How OpenCode became the world's most popular open source coding agent

Jay V, OpenCode's CEO, applied to YC nine times and worked on startups for almost 20 years

Since the start of the year, OpenCode — an open source alternative to Claude Code and Codex that works with any model — exploded to 4.6 million weekly active users, 13 million monthly actives, and roughly $40M in annualized revenue.

In this episode of The Lightcone, Harj, Jared, and Diana talk with Jay V, OpenCode’s CEO, about what’s driving this wild growth, the Anthropic clampdown that inadvertently fueled it, and the almost 20 year founder journey that led him here.

Watch on YouTube

Timestamps

00:44 — OpenCode’s Explosive Growth

01:16 — 20x Growth, 13M Users, and 7 Trillion Tokens

03:39 — The Anthropic Controversy That Changed Everything

05:43 — Bringing AI Coding Agents to the World

06:39 — When Open Source Models Became Good Enough

08:56 — What Millions of Developers Are Actually Using

13:31 — Why OpenCode Is Huge Outside the US

15:27 — Why Fortune 500 Companies Choose OpenCode

16:36 — The Economics of AI Tokens

20:02 — How Enterprises Are Using Coding Agents

22:58 — AI’s New Unit Economics

24:56 — Why Model Choice Matters

29:55 — The Product Decisions Behind OpenCode

34:21 — A 16-Year Overnight Success

41:16 — Why Jay Never Gave Up

Transcript

Harj: Welcome back to another episode of The Lightcone. Garry’s out traveling today and will be back next episode. Our guest today is Jay V, founder and CEO of OpenCode, an open source alternative to Claude Code that works with any model you want. OpenCode has been growing at an astounding rate this year. They’re now at 4.6 million weekly active users, which is actually pretty close to Codex’s numbers. Today we’re going to talk about what’s driving their wild growth and also Jay’s winding road to get here since the company went through YC back in 2021. Jay, thanks so much for being here.

Jay: Thank you. Thank you for having me.

Harj: Why don’t we just start with the crazy scale you guys are at. Maybe tell us about any stats you can share with us.

Jay: Yeah. You mentioned the weekly actives. We like to post a lot of our metrics on Twitter. Our monthly actives—I think we ended June with around 13 million or so, and that’s around a 20X increase since the beginning of the year. We also recently started processing around seven trillion tokens per day. For context, OpenRouter does a total of around six trillion. I think at the beginning of the year, we were probably at around 300 billion or so. In terms of our revenue from our subscription product, if you were to pay per token with our inference and you took, let’s say, June’s data and extrapolated for the year, that would be around 31 to 33 million or so. If you took last week’s data and extrapolated for the year, that’s around 38 to almost 40 million. And that’s just the inference part of our business.

So the way we make money—we launched that, call it end of September, early October last year—so about eight months to getting to around 40 million or so. To round out the numbers, our subscribers—these are people that pay for a monthly subscription with OpenCode—we launched that product early March, I think end of February, and that’s grown to around 160,000 monthly subscribers. That accounts for about 18 million of the annualized revenue.

Harj: I saw a tweet from—I think it’s Thibault, I think he’s a lead engineer, at least one of the main engineers on Codex—saying that something like 5% of all Codex subscribers choose OpenCode as the main harness to actually use the underlying API.

Jay: Yeah. So Codex officially supported OpenCode. What that basically means is that you can use Codex’s subscription in OpenCode, and a bunch of their users use OpenCode directly to take advantage of their subscription plan.

Harj: And I think they did that right after some of the back and forth you had with Anthropic and Claude Code. Tell us about what happened and how that seemed like it really fueled your growth—was an inflection point for you guys.

Jay: Yeah. To contextualize this, we were at around 650,000 monthly active users at the beginning of the year. The first week of January, we started to hear some rumblings around Anthropic trying to clamp down on people using OpenCode but using Claude Code subscriptions on there. This was a very common way to use the Claude Code subscription. The way they tried to block it was if the system prompt mentioned literally the word “OpenCode,” they would reject the request. From our perspective, that sort of makes sense—they’re subsidizing usage, that’s what they want to do. But of course, a lot of users weren’t happy. When they did that, what it inadvertently did was put OpenCode and Claude Code on the same pedestal.

It equated the two products in some ways. Even for the people who weren’t using OpenCode at the time, they took notice of the fact that Claude Code was taking that kind of action against OpenCode.

