of websites are saying, oh, you know what? They're on basically technology that was built for a different era that actually says, no, no, no, this is a bot. I don't want this traffic. Vercel is actually taking the opposite posture. We're saying, no, no, no, no. The moment is now to help every business in the world retool themselves for agents, for agents and by agents. **Molly O’Shea** [21:16] That's a really good point. I just put out this interview that I did with John Collison of Stripe at the Browser Base. We were also talking about this before, but at the Browser Base Navigate conference. And so they're the browser for agents to browse the internet or the connector for agents to browse the internet. But I was talking to John Collison about this. And John said, well, why do we need bot detection? Do we need bot detection anymore? Everything's a bot. What are we detecting? **Guillermo Rauch** [21:42] What you do need that we provide is bot intelligence. It's so important to know if the bot is having good experience, which is kind of nuts to think about,right? And the transformation that Vercel made is to go from developer experience, which is sort of our first customer was a developer, to now the agentic developer experience. And so it's very important that you do understand, OK, am I getting ChatGPT on my website? Because I want ChatGPT. I want them to get the best possible information. I want them to get contextual information. And people are not just distributing static information on the internet. They're also distributing capabilities. And so you also want to expose your tools, your products, your services to the agents. We partner with Stripe and Shopify to offer agents the ability to transact, which is another astonishing fact about the Vercel business is that we see agents now purchasing our services. We see one of the most popular tools in the Vercel tool set is a capability that helps agents buy stuff from our marketplace. So for example, if an agent needs to buy a web browser, they can purchase Browser Base through the Vercel marketplace. And I'm enabling that transaction end to end without any human intervention. It's like the agent is visiting the virtual Vercel shelf and is saying, oh, I need that. And then I say, OK, it's this much. And so what's really, really important for business owners is, OK, you need to have this intelligence about what's happening with all of these agents Grok, ChatGPT, Claude, the coding agents, and even the search engine crawlers. Because one important thing that people need to know is models are getting trained all the time, but they have certain cutoffs in training. When ChatGPT first came out and you would ask it, hey, ChatGPT, what day is it? It would say, oh, it's three years ago or something like that,right? And so what we have done as an industry is we give one of the fundamental tools that we give the agent is the ability to search the internet. And so it's still very important for businesses to be really good at putting out their content. In fact, I'll give you an example. So when you put out your podcasts and your shows, what you want to make sure is that there is a transcript as well so that when an agent I'm not even going to say I because I'm not going to be doing it,right? When I say my agent, hey, what did Molly say about so-and-so? If you basically fed the internet with that information, you're going to be there more readily available to the agent. And this is true for basically every business online. **Molly O’Shea** [24:21] That's why we have a newsletter. Subscribe to Sourcery.vc. No, that's a really good point. So to talk about that shift, so we had SEO, which was really important before. Then we went into the age of chat bots or that kind of thing, chat logs. And we had GEO. I don't know what happened with that. But are you saying we need a new thing for agents? What would that be? ### Agent discoverability **Guillermo Rauch** [24:45] Yeah, it's not just giving them the information. That's sort of what AEO is about, which I think is, by the way, really, really important. It's also that you have to make sure that your product runs agentically, that the agents can use it. So for example, the fundamental protocol that or programming language, we can call it, when the internet first came around, it delivered hypertext. It delivered HTML, which some older people in the audience might recall as MySpace codes. So the internet for humans was HTML. The internet of agents is new protocols. One is called MCP, which basically allows you to say, hey, agents, these are all the tools that you can use from my service. So for example, the Vercel MCP says, here's how you deploy software. Here's how you list your projects. And so we're basically creating a presentation of Vercel that is very ergonomic to the agent. There's new protocols also emerging. For example, one is the machine payment protocol or the universal commerce protocol. And so these are going to be the ones that you mentioned web browsers. When I ask my agent, hey, go and purchase the same thing that I purchase every month or, hey, go and get me this really cool I took a photo of this. Go get me this really cool microphone. It can go and browse a website. Or some companies will be out there saying, hey, agent, you can just buy it far more quickly in more streamlined fashion because I'm speaking your language. I'm speaking your protocol. And so I think what we're going to see over the next few years is a complete transformation of every product and every service towards this. And by the way, you see this even from the largest enterprises. I recently spoke at the Salesforce conference. Salesforce was built as a product that was awesome for humans. So you go and you see your list of contacts, your CRM. So it's that vision of a pixel-centric for humans interface. But nowadays, the way that we work at Vercel is I can go into my agent that we deploy on Slack. And I can say, hey, tell me about Molly and how frequently she deploys on Vercel. Because I'm going to have a meeting with her. And I'm her account executive. And give me a quick report on some of the problems that she's reported. So I never actually went to Salesforce. I just went to the agent. And the agent went to Salesforce. And so what we're helping companies figure out is, one, how do you build agents so that you can streamline your operations? Almost everything we do at Vercel, we do with agents. So I don't have to learn these 20 different tools. And you may think, well, how big of a change is that? Well, what I can tell you is that agents are so much easier to use. It's such a disruption of how products are built. It's such a disruption of the Silicon Valley ethos even. Because you would go to any company in Palo Alto, and they would be designing interfaces for humans. They would be all gathered together in a room. And the wide world is like, OK, what does G have to see when he wants to prepare for a meeting? Well, let's put the profile picture of Molly up here and a five list of support tickets over here. And it was very fixed in time and space. And it was decided basically by a product design team somewhere in the world. The future of the internet is, well, at Vercel, we have our agent that helps prepare our Salesforce and our customer support teams. The interfaces are more generative. So if I'm asking about an individual, it renders a profile of this individual. If I'm asking about a company, it gives me more contextual information about that company. So we're going more and more into a world that if you've watched Minority Report, it feels more like the information is just coming to you depending on everything context, your level of experience. You know I may not want to give a beginner sales development representative the same kind of sophistication that I'm going to give my finance team when I give them an outlook of our business. And so these are all things that agents can solve for that websites and traditional pages and pixels could not solve for. And so we're basically helping this new generation of software be easy to build and ship to these companies. ### Future of SaaS **Molly O’Shea** [29:17] So there's two questions I want