Harj: Oh, so you think you actually got new users because people heard about you for the first time and they’re like, oh, Claude Code’s banning this thing—

Jay: —or clamping—

Harj: —down on this thing.

Jay: Yeah, or that it’s worth looking into, that it’s not just one of the other dozen or so coding agents out there.

Harj: It’s funny how often this happens in startups. Same thing happened with Instacart when Amazon bought Whole Foods. It was like, oh, this is the death of Instacart. And the death of Instacart became this meme, but the meme actually just drove all the grocers to check out what Instacart was. Then they went through this explosive growth of signing up every grocer in America. So it seems like actually a lot of your growth is global across the world. Tell us a bit about that.

Jay: Yeah. So the premise of the product is that most people in the world still haven’t experienced the magic of a coding agent. It’s been almost a year, and I’m sure it’s hard to remember for you guys as well, but the first time you had this experience, it’s a magical experience. When we felt that for ourselves, we recognized how important that moment is in tech history. It comes along once every generation or so. The way we looked at that was, let’s take that to as many people in the world as possible, have them experience something similar, because the frontier models and the frontier labs charge so much per token that it’s going to be hard for a lot of people across the globe to have that experience. We wanted to make sure they had that with us in some ways.

Harj: I guess at a point early on, was there a big gap between the open source models and the frontier models? Was that true when you were first launching the product?

Jay: Yeah, for sure. I think when we first launched, it was mostly, “Hey, you’re using your Claude Code subscription with Claude Code. Come try that with OpenCode.” This was back in June of last year, and by about August or September, we started to see the first crop of open source models. You had this sense that, “Oh, they’re maybe six months behind.” Of course, there’s always been a gap between them, but that was the first point when people were like, “There’s the GLMs of the world, the Kimis of the world, the MiniMaxes of the world, and it seems like there is now an alternative.” As that started to happen, it triggered a wave of users coming in to try out OpenCode, because that was probably the only way you could try out some of these multiple models. As that gap started to shrink, or as the open weight models became good enough for real work, it became viable to use OpenCode with them.

Harj: I think I remember a few months back you were saying that, let me get this right, even though the open source models are obviously cheaper to use through OpenCode, you still saw more usage from the leading frontier model—except was it when KIMI 2.5 came out?

Jay: Yeah.

Harj: That was the first time that it was equivalent or it flipped?

Jay: Yeah, that’s exactly it. I think it was probably 2.4—I forget the exact model—but yeah, this was February of this year where for a four-week span, and this had never happened in our data before. I think we’re fortunate to be able to see this global usage. A lot of times we see the comparison of these open source models versus some of the frontier ones. We noticed for the first time that a bunch of users were just using KIMI way more than they were using the Anthropic models. It was Sonnet plus Opus at the time together. That’s the point when we were like, “Oh, I think we should launch a subscription product because now maybe these models actually do make sense for real work.”

Harj: Speaking of that, you guys have this incredibly unique insight. You have the best data on how these models are being used by engineers across the whole world, and you released a bunch of this data. Maybe we could just look through some of it and pluck out some more interesting insights.

Jay: Yeah. You can go to opencode.ai/data. We started to publish this about a month ago. This is basically taking all of the usage on OpenCode Go, which is our subscription plan where you pay $10 to use any of these open source models. This breakdown specifically is by token volume per day across the different models. What we see here is that DeepSeek Flash is the one that is used a lot. There are some details here—I can go into why that is the case. But if you just look at the top three, we’re seeing the two DeepSeeks plus GLM, and you can see all the hype that GLM has been getting lately being reflected here in these charts.

Diana: Which the data says otherwise, as opposed to all the Twitter chatter about GLM taking over DeepSeek. This is telling a different story.

Jay: Yeah, it is. I can show you a different breakdown here. This is by unique users, and this was a thing I heard you were alluding to. We get to see actual usage data for each of the users, as opposed to maybe an OpenRouter where you’re seeing it aggregated across a bunch of services or other products that are internally using it. In this case, each one of these is an actual user. What’s fascinating is if we look at the market share graph—this is breaking down for each of these labs the amount of token volume they’re doing per day, comparing that across them—you can see the DeepSeek one dip around the time GLM comes out, but it seems to bounce back up after. There are some theories around why that’s the case, but that’s an interesting fact.

And then if I go back up to the users one, this is also kind of fascinating in that it might be a little bit hard to see here, but if you look at the top three unique users per model, you’ve got Flash at 38K, DeepSeek Pro at 31K, and GLM 5.2 at close to 30K as well. That’s interesting because GLM is on par with one of the DeepSeek models, but the fact that there are two of them makes it a little bit different. I’ll caveat this by saying that DeepSeek Flash being as cheap as it is allows a lot of users to extend how much they can use a coding agent, because as they get closer to their daily or weekly limits, they can switch over to one of these very cheap models—in this case, DeepSeek Flash—to get the rest of their work done.

Again, this is very different from the way we think about coding agents and LLMs here in the Valley and SF, and in the West in general.

Harj: What are you seeing? We definitely are at the other extreme end of where it’s like—

Jay: Token maximum max is towing money.

Harj: But for the users you see, which to be fair, it does seem like there’s a general vibe shift, especially in the enterprise world towards more token budgeting. What do you see? Is it as simple as people, once they approach their usage limits, switch over to one of these models, or is there more going on?

Jay: Yeah, there’s a few different things. I think people do try and optimize for things. One of the reasons why originally Kimi had taken off was it was being hosted in a way that the tokens per second was a lot higher than what you would get out of even Opus. So it was just a drastically faster model. It felt like you were working almost in real time. Some of these models tend to exhibit those qualities, which makes it characteristically a little bit different from using some of the frontier ones, for example. So there’s a little bit of that, but cost is obviously a big driver. Another thing that pops up every once in a while is people get a sense that GLM 5.0, for example, is better at front-end design compared to some of the other models. That ends up driving some usage as well.

Jared: You also have a really interesting breakdown by geo to show where the users are. So who are your users? Where are they coming from?

Jay: Yeah, so some context here is that we had launched the OpenCode Go plan to be able to serve the global audience. You can see that with China being number one at 17%.

Jared: That’s crazy. I feel like you’re probably the only YC company in history that has meaningful usage in China.

Jay: It’s also fascinating because a lot of these models are Chinese. In their situation, they’re trying to use the ones that are being built in their country, and OpenCode gives them the choice to do that. That’s interesting as well. I think the US is actually interesting to us because when we built this plan originally, we weren’t thinking about the US. We weren’t thinking about the States because, as we were saying, people here just throw money at it. But it turns out it’s growing really well in the States too. Maybe that speaks to the vibe shift of being a little more conscious with tokens. But there’s also the other side, where people want to use some of these models. So when GLM 512 is getting popular, a lot of people are trying out our subscription plan because it’s one of the options to do that.

Jared: So because it’s so much cheaper, it’s huge in developing countries—like Indonesia is 4% of your traffic, Brazil is 5% of your traffic, places like Vietnam, places where a $200 a month Claude Code subscription is very expensive. That makes sense. But what you were telling me earlier was that in addition to that, there are a lot of large US companies that have basically unlimited budgets for tokens that are also using OpenCode. Can you talk about that—who’s using you and why those people are using you?

Jay: Yeah, this is what we had seen early on before some of these open source models even took off: a lot of companies would start using OpenCode because they didn’t want to be locked into using a specific model or a specific harness. Some users just wanted more choice, and this was a good neutral option for them that allowed them the flexibility in the future to switch to whatever they wanted.

Diana: And I think you had a crazy stat. It was something like a dozen of the top very forward Fortune 500 companies are using you and have a significant footprint.

Jay: Yeah. It’s funny—we get DMs, we get emails internally about, “Hey, XYZ company here. We’ve got a few thousand people using OpenCode now. Please don’t share this publicly.” But I think most of what’s going on is there’s just a bunch of choice there. We can probably talk about this later, but we’re very intentional with our product design. We want something that everybody uses every single day, and we hold that bar fairly high. In those instances, I think that’s what’s resonating with people as opposed to just the cheaper tokens thing.

Diana: I’m very curious on a slightly different topic. There’s been a shift for all these companies like you in terms of the AI token economics. In the old world of companies—other B2C or B2B companies—a lot of CAC used to be based on ads. Now the equation is different; it’s based on tokens to acquire users. But even more weird, there’s this thing you’re describing where there are some whales that pay for a lot of it at some point. There might be a lot of churn, but it doesn’t matter because as long as you have the power users and experts really using it, they convert these large orgs. I think the big labs can subsidize that, which is what effectively Claude Code and Codex have done. They can subsidize it, but you have a magic formula to skip all this.