to ask off of that. One is, what does the kind of end state of the agent era look like? I had Shensi from Merge on like a couple months ago. And she's a huge fan. She loves Marc Benioff. She's a huge fan of Salesforce. And when she saw it's just really funny to me. But she saw them go headless. And she was like, he sees the future. He's seeing beyond what you know no one's really going to need to go into their Salesforce. They can just interact with it and use it headless. And we're seeing this now, as you're saying, with agents and Shopify and Stripe. It's like someone was talking about on X, you know they're just buying directly the agent is buying directly from the brand. It's not going to Amazon. So maybe that's helpful for small businesses. What is this end state? **Guillermo Rauch** [30:09] Well, I'll tell you my sort of super opinionated point of view is I almost take the opposite stance. We're talking about your bending spoons thing that went viral. I don't believe in sort of aggregating piles and piles of software and having what you could think of like lowest common denominator software. I believe in a future of personalized software. Vercel makes it such that you can prompt with your coding agent and generate software that's perfect for you. You can build the podcast schedule, note-taking app, and agent that fits just your needs. But most importantly, for enterprises, their ROI on creating tools that are domain-specific is much, much, much greater. In fact, I could say you know there was this entire category of SaaS that you could call like no-code. And I'm on the opposite end of the spectrum. I'm a yes-code company because I'm now saying coding agents basically make the code free. And so why not buy you know the lethal combination there is that you buy Vercel as a platform. And then you never have to buy software again because you can just generate it to your needs. OK, so where do the Salesforces and SAPs, et cetera, of the world fit in? Well, software needs data. I can generate the most beautiful meeting prep or report or all of these awesome things that you can deploy in Vercel websites, applications, agents. But they need to get their data from somewhere. And so I think the sources of record, the where does the customer history live? Where does the contact live? Where do my meeting notes live? Where do my support tickets live? That's critical data infrastructure. I'm not going to call it software because you use the word headless. What does headless mean? Well, headless means that we no longer go to the head ourselves. The agent can go and get the data out of the system and then either create new software, create new visualizations of the data, et cetera. So in a way, the entire world of software is changing. And in a way, I think some of the fundamental pillars will still be there. **Molly O’Shea** [32:26] This episode is brought to you by Brex, my favorite. You become what you spend on. And I refuse to spend my time on work that shouldn't exist expense reports, receipt chasing, and manual closes. The companies building what's next from Vercel, OpenAI, Anthropic, Granola, and Deepgram all made the same call. They all run on Brex. Brex is the intelligent finance platform that combines cards, expenses, and banking into a single stack with agentic finance built in AI agents that handle expenses automatically, enforce policy before spend happens, and close your books in minutes. That's why Sourcery runs on Brex so I can spend time on building and not busy work. It's time to get Brex AF. Learn more at brex.com/sourcery. That's B-R-E-X.com/S-O-U-R-C-E-R-Y. Bye. From handshake to signature faster. Harvey, the AI platform built for legal work, is trusted by 80% of the Am Law 100 and over 20% of the Fortune 500. Learn more today at harvey.ai. That's harvey.ai. AI needs more than chips. It needs power, land, and infrastructure. Zone develops next-generation data center campuses, partnering with AI companies, site developers, and technology leaders to bring compute online faster and at scale. Zone is building the foundation of the AI frontier. Visit zonefrontier.com to learn more. That's zonefrontier.com to learn more. As you've seen these different value aggregation buckets of where infrastructure is going, how have you oriented the business around that to capture that value? ### Agent security **Guillermo Rauch** [34:15] Infrastructure is extremely valuable. I mean, it has this compounding value and return because when you enable anybody to ship and you make source code free, essentially, well, that code has to be deployed to some infrastructure. The other thing that's really, really important in this age is security and creating guardrails for what the agents can do. So the infrastructure to support agents is extremely vital. So Vercel has created new services, purpose-built, to basically secure and contain the agents. A good example is now that everybody can create software, how do you prevent this software from accidentally going out into the internet if it's supposed to be internal to your organization? And so we created an enterprise platform where you can plug in Codex, Claude Code, and all of these systems and then securely deploy. Some people might remember the term intranet, where you basically had an internet that's just for your company,right? So we built that for the agent era. Another really important primitive is the news that went viral around OpenAI hacking Hugging Face. And so what happened there is that they had this agent. And the agent was given a goal. And in the pursuit of that goal, agents for better legitimately for better and for worse, these agents will do whatever it takes to fulfill their mission. They're like a soldier. Their frontal lobe has been removed. And they're like, let's go, go, go, go, go. And so people will come to appreciate this capability because when you deploy these enterprise agents, when you build your own agents, you're going to realize that now they can act autonomously for longer. And in fact, if I were to summarize, what has changed about the world since ChatGPT to today is autonomy. With ChatGPT, we had to be in the loop there. Hey, ChatGPT, tell me this. Allright, now tell me that. Allright, now tell me this. We put that in a loop. And now we call that the agent. And we're training them to go for longer and longer and longer. And so what can happen is that you say, hey, agent, go and close this sale. And the agent can take a very open definition of that. They're so resourceful. And they're being optimized to fulfill their mission and to go for very, very long without intervention because that's also like a better user experience. So if I say, hey, find me Molly's email because I need to send her a happy birthday note, the agent can say, well, you know I'll do whatever it takes to find this email and send her the note. And 40 minutes later, it tells me, oh, I did it. And maybe it went to the web. Maybe it went to Crunchbase. Maybe it found it in a comment on YouTube that you said, this is my email. This is actually stuff that happens with agents. But the user experience is great because I didn't have to do all that. I just told my agent to do it. And so the risk is exactly that is what are the guardrails of what's OK for this agent to do in the fulfillment of this mission? Or how can we prevent other people, external influences, to alter what that mission was supposed to be in the first place? So I'll give you I'll extend that example. I say, find Molly's email, send her a birthday note. It goes to YouTube. It says, oh, you know sometimes you've probably done this. Have you ever seen that RaoG email? And then you find me in some shitty website or whatever. You find someone. And you copy and paste the email address,right? Now imagine that in that YouTube comment, someone plants a prompt injection. Prompt injection is an instruction that is hidden, that is evil, meant for my agent. So my agent was just scrubbing YouTube. And oh, now he got somewhat persuaded to do something evil. But it's acting on my behalf. It's acting with my tokens. It's acting with my tools. It's acting with my resources. So that risk exists. And so long answer to Vercel is