Jay: The broader context here is that for you to use AI—and especially these coding agents—because of the amount of tokens they use, to use them well, you have to really understand them. This is the experts thing you’re talking about. And to get there is fairly expensive from a per-token perspective.

Diana: Token maxing is expensive.

Jay: Right. It’s very expensive. And that is a chasm that is very hard for a lot of people and a lot of companies to cross. What the Anthropics and the OpenAIs of the world do is subsidize it so that people are able to cross that. The whales idea here is that a certain percentage of them cross it, get to that point, start spending the crazy amounts that you see, and then it makes sense. The entire funnel then makes sense. When we approached this, we thought about it from a product perspective. We talked about wanting everybody in the world to have that aha moment with a coding agent. So that’s our free tier. And then, because the open source models were now cheap enough and good enough for real work, you want a subscription plan that allows them to do real work with it.

And that is the thing that allows somebody to buy into, “Okay, now I can justify spending so much more to potentially redo some of the processes within my company to take advantage of these coding agents.” And that’s when hopefully some of these turn into whales.

Harj: And that’s what you’re seeing. You’re seeing people come in through the free tier to try it out and then become advocates, saying, “Oh yeah, we should adopt this at our Fortune 500 company.”

Jay: Yeah. In the older SaaS enterprise world, you would have this procurement process that a lot of them go through, where someone in a region starts the whole dance. In our case, the inbounds that we get are typically just, “Hey, there are a bunch of people at our company using you guys. Can you fill out the security questionnaire?” And we’re just like, “Oh, first off, I didn’t know we had users there. But secondarily, give us a second.”

Harj: That’s when you really know you have product-market fit—when enterprises are bugging you

Jay: To sign the

Harj: Security

Jay: Agreement

Harj: So they can

Jay: Use your product. Just like, “Please do this because I don’t know who you guys are, but a bunch of us are using this.” So it is a little bit different now.

Harj: Once you’re through the procurement and admin side of things, what are the enterprises asking you for? Are they trying to pull the product in a different direction? Because that often tends to be something that open source companies have to think through and be careful about.

Jay: What’s interesting here is these coding agents are at the core of how an LLM does work. A lot of times when we get these enterprises reaching out to us, it’s because they’re trying to figure out where else they can use it. There’s the obvious one: we’ve got a bunch of developers at our company using OpenCode, just figuring out how we can officially use it. And then on the flip side, there are some non-technical people who would want to be able to use this as well. Secondarily, our product is probably going to be using a coding agent as part of its core loop. Can we use it there? So we get some pull there. The other bit of pull that we see is mostly around just managing tokens a little bit, with people saying, “Hey, there are certain organizations within our company that don’t need the frontier models. Can we limit access there or have some more creative ways of managing token spend?” We see pull there. So there are some questions around that. The strange one we got recently was somebody wanting to have really good visibility of exactly what everybody’s doing at the company. That gets into questions of whether that’s something we want to build.

Harj: Didn’t Ramp use you in an interesting way?

Jay: Yeah, I think Ramp was very forward-thinking. They published a blog post—I think this was in December of last year—but they had reached out prior to that, where a team within Ramp had built this Slack bot that was essentially running OpenCode behind the scenes. It was just incredible to see. It was mind-blowing, partly because we hadn’t even done that yet internally, and they were showing off a use case that was definitely pointing towards the future.

Harj: What does that mean exactly? I think it can be hard for people to get their head around because they think of you as an open source Claude Code. So what does that mean that their Slack bot was running on OpenCode?

Jay: Yeah, so you can think of OpenCode as a two-part product. There is the UI and the application part that you see and interact with, but then there is the agent loop—the thing that’s actually doing work while calling the LLM. That is behind the scenes. We call it the server. That server can be embedded separately from the UI. In this case, they were embedding that and running their Slack bot off of it.

Diana: Can you tell us a bit about these numbers and how the unit economics work? There was a shocking stat mentioned in Dylan Patel’s podcast that Anthropic is profitable and by a huge margin—in Q2 they are on track to be doing $50 billion annualized revenue and around 70% margin. But the previous year was not profitable and they crossed this chasm, which sounds like where you’re heading, but you don’t have to subsidize it, which is special.

Jay: Yeah, we do subsidize a little bit, the subscription plan that we have. And you

Jared: Have a free tier too.

Jay: And we have a free tier. Yeah. I think that’s the CAC part you were talking about early on where—

Diana: Rather than paying for ads, the CAC is paying for tokens.

Jay: Yeah, it’s because that’s how people experience that kind of magic moment. That’s the way we get them into using the product, understanding what a coding agent is. All this stuff is definitely a little too much for a lot of people. I think the part here in terms of the unit economics that works out is if you’re actually paying per token—in the case of Anthropic and in the case of how we operate as well—we’re able to get, at least in our scenario, volume discounts because of the amount of tokens that we serve. So when you pay per token, that effectively turns into our margin. Whereas when you’re subsidizing, obviously you’re eating the cost there, or in the case of the free tier, you’re eating the cost as well. But as you were talking about before, as you start to get more and more of these whales, those whales are paying per token and it’s directly feeding into your margins.

Harj: Yeah. The discounts you’re able to get by aggregating the tokens is interesting because it’s something we’ve talked a lot about over the last year in particular. Everybody in the Valley is at this point, so where’s the value going to accrue? Will it be the frontier models that are going to make all of the money and everything at the app layer is left for dust, or will it go the other way? How have you thought about that? Because you’re in an interesting spot since you actually really do own the relationship with your end user and you’re effectively making it easy for them to pick and choose the models that they want to run with. So what are your thoughts on where this all plays out and how it will hopefully play out in a good way for OpenCode?

Jay: Yeah, it feels a little bit like a marketplace where a user is able to make the choice of the model that they want to use, and different models have different attributes and different cost characteristics. Our claim here is, look, we want to showcase that diversity as well as possible. That also creates an environment where the labs are aware of each other, and that competition ends up being good for the consumer in this case. The flip side is if you’re locked into a specific vendor, then you don’t benefit from the competition that would otherwise come with it, and you are likely helping their margins in some ways.

Harj: Yeah, I feel like your growth is a pretty decent proxy for the fact that choices have only improved over time, right?

Jay: That’s exactly it. I think we could probably track back our bump in monthly active users to some sort of bump in the open source model market. The way we think about this is that it’s not that we’re picking a winner in terms of a model lab—we’re just betting the field. We just think the rest of the field is going to do well.

Diana: The next logical conclusion of this is that models are becoming commoditized utilities.

Jay: Yeah. I think what’s interesting here is that the market is so large that it’s hard to imagine people or model labs not picking off niches and chunks of their own, specializing for specific areas or specific characteristics. In our case, when you see some of the data that we were looking at before, you’ve got the Frontier Labs—we know them well—but when we see the success of DeepSeek, it’s very clear that they have picked the cost part of the quality-cost-performance axes. They’re basically saying, look, we want to be very, very good at that part. It’s hard not to imagine that happening across the board, across all these characteristics.

Harj: Yeah, it feels like if you truly believe the sales pitch that the labs themselves make—which is this is just an unprecedented market, the market for intelligence has not existed before—everybody should be thinking a positive-sum, grow-the-pie mentality. In which case, they should really want you to grow and succeed because you’ll end up just being a huge customer for all of them.

Jay: Yeah. I think that’s sort of true now with these open source models. We’re the largest customer for most of them.

Jared: You’re the largest customer for most of the open source models.

Jay: Yeah. Just in terms of the token volume, the amount that we do.

Diana: Wow.

Jared: So that means that you guys must be locked in this symbiotic relationship now where you both need each other for this machine to work.

Jay: Yeah. The other half for us is the ecosystem. We look at the open source ecosystem as a whole and we’re going, look, we need to make this entire thing work. And again, with all the talk of open source models lately, that’s essentially our pitch.

Jared: Since you’re the largest driver of most of the open source model companies, you must have this incredible insight into the GPU market and where all this inference is actually happening because you’re the ones driving the inference. Are you seeing anything interesting in the GPU market? Where’s all this OpenCode-powered compute happening?

Jay: Yeah, so we rent GPUs. We work with providers that provide just straight inference. We work with the model labs themselves. One thing that we started to notice that was interesting at some point because of our global usage was that the peaks and troughs over the course of a day for GPU utilization weren’t that far off for us. Given the fact that when the East is working, maybe the West is asleep, but when the West is working, the East is… And so as a result, we have a reasonably stable 24-hour GPU cycle, which allows for pretty good utilization and it helps the unit economics for us in running these models a little bit more efficiently. That ends up being a competitive advantage when we think about some of our counterparts that maybe just serve one part of the world.