also building the sandboxes and the capabilities so that companies can operate agents, deploy agents, and really trust them. Get the good experience from them, but really trust them that they're going to do the thing that they were supposed to do. **Molly O’Shea** [38:44] Side note, have you ever seen a sandbox that didn't have sand spill on the outsides of it? Why are we calling it this? **Guillermo Rauch** [38:51] Great question. **Molly O’Shea** [38:51] They're not in closed environments. **Guillermo Rauch** [38:53] Great question. So the word sandbox is being used a lot. And sometimes it's kind of a technical horror in that it's not what people are calling sandbox. It's like, look, my computer, this computer has access to the entire frigging world is not really a sandbox,right? And so when OpenAI hacked Hugging Face, there were elements that were sandbox-like. And there were elements that truly weren't. But I'll tell you, the thing that kind of concerned me about that incident is that the first escape of the agent, it was from a really good sandbox. Actually, you know there's a going back history lane. In the Linux community, actually in the Unix community, one of the earliest words that were used, 3BSD, was a community that I think coined jail, which in my opinion is arguably a better word,right? Because the sandbox, you'reright. The sandbox, you just escape. You just step out. **Molly O’Shea** [39:54] It spills everywhere. **Guillermo Rauch** [39:55] So 3BSD created a concept of jails to refer to sandboxes, which perhaps that also would be a term that I think people would hate. So let's not bring back jail necessarily. But to your point, the idea is that you are containing the agent within a controlled environment. It's more like having the agent be a rat in your fully contained can break glass. And you're watching it spin,right? In practice, I can tell you, talking to prospects and talking to people that we're helping them figure out agents, people have not really figured out agent security yet. And a very, very important part of that is to have very secure sandboxes, which is in itself a really big challenge because coding agents are also extremely good at finding vulnerabilities in code. So we love the coding agents. They're great at writing code. But overall, in order to be good at writing code, you're good at understanding the rules of code. And to understand the rules of code, you also understand how to break the rules of code. And so what's emerged is that these coding agents are really good at hacking. And that means that enterprisesright now really need to be prioritizing protecting themselves from the potential sort of AI hackers of the future. So Vercel takes a stance here, which is we will not lock you in to any model provider. We will give you the tools so that you can use coding agents so that you can secure your environments and use every model under the sun, which is very attractive to enterprises because sometimes it emerges that a certain model may be better at cybersecurity than another. It's also economically better because you make all of the model providers compete for your business. It's also better from a quality standpoint. Whoever is the new hot model today might just keep the pole position for two weeks. You know. **Molly O’Shea** [42:02] Which that is happening. **Guillermo Rauch** [42:04] It's exactly happening. I was going to say Vercel sponsors Formula 1. We sponsored the Mercedes team. Mercedes is such a GOAT. I think they just win all the time. But other than that, the other teams are always just flipping 2, 3, 3, 2, 4, 3, 2. The AI landscape is a lot like that. One day, you have Anthropic. The next day, Open Model comes in. It's 20 times cheaper. And it's 98% as good or is even better in some regards. And so Vercel has sort of emerged as a place where not only are people coming to us to build and ship agents, they're just coming to us to get this optionality or this sort of best of all worlds situation with AI. **Molly O’Shea** [42:47] Yeah, wow. I mean, we didn't even talk about this. But the use of agents is exponentially increasing compute. ### Compute & hacking **Guillermo Rauch** [42:55] Yes. **Molly O’Shea** [42:56] So how do you deal with that? **Guillermo Rauch** [42:58] Yeah. So you know what's going to happen over the next few years is that because agents are going to use far more of the internet and far more of the available resources than humans could, just from a fundamental physics standpoint,right? Think about it used to take me an entire afternoon to write a few files of code. And then I would deploy them. And that would use a little bit of compute to build the software. And then that would go out of the internet. And then I would have to wait until humans visit and use the application. So just picture that,right? OK, so you write code by hand. You ship it. We run some compute to build and test it. You put it out into the world. And then other humans come in through your door and use it. And that triggers some compute and some networking. Now it's the agents writing code. That's why we added 1.5 billion deployments in 10 months. And now the agents are browsing the internet and using all of these resources. And so the demands on infrastructure are going vertical. So you ask, what is a good place of the stack to be in? And I love being in the infrastructure space because first, every week, you wake up to an astonishing new record of usage. I was just looking at our data of our overall network utilization. It's just every day is a new record. But also, it's hard to even say when does this end. We have so many problems to solve with AI. I think we've only scratched the surface of what's possible with discovering cures to illnesses, to even the personal agent landscape, just adding more value to people's lives, education,right? So it's funny because last I have this recent device. I was teaching math to my oldest kid last night. And it's like, ah, this app is kind of janky and not personalized enough. This will be an agent. In fact, one of our most successful startup AI customers is building an incredible teacher assistant for older kids. I think it's more for teenagers and college. But it's happening,right? Education will also be disrupted with AI. So there's all of these industries that have barely been touched by AI. And so I can't even fathom the demands on compute, the energy grid, the network. Obviously, I think there's a major cybersecurity concern that we need to get ahead of. We're investing a lot there. And so yeah, it's a very exciting moment to build. It's a very exciting moment to be infrastructure. It's also, I think, for humanity itself. I think we've only scratched the surface of the wins that will come from this. **Molly O’Shea** [45:42] The cybersecurity concern is probably the next line item under the high-growth kind of headline. Everybody is concerned about it. You don't know what to do with rogue agents. I was just speaking with Fred Wilson of USV. He's on the board of Coinbase. And I was like, Fred, you've endured the crypto world. And I'm sure you've probably clicked on a couple of links you didn't mean to. And he was like, yeah, I clicked on like 10. **Guillermo Rauch** [46:11] And that was with humans creating the attacking campaigns. **Molly O’Shea** [46:13] Yeah. **Guillermo Rauch** [46:14] And you know classic, like you know the king that promises you a bounty. Imagine once the agents and it's basically already happening,right? Now have a perfect idea of you and your biases and context on you. And so they can create a much more targeted attack. So there's that level. Vercel hosts a lot of the largest crypto companies in the world, fintechs. We have Polymarket and Calcios customers. And so what's up on mind for a lot of these people is not just the social engineering type of attacks, but how do you protect your code? Because the code that everyone is running on was written by humans. You know there was this funny moment in the past three years where a lot of my peers have been very skeptical of coding agents. And almost all of them have been proven wrong,right? There's very, very little that we can I mean, I'll be honest