Harj: Do you think it’s just such an interesting spot to be in? Do you think there are specific product choices or design decisions you made when you were first launching the product that have led to the fact that you’re growing and winning so much?

Jay: Yeah, it’s funny, the name OpenCode literally comes from that. I think we’d done a similar project in the past. It was called OpenNext. The idea was when you’ve got a dominant, or in this case, two dominant players in the market, the rest of the market coalesces around an open alternative, and picking that position ends up being really valuable because if you pick it, it’s very hard for somebody else to displace you. And if it’s open, you should try and become the default as quickly as possible. So when we launched, the name was a very deliberate choice. The fact that we wanted to support, even at the time of launch, we claimed that we supported 70-plus models and providers. Just to make that happen, we had to create a separate open source project called models.dev that built up this entire database that didn’t exist at the time.

Now you can contribute to it, and this is probably the best dataset of all the models and providers out there in the world. But again, just to make that happen and to occupy that position, it was a very deliberate choice.

Harj: Can you think of any other examples of intentional product or design choices you made that you think have really helped you hit this inflection point?

Jay: Yeah, I think the name came a little bit later. It happened within a span of a few weeks. But the first thing that happened was with our past product, we had just hit breakeven. Around that time, this was February or so of 2025, Claude Code comes out. We look at Claude Code and we go, “This is fundamentally different from using autocomplete, using AI in that form. And this is something that actually does make sense for us as a core developer.” We were NeoVIM/VIM users at the time, and so Cursor didn’t necessarily resonate with us as much. But watching Claude Code and using it a little bit, we realized that it wasn’t up to the standard of some of these other terminal UIs like the NeoVIMs of the world. We wanted to build something like that.

That was really the first bit: when you open this up, it should feel like a modern terminal experience to a lot of the core developer audience, the ones that we were going after very early on. It should feel like one of these other things.

Harj: That’s so interesting because I feel like a lot of Claude Code users, because you have such low expectations for what you can get out of a terminal UI, people are like, wow, this is going crazy how much you can get done. And these graphics and effects are really cute and cool and that’s awesome. But the fact that you were terminal UI connoisseurs, it sounds like you were looking at it thinking, oh, you could actually do so much more in the terminal.

Jay: Yeah. And we had built a couple of terminal UIs in the past, one as a part of our core product with SST. The other, Dax had built on the side with a couple of his friends. You could buy coffee online. It’s a terminal UI, it’s a complete storefront. It was a fun pet project, but it showed off what we could actually do.

Jared: Wait, wait. So your product was a terminal UI for buying coffee in case you wanted to buy coffee without leaving your terminal.

Jay: Because the idea was you’re a hardcore developer, you’re in the terminal all day, you can’t be bothered to open a browser. So you go to sshterminal.shop and you order coffee over SSH.

Harj: It makes sense. Way easier.

Jared: I feel like this is an example of that PG essay where if you’re a developer and you build things that you want yourself, even if it seems really goofy to other people and a VC would turn up their nose as like, “Well, that’s a horrible business. What are you talking about, a terminal UI for buying coffee?” It has this tendency to pull you in an interesting direction.

Harj: Yeah. It generalizes—just having eccentric tastes does not always lead to something interesting, but it’s kind of how you get to these outlying things.

Jay: Because we were so embedded in the open source community and because most of the people that surrounded that community were people that looked up to products that were really good in the terminal, we knew that if we built one of those, it would resonate with them almost right away.

Harj: It’s pretty easy to think that you guys have just come out of nowhere and have exploded and are this overnight success story. But the company’s actually been around for a little bit longer than six months. Tell us a bit about that backstory because you went through YC over five years ago now, but even before that.

Jared: Yeah, but the story starts way before even that.

Jay: So it was second year university and I had just done a co-op term at Waterloo and I was like, “Wow, I don’t want to do this.” So I came back being very naive, found the smartest people I could get ahold of. It ends up being Frank, my college roommate, and the two of us were like, “Yeah, let’s just start a company.” Again, I’m a 21-year-old, and so I picked the name Anomaly as the name of the company because I thought, “Ah, I’m special.” But yeah, it literally starts off with me reading. How

Harj: How long ago was this?