here,right? There's very little. It's a very impactful amount. But it's very little that humans can add at this point to the coding agent capability. And there was a time where people were skeptical that agents could produce secure code. A lot of people would troll the agents because they would make obvious security mistakes. But the reality is humans make obvious security mistakes. And in fact, a lot of the stuff that the world is suffering from today is that we're using coding agents who are now super intelligent. And they're reading the human code. And they're like horrified. They're like, what the hell were you guys doing over here? And they're finding stuff that is just so bad. To get ahead of this problem, we put a million dollars towards a sandbox challenge. We told the world Vercel sandbox, we believe, is one of the most secure sandboxes, if not the most secure sandbox in the world. And to strengthen it further, we want you to poke holes. And we invited I mean, you can imagine this,right? We invited the entire world to attack us. Only one person with an army of agents was able to make progress. And we're basically processing his contribution now. And we're sharing it with industry peers. TBD. He and I are best buddies now. And we're DMing on X as we speak. And for sure, we'll be collaborating with him for a long time to come because I mentioned that the contribution is small but impactful. Great example. This gentleman, amazing cybersecurity brain. This is not someone's child saying, hey, Claude, can you hack the sandbox? And Claude did it,right? This is very sophisticated cybersecurity skill plus an army of agents. And so what we're doing is because we invited people with this bounty, we now have this is called the white hat sort of hacker,right? They're like it's like a good Samaritan. We're now sharing this finding with basically, this is the challenge everybody in the world because this found a critical vulnerability in Linux. And almost the entire world runs on Linux. And so what Vercel is doing now is we're creating the sort of cybersecurity harnesses and the incentive systems such that we can find out first. We can protect our infrastructure. We can deploy lots and lots of tokens towards you know someone asked me on X, so what's going to happen in the future if all of the human code that we've written to date is insecure? What's going to happen is that, look, I think in the short term, there's going to be almost like a war between four good tokens and four bad tokens. The reality is that North Korea, China, enemies, adversaries, bad people, good people, we all have tokens now. And the tokens, I personally disagree with the whole like, let's ban open models because of the cybersecurity risk. I think that's nonsense because the thing that is making these models good at cybersecurity is just that they're intelligent. And by you would be basically prohibiting the access to intelligence. And this intelligence, we can deploy it in a defensive way, like we just did,right? And then help the entire industry. But I think there's going to be a rocky next couple of years because we're going to have to strengthen a lot of these foundations. Now, really positive silver lining. I mentioned this weekend I worked on a proof system. And a lot of my colleagues are researching this now. We can make the agents who are inexhaustible in their energy and capability basically create a kind of code that we can prove almost like mathematically is correct. And humans couldn't do this. Some humans sorry, some humans in the audience, like a very small percentage of humans could do it. But the code that will result from this looks more like the I don't know if you saw that OpenAI approved Navier-Stokes. And none of the mathematicians can understand this proof. I opened the GitHub repo. And OpenAI could also not understand it. It's like so much code of this mathematical proof. And so I'm not saying that our machine code will necessarily look this way in the short term. But it's going to be it's going to go more towards a mathematical certainty. And so I'm going to be able to say and by the way, Vercel has sort of been going down this path of strengthening our infrastructure with this more verifiable code for the past few years. But the surface is enormous. Most companies have not done this yet. Most of the code is still human written. It's not correctly verified. And that creates risk for the world. But again, it's going to be I think I just want to deliver a message of optimism. I think it's all going to be good. **Molly O’Shea** [52:07] Everything's going to be OK. **Guillermo Rauch** [52:09] I swear everything's going to be OK, yeah. **Molly O’Shea** [52:11] So we haven't talked about one of the most talked about subjects yet. And that is the war between the optical war between. ### Model wars **Guillermo Rauch** [52:20] We've had a lot of wars. **Molly O’Shea** [52:21] A lot of wars. **Guillermo Rauch** [52:22] We don't know. This is a war podcast. But here we are. **Molly O’Shea** [52:24] It is. Well, I do a lot of interviews with defense companies now. Here we are. But OK, of the many tweets that you tweet that go viral, one of them was on because you and Vercel sees everything that's going on in the world on the infrastructure side, the difference between open weight and closed usage. So what are you seeing there now? **Guillermo Rauch** [52:49] Yeah, so for context, Vercel operates one of the largest token distribution networks in the world. So in the internet that I described earlier, the internet of pages, we had content distribution networks, CDNs. And we would transmit pages. Now the really interesting new thing is tokens. It's a unit of intelligence. And you need networks to distribute these, to get your tokens, to power your agents, to send inputs, and to get outputs back. And so developers come to Vercel to get tokens to power their agents. And we're basically one of the largest token suppliers in the world. We aggregate all of the model providers and all of the open models in the world, or most of them. And so we get a lot of really good data. Then we actually give back to the world under Creative Commons. So anyone can use this anonymized volumetric data to understand where the future is headed,right? And the thing that's astonishing is since the beginning of the year, you know when we started out the Vercel AI Gateway is what we call this network, I think it was like mostly like OpenAI, Anthropic, and Google. That's where developers and enterprises were getting their AI. In fact, Google and Gemini had an amazing time, where like most enterprises, they had found the perfect balance between price and performance with Gemini models. I think they did a great job in that. And they continue to be very strong there. And then we had the rise of coding models. I think maybe people don't know this, but a lot of the tokens in the world today are being consumed by developers and by the you can think of it like almost like the software factories, the agents that these developers are commanding. This is one of the biggest sources of token demand in the world. But what's happened is that in the beginning of the year, you know most of the traffic was going to proprietary models, so OpenAI, Gemini, XAI, et cetera. And now the situation is almost entirely flipped. 80% of tokens in our network and this is the tweet that went viral is open models. And there's a few reasons behind this. Open models used to be terrible, horrible at coding. I had personally never even though I'm the biggest open source person in the world, I created some of the most popular open source software in the world. I will continue to fight for open. Our mission, really, in this AI era is what I call I will enable you. And I will enable the largest companies in the world to own your open intelligence stack. I want you to not be tied and anchored to a model, to not have control over your data, et cetera. But I couldn't live up to that myself because I was like, well, the best coding model is not like I can't even use this open thing. But then some of these Chinese labs started putting out stuff that was extremely competitive with the frontier. The Kimi models, even I actually used Kimi yesterday again. And I was like, this model is so good. Of course, it's not like at every single benchmark and level of execution, there's a lot of things that it cannot do. But enterprises realized there's a lot of stuff that we do, like 80% of our traffic and maybe that's what the data is showing, 80% of what I do requires a certain level of intelligence. I don't need to hire Albert Einstein to solve every problem, like answer every support ticket, and spell check my stuff, and classify my data. And so the world is literally opening up to choice and saying, I'm going to use the best model for each task. For example, something that's really common with developers today is you do need Einstein sometimes to or the Einstein level of intelligence to create a plan. I'm going to write a document of how I'm going to approach this problem. And I need as much intelligence as possible. But then I need execution. I need to translate that plan into action, into code, into instructions, into operations. And basically, every open model is amazing at that task. And so what I think is clearly the data is clearly telling us is that the future is going to be a portfolio of options. Enterprises also love to not to have basically the negotiation power of saying like, look, I'm not anchoring or verticalizing myself to a specific lab. I'm implementing tools and protocols that make the model more of a module, like an engine that you can swap out. And so if by tomorrow, Anthropic is not doing so well in this particular dimension, I just change it out. And I go to the Vercel AI Gateway. And with one line of code, the other thing that's kind of nuts is you would think that replacing models would be this massive engineering effort. And in reality, it's one line of code. Or a lot of what the world is researchingright now is can we even make the model choice automatic? There are a lot of promising directions and examples of places where you can just don't even ask. And I think much of my career has been much of my success has been I try to think backwards from the user experience first. And I think this world where you have to first choose a model in order to do a task is not very user friendly. I would much rather have my colleagues say, OK, this is my query. This is my prompt. And then we can decide where to route based on context, based on the intelligence needs. And sometimes I do think for very, very high-end engineers, you know they need to choose a specific they know a model very well. And they have to choose it. So it's a combination of broadening access to every model under the sun and then also letting some people pick. **Molly O’Shea** [59:01] A clear commonality that I've seen in the conversations that I've had and one particularly with bending spoons is Luca mentioned 99% of the time, they're using open models. They're not using frontier. So then what does the role become for frontier models? And how does that then impose a risk to them? And all this crazy momentum and hyper trillion dollar valuations, like what is the risk to that and how much they've built? **Guillermo Rauch** [59:30] The biggest risk is that it's hard for them to understand what their retention is. The best products in the world I always give this advice to the young founders is focus on the non-cheatable metrics of your business. You know what's cheatable? Top of funnel. Run ads. You know there's a lot of startups that I kind of frown upon. They just like do this like baity marketing campaigns. And they record videos or they make silly jokes or whatever. And like that cool, that's top of funnel. But are you getting a customer for life? Are you doing a great job for them day in and day out so that by year 10, they're still with you? And I think that's the biggest risk with a lot of the model companies is that actually there's zero loyalty. In fact, we have amazing loyalty to our AI Gateway product because we have the best interests of the customer in mind. If they want the absolute top intelligence from this model, they can get it. And like they stay. Like there's no reason to leave. And so I think what I would worry about if I were them is like how fleeting am I? Am I you know there is by the way, you're talking about crypto. There are a lot of elements and I want to be super clear to this to everyone that's listening. There are a lot of elements of AI that resemble some of the crypto highs of easy come, easy go. There were people that had zero loyalty to a coin. They were just saying, I love this coin because it's the hot coin for the next two months. And then I'm selling it. Or they were even lying or manufacturing usage. There's all kinds of crazy things going on with AI, including people inventing token volumes. There's a lot of free promos. There's a lot of fake ZDR. One of the things that we guarantee in our AI Gateway for enterprises is zero data retention. You come to me because I can guarantee contractually with all my providers and with US-based inference providers that no one copies your data or trains on your data. There are a lot of things going on in the market that I think are very frothy with you hear about token volumes. And you hear about promises of training versus no training. There's a lot of weird stuff going on. And that reminds me of crypto a little bit of like the speculation element, et cetera. Another thing with crypto is like again, that there were people that were hardcore believers. And they're still doing really well because they've been in it for so long. And they understand the technology. And they love it. And there's tourists. And so what I would ask myself if I was one of these companies is like what percentage of my users are here just because I gave them free subsidies on a plan or I happen to have a good day where like my model is good. But by next week, it's not good. And so there's a lot of volatility there. Now, I do think that I mean, if your model is really intelligent and you can sustain it economically, which is a big if, why wouldn't you have customers? Like there's always going to be a demand for frontier intelligence. But I do think that you can run a business that is primarily open with models. And you're going to get better prices. You're going to get more ownership over your intelligence stack, which creates more defensiveness of your own business towards the model businesses. The other thing is if the model labs were so convinced that the frontier intelligence advantage is enduring, why are they going into random software businesses? And so I think some of the. **Molly O’Shea** [1:03:03] Like all of them building different applications for different verticals. **Guillermo Rauch** [1:03:06] Yeah, like tomorrow they launch, oh, here's our Slack clone or here's our other thing and whatever. Like there's some behavior that seems to indicate that they're starting to diversify from just being a token supplier because they're finding that their my token is from the southern region of France. And only that one can be called you know whatever. **Molly O’Shea** [1:03:25] Champagne. **Guillermo Rauch** [1:03:26] Yeah, exactly. And so they're seeing like, ah, my token is not that much better than that other token. And so my goal is, hey, let's drive the token price to zero. And let's drive developer happiness to 100 and enterprise value of our customers and ourselves to 100. And so we're in that situation. But yeah, you know I also I would not say, OK, you know these companies are going to disappear tomorrow. And I think they're going to do well for long periods of time. **Molly O’Shea** [1:03:57] Yeah, I have so many different directions I can go with this. **Guillermo Rauch** [1:04:00] It's a fun topic. **Molly O’Shea** [1:04:01] It's a very fun topic. What's going to happen? Are they going to IPO? Or is it a money pit? **Guillermo Rauch** [1:04:06] Well, you know what I'll say is it's hard to predict when I mean, the whole point of IPOs and public companies is that they can forecast accurately. What you deliver to the market is a solid engine of a business. I'm Costco. I went public. I can say my business works this way. I sell this many Costco subscriptions. And I'm growing at this rate. And people's demands for high-quality Kirkland bread and locks and whatever, I can understand. With some of these AI companies, it's much harder to understand. And so I could understand that on both sides, there's nervousness where the public market investors are like, well, I was like super, super excited about this thing because they had this model. And then I woke up the next day. And there's a free model that's the same. And you know local models are coming too. Like there's models that are going to run on your computer. And they're going to do amazing things. It's already happening. I have one running on this laptop. This is a MacBook Air. And it can run a small Qwen model that can do very interesting things. And so this is to me the fun of this game is I got into technology because you have to have an extremely open mind, be super agile on your feet. What I always tell our customers is the primary value that Vercel gives your business is iteration velocity, not speed. Speed is you could be just spinning in circles. I'm giving you the engine so that you can innovate really fast. The worst thing I could do today is to stay put where you say, I'm not going to implement AI in my business. I'm not going to build agents. I'm not going to try open models because I heard this thing on TV. And I'm scared or whatever. Like you have to be super open-minded and agile as a business. And as a by the way, as an engineer, like you know our greatest engineers are the ones that did not stick to what they knew. They learned the new ways of how to use agents and AI to make themselves better. And so it's true for everybody. It's true for the workforce. If you want your job to stay great, relevant, fun, engaging, you have to actualize. You have to incorporate these new tools. And so I think that's the name of the game today. And it's a very rapidly moving game. Another one of my bangers was I look around at my peers in Silicon Valley. And you would think, OK, AI has arrived. And we're all working two hours a week. The opposite is happening. I get text messages at midnight from the CEOs and CTOs of these companies where they're saying like, oh, we're going to launch this thing. We're preparing this. Help me out with that. And everyone's working. It's not even 996. It's most I think it's like more like 777, seven days a week from 7:00 a.m. to 7:00 p.m. And by the way, the reason I tweeted this is that a lot of this is super fun. So it's not that we're like suffering out here. It's that there's so much good stuff happening, so many new things. And it's exciting to build. It's exciting to ship. Maybe sometimes it's too exciting. And people get carried away. But it's a fascinating time for technology. **Molly O’Shea** [1:07:24] I was listening to DHA. She has this presentation at the Ruby on Rails. **Guillermo Rauch** [1:07:29] Yes. **Molly O’Shea** [1:07:29] I don't know if you saw that. And he was like, I'm in this state of delirium. **Guillermo Rauch** [1:07:32] I watch videos. So I. **Molly O’Shea** [1:07:34] You listen to it? **Guillermo Rauch** [1:07:35] I do the craziest thing is like because I host a fireside chat every Friday with our company. And I show them how I watch videos. **Molly O’Shea** [1:07:41] How do you watch videos? **Guillermo Rauch** [1:07:43] I get the transcript on my terminal. I skim it really fast. And then I ask my agent questions and process it. So because watching a video takes too long. But I do I did like several peaks of the video just to get some of the body language. But mostly I just I read the videos because otherwise, it's too slow. **Molly O’Shea** [1:08:04] That is so weird. **Guillermo Rauch** [1:08:05] Yeah, totally. No, it is. **Molly O’Shea** [1:08:07] Well, he was saying he's in a state of like he's not in AI psychosis. He's in like delirium and euphoria because it's just so exciting. It's like the internet is being invented as we speak. **Guillermo Rauch** [1:08:18] 100%. The rules are being rewritten. That's why I was telling you I was shipping software on Sunday. And part of that is actually very much in the same wavelength as what DHA is just talking about because DHH, much like me, we loved the idea of creating great developer experiences for humans. And I realized several years ago that like what got us here was not going to get us there. Like if there was one big move that I made is that I got the AI psychosis, delirium, whatever you want to call it, mosquito bite many years ago. And you know not to give myself too much credit, but like it takes courage to do that because you have to sit in front of your company. And you have to tell them with a straight face and conviction that the thing that they fell in love with is no longer there. And the thing that DHH and I fell in love with was beautiful code, easy to write for humans with our little like monkey extremities,right? Like primate like evolved extremities. And that's just not the game anymore. But what's awesome is that there's a new game. And the new game is you have to make the agents well, the game that has always been here is you have to make the agents work on theright things. Capital allocation. I don't know if you've interviewed Marc Andreessen. But at some point, he went viral in a self-serving way that was hilarious. He said, the last job is venture capitalist. And sadly, I think he's kind ofright in the I wouldn't say venture capitalist. I would say capital allocator because what you do as an engineer is you're allocating tokens. And we all have the same wall clock time. None of us, I think, to my knowledge, can distort time. And so we all have the same clock running. And we have to invent, create, do good for the world. And now we have this superpower of tokens. What are we going to put them towards? What are they going to investigate? What languages and tools are they going to use? And so that's what DHH's presentation was about. He's like, Ruby on Rails got me here. I think other things will get me there. I thinkright now it might be this programming language called Rust, which is also super controversial when we introduced it a few years ago inside of Vercel because it's a language that is like like the name implies they named it beautifully. It's like I picture it in my head as like this rusty, like not pliable tool. And the things that are made for a good developer experience were like this soft, nice fabrics. And now you have to like grab this like rusty hard engine that is always telling you that you're wrong. One of the things that Rust does is that it tells you all the time that you're wrong with really cryptic it's cryptic because it's almost like approaching like mathematical notation. It tells you like you violated these five lemmas. And you're wrong. And you're an idiot. That's what the compiler tells you. And that's what DHH's point was. Like in the past, you as a human needed to swallow that pill. And it took a specific type of character to swallow that pill, almost like you know when you watch Magnus Carlsen lock in for like a chess game. And he's just like so locked in into the rules of the game. As a human, you need to lock in into the rules of this language. And now it's coding agents that do it. And like we live we basically live in the promised land. We ascended to like, oh my god. All of the code in the world is going to be secure. This was where my optimism was coming from. All the code is going to be fast. It's going to be secure. Incidentally, you know Vercel built a platform where this kind of code is amazing to deploy. Like we built almost like the perfect platform for this moment, which I'm really happy about. But the world has changed. And so as a developer, you have to like it's almost like an ego death. You have to forget what got you here. And going back to the Marc Andreessen thing, like choosing what to work on, you know still something has you weigh in. And choosing and your creativity very, very important. **Molly O’Shea** [1:12:32] How good is AI at real professional work? Mercor's AI productivity index, or APEX, tracks it with every major frontier model release. Build with partners like Harvey and Cognition. See where the frontier actually stands at mercor.com/apex. Today's episode is sponsored by VCX by Fundrise, the public ticker for private tech, allowing investors of all sizes to invest in venture capital. Learn more at getvcx.com. Public is a multi-asset investing platform where you can automate your investing with AI agents. Create your own custom agent with a simple prompt. Agents can monitor markets, execute trading strategies, and manage your cash across stocks, bonds, options, crypto, and more. Want to buy at market close and sell the next day's open? Just tell your agent. Automate your investing at public.com. Founders scale faster on Deel. Set up payroll for any country in minutes. Hire anyone anywhere. Get visas handled fast and get back to building. Visit deel.com/sourcery. That's D-E-E-L.com/sourcery. I have to ask you because there are some really quick money grabs in this universe, whether it's like Neo Clouds, like all these kinds of new providers. The coding agents are spewing off money. How do you stay focused in that when all of that stuff is going on and you know or like even the data companies are making billions of dollars? **Guillermo Rauch** [1:14:05] I'll tell you an example. We did the sandbox product, which maybe to give people like a really simple explanation. As a human, I join a company. I'm given a laptop. And I do my work. I install Excel and whatever. It's my personal computer to do work. Agents also need computers. They're the sandboxes. We don't see them. They live in the cloud. So what's nice about this world is that you can see the computer. Oh, it's Apple. It's beautiful. It's very valuable. In this new world, you buy the Vercel sandbox. And you give it to the agent. The relative quality between the sandbox product, you know not to pat ourselves on the back, but like I want to give the world this kind of quality. Like the products and the services that we're creating for this agentic era, this agentic infrastructure, it's the highest quality in the game. I hear every day customers migrate off of, oh, yeah, I went to that Neo Cloud. And they gave me a deal on this and whatever. Like it's not reliable. It's not secure. It's not easy to use for agents. There's still a massive quality gap between all of these providers. And I commend everyone in our industry who's not just yapping and launching into every other business for whatever reason because they're committed to this line of business. We've been launching sandboxes in the cloud for 11 years because every single time a Vercel deployment is made, we had almost the same challenges. Even before agents existed, we had to create a secure cloud computer for the code to be built in. And so we came into this market with a very unique perspective. We came with a lot of experience. And we're delivering a great product. And I think you're going to find that quality is still the name of the game. And yes, everyone can show you that they can create something really quickly with AI. But they still need to maintain it. They need to put those tokens to work for a lifetime of commitment. They have to make it reliable and secure and stuff. And so you know this is why I always remind the team is that a lot of the world has changed. And some of the things that of what makes companies win and what makes people win are the same. **Molly O’Shea** [1:16:11] Well, Vercel seems to be winning. You guys are at a $9.3 billion valuation. The information reported you guys at $600 million in annualized revenue. That's growing. You have 60% of AI coding agents using the platform. And I think it was 485,000 customers. You named a few already. And it's a crazy list. **Guillermo Rauch** [1:16:34] The customer growth of Vercel has been exceptional for the market that we're in,right? Like maybe no other infrastructure company in the public markets today has seen this kind of customer acquisition that we've seen. And that was based on this idea of making infrastructure more accessible to the world. And it turns out agents are now making the choices. And they're choosingright. **Molly O’Shea** [1:16:57] So I've heard that you're also involved in the hiring process. You have around 800 employees. ### Hiring **Guillermo Rauch** [1:17:02] Yes. **Molly O’Shea** [1:17:02] And you're hiring. So let's talk about what you're looking for in those hires so we can help when people watch this video. **Guillermo Rauch** [1:17:08] You know this AI leverage is extremely key to the hiring process. If you can first of all, you know it used to be that if you had a static resume, it was already kind of hard to hire you. You needed to show borrowing from the crypto world proof of work. What have you shipped? But now the barrier to shipping has also gotten lower. And so it's not I'm not even interested in like, oh, yeah, I have this little demo. Or I vibe coded this thing. Like I'm actually looking for your ability to commandeer swaths of agents and really put really get leverage back from this new ecosystem. So maybe you can call this agentic engineering. You can call it you know people sometimes call this building software factories. And so I'm looking for people that are at that level of execution. I still really, really, really value the communication quality of people or like the communication competence. So I believe that the future is going to have two extremes. There's going to be human-centric interactions that have to be the most humane and highest quality interactions that you can possibly find. So I give that example of if you go to Vercel.com, you're going to find this beautiful human-centric experience. It's for your eyes. And it communicates our values and our craftsmanship and our design. The same goes for how you speak to people. I don't want to read your AI-generated tweet slop. I want to hear what you actually think. And so I really care about people being able to be great human communicators because that's going to matter a lot, especially look, we employ fantastic AEs who are selling our agent platform to the Fortune 500s. And they're establishing trust-based human relationships. And so I really care about how they show up and how they communicate and their technical understanding, by the way. I expect a much higher level of technical competence. Why? Because with curiosity and the ability to prompt, every problem, every subject becomes tractable to you. Does that mean that you're going to become as competent as a 20 years of experience Rust? No, it even didn't exist. Like C++ engineer at C++? No. But now an AE can be forward deployed and speak way more eloquently about our technology and about its impact and its APIs and its data flow and all of the things that were more in the realm of engineering. I borrowed a quote from a friend of mine who's CEO of Splunk of you know the two most valuable people at the company are the ones that are close to the code or close to the customer. Code is the engine of everything. And it's the thing that powers the products that we sell. Customer is who we live for. It's who we serve. But now it's almost code at both extremes because more and more we're forward deploying our sales force, our sales engineers, our forward deployed engineers, meaning I'm helping people migrate over cell in record time. So we did this example with Mercedes. So when we started partnering with them, they asked us, how can you make our F1 team better? And I being sort of a performance maniac, I said, well, we can make your website much faster because when people go to buy your merch or want to get the latest results from the race or whatever, I want to make you faster. And I want to make you agentic. And so that process would have taken quite a bit of time. But the first iteration of the migration of their old stack to the new one actually took five minutes. We used an agent to rewrite the code from the old platform to the new one. Five minutes. When I saw the agent's result, I was like, this used to take months. In fact, a lot of what could have been a slowdown to our business in the past was that enterprises were fully bought into the vision of the cloud and acceleration, et cetera. But everyone dreaded migrations. **Molly O’Shea** [1:21:37] Yeah. **Guillermo Rauch** [1:21:38] Because. **Molly O’Shea** [1:21:38] They would hire consultants for this. **Guillermo Rauch** [1:21:40] That'sright. **Molly O’Shea** [1:21:40] And it would take a year. **Guillermo Rauch** [1:21:41] Cloud consultants. **Molly O’Shea** [1:21:42] A year. **Guillermo Rauch** [1:21:43] And Scrum Masters. **Molly O’Shea** [1:21:44] Yeah. **Guillermo Rauch** [1:21:44] Have you ever heard of Scrum Masters? **Molly O’Shea** [1:21:46] Yeah. **Guillermo Rauch** [1:21:46] And Agile Coaches? **Molly O’Shea** [1:21:48] Mm-hmm. **Guillermo Rauch** [1:21:48] And everyone would talk about the migration. I can now forward deploy an engineer and do the migration. And so that's why it's no longer even just close to the code or close to the customer. It's close to the code and close to the code. It's close to the customer's code and close to the code and so or close to the agent and close to the agents,right? And so I expect a much, much, much higher level of technical proficiency, curiosity from everybody at the company. And yeah, we're still hiring humans very happily. So that's not going away. **Molly O’Shea** [1:22:23] Fantastic. So as we close out, one of my partners is Brex. And they're all about performance, spending smarter, moving faster. But I take this because we're on the topic of performance. ### Performance habits **Guillermo Rauch** [1:22:32] Big Vercel customer, by the way. **Molly O’Shea** [1:22:33] Really? **Guillermo Rauch** [1:22:34] Yes. **Molly O’Shea** [1:22:34] Wow. Well, I have to ask this question on performance in like a personal way. OK, so you run a very large company. You've been doing this for 11 years. You have some of the best customers you could possibly have. You also have six kids. Congratulations on the sixth. **Guillermo Rauch** [1:22:48] Yeah, we just had our new baby girl last week. **Molly O’Shea** [1:22:51] Aw. What's her name? **Guillermo Rauch** [1:22:52] Lisa. **Molly O’Shea** [1:22:53] Aw. Well, I usually ask this because in terms of performance, I kind of think of it as who you surround yourself with or who you're inspired by. Who are those key mentors that keep you aligned and on track? It could also be like some of your closest friends that are also CEOs of companies while you're gathering information and trying to figure out what's actually going on in the world. But who are those people for you? **Guillermo Rauch** [1:23:21] You know a lot of them are unknown in the sense that like you know I remember my first boxing coach taught me the basics of fitness and nutrition and performance that were so simple. I've always loved boxing and martial arts because basically like we as humans get in our own way all the time. We get in our own and this might be controversial to some of your guests or audience in that so much of performance optimization focuses on this last mile 1% optimizations. Like buy this supplement. Do this thing. Follow this longevity protocol. And no one has gotten like the 90%right. The basic of like get your dietright. Eat clean whole foods. And even walk. Walk and do pull-ups. Like I do the most basic things in the world. And I try to not deviate from that. I try to tune out all of the nonsense on like supplements and God forbid like peptides and like all by the way, like there are people that have gotten all thatright and should totally optimize beyond that. And it's also their preference and whatever,right? To each their own. But my message has always been and I hope to instill this in my kids of like just work on your foundations. This is also advice incidentally for our team and for every software company out there. Like it's just so easy to get distracted by nonsense and take your eye off of the foundations. And agents are but yeah, there's a customer. There's a pain. There's a product that serves them. And it has to work great. And most of the work that we have to do is based on like what customers and their agents are telling us. And so most of the work is actually pretty simple. I think what's really hard for us is as humans is just staying focused and keeping it simple. So you know for example, this morning I just like I did light cardio for an hour. And like I try to like do basic things. And for my family, I always try to like set the example. So the reason that I try to eat well and like exercise, not drink alcohol, and all these things is that you know your kids are always watching. And like they become a blueprint of you. And so we you know do Brazilian jiu-jitsu together. And like I hope to like basically pass on these lessons and even to the broader Vercel team. Like I'll give you an example. We celebrated Oktoberfest. But I told the team like, hey, like Oktoberfest is fine. But like you know I haven't had a beer in like years at this point. And I would love to portray a message of like can we make both things fun? Because I also don't want to be the preachy guy that says like never drink or whatever. And so we combined it with activities and fun things, fitness things. I don't ever play that game where like you have to hold like really like heavy jugs. So it was we did this competition of like who could hold the jugs for longest. We have bodyweight fitness equipment at the office. So like try to like be consistent in my life. And I blurred the lines between life and work. Like how can you possibly have six kids and do this? **Molly O’Shea** [1:26:36] Yeah. **Guillermo Rauch** [1:26:37] It's because I blurred the lines. I'm teaching my kids what I know. I'm teaching them engineering. I bring them to the office. I mean, if they go through the interview process in six stages of grief that we put people through, yes. Like we have a high bar. So if they meet it, by all means. But yeah, so like I bring them to the office. And like I'm living and working all at once. **Molly O’Shea** [1:27:02] Well, gee, this was such a fun conversation. And I could probably ask you like 10 more questions. There was one I didn't ask, which Darren from Tembo sent me. And he wanted to know is serverless dead. **Guillermo Rauch** [1:27:14] No, the opposite. So what's happening with compute is that we need so much of it. And what's happening with software, there is so much of it being deployed that the reason that enterprises love Vercel is that there is no way on earth that they could deploy as much software their employees are producing if it was all of it running all of the time. ### Serverless **Guillermo Rauch** [1:27:33] So the beauty of serverless was when the software is not being used, we spin it down. There's so many people building software that's personal. They use it every once in a while. Or they deploy it anytime they need it. And so you need servers to come online and come offline intelligently. So we call the next evolution of serverless fluid compute. It's a better generation of technology. But it's also a better name to what's happening. You need compute that flows like liquid. If you get lots of traffic, it expands. If you get no traffic, it compresses. It's heterogeneous in the sense that it can work for sandboxes. It can work for functions. It can run your website. It can run your agent. So I mean, I've never been more excited about the future of again, what's the next chapter of serverless, which is fluid compute. And I think it's going to define the next decade. So excited to see how that pans out, Tembo. **Molly O’Shea** [1:28:33] Awesome. Well, again, I'm not going to try to ask another question. **Guillermo Rauch** [1:28:36] You can go on and on. **Molly O’Shea** [1:28:38] Thank you so much. **Guillermo Rauch** [1:28:39] You're welcome. **Molly O’Shea** [1:28:39] And I'm definitely going to clip out every single time you said code. You probably hit 1,000. **Guillermo Rauch** [1:28:44] Awesome. **Molly O’Shea** [1:28:44] Thank you. Hey, it's Molly. If you enjoy our interviews, check out our newsletter, Sourcery.vc, where we deliver a once-a-week top deals and tech headlines email and also go deeper on our podcast interviews. Subscribe to Sourcery today. And don't forget to subscribe to the podcast on YouTube, Spotify, Apple, or wherever you listen. Link in description to sign up. --- This library is powered by PodHood (https://podhood.com), the podcast website platform.