Jay: 2006, 2007. I think I started reading PG’s essays because I wanted to start a company. I wanted to do a startup. This was effectively the best thing you could find on it at the time. I think the first YC application was probably around that time, but the first interview and what brought me the first time to Silicon Valley—I forget the exact batch, but I think this was the Mixpanel batch because I know the day that I interviewed was literally after Mixpanel. So I was interviewed right after them. I think this was right after Airbnb’s batch, and one of the Airbnb founders was hanging out in the room when we were waiting to be interviewed. This was obviously with PG back then. That was one, and then I think the first of many interviews and applications to YC. It took more than a decade to get in, let’s just put it that way.

Jared: And you told me something crazy, which is I assumed that during those years, these were different startups, but apparently it was literally the same startup, the same legal entity all of those years.

Jay: Yeah. Yeah.

Diana: So this legal entity is almost 20?

Jay: Yeah. We incorporated in 2010.

Jared: So it’s a 16-year-old legal entity. And you’ve had successful products before, but OpenCode is the most successful. So it took 16 years since you started the company to have a truly runaway success product.

Jay: Yeah. I think part of it is also to do with your level of maturity, your level of ability. We were maybe too young for some of these other waves—the mobile wave, the cloud wave, whatever. We did build things that did reasonably well. But this time around, it feels a little bit different in that it is a sum total of our experiences. That’s maybe made things a lot easier to navigate, especially given the chaos in the space that we operate and the competition.

Diana: I looked at the numbers and you did nine applications from 2016 up to 2021 when you got accepted.

Jay: Wow.

Diana: And you did four interviews. For all of these, they were all different ideas, but it was the same legal entity.

Jay: It was the same legal entity. Yeah.

Jared: And the same founders.

Diana: And the same co-founders.

Jay: It’s just been—

Jared: You and Frank doing it the whole time.

Jay: Yeah, doing it the whole time. Then after we did YC, our third co-founder joined, and it’s been the three of us for the last four years. Another chunk of time flew by.

Harj: What was the idea you applied to YC with for 2021 that you got in with?

Jay: Yeah, so we were building a serverless platform at the time. It was like Heroku, but for AWS and serverless. We wanted to do a better job in that space and maybe grow the market. We built basically a serverless framework. That was our first big open source project and our first move into building open source products, building in public, doing the whole thing. Eventually, all of that ties into OpenCode.

Harj: Where does the building in public come from? Because you’re clearly doing it now. You’re quite transparent with your metrics and growth, which is awesome. But where does that come from?

Jay: A little bit of that came from Dalton. I think he was pushing us, just in general, saying, “You should probably do things in public because you’re an open source company.” Then it sort of dawned on us—this was maybe 2022 or so—where it was actually Dax, one of our other co-founders, who said, “Look, all your code is public. You work basically in public. If you don’t talk about it publicly, you’re probably doing yourself and your product a disservice.” Ever since then, it’s just been a part of our identity. For a lot of our community, the people that follow us on Twitter, it sounds funny, but to them, it feels like watching a reality TV show of this team and the company and the journey that they’re on.

Diana: I think the thing that’s fascinating is that even though it sounds like such a winding road, and if people just heard the beginning of the podcast, the company sounds like a lightning-in-the-bottle moment, like you just got caught by Stripe and got so lucky. But the reality is you’ve been grinding for a good 10 years and never gave up, which is so impressive. All those winding destinations that ended up being dead ends actually taught you different things because you also ran a consumer company. So you got really good at consumer acquisition, tracking numbers, and all of that, which really plays right now and the level of detail you have for OpenCode and, of course, open source. All of these were not wasted, quote unquote. It was really more a journey that took 10 years to get to zero to 30 million in eight months.

Jay: Yeah, it’s crazy when you put it that way. I think what’s fun is that with OpenCode, we feel like we can go out and address the entire market. That includes consumers, individual users, small teams, obviously the open source community, mid-market, all the way to the enterprise. In our past iterations of all the different products we’ve worked on, we’ve probably done one of each. Now it feels like we get to do all of them together in one part. It’s a lot more fun because you can think about the customer journey through that entire thing, and it makes a lot more sense because you’re not hyper-optimizing for specific parts. I’m not trying to build a very specific enterprise company or a very specific consumer company. We’re trying to just do the whole thing.

Diana: What got you to not give up?

Jay: Partly maybe being a little stubborn. It’s funny, maybe we should put a little warning that says, don’t try this at home. Because I think the prudent thing would’ve been to shut down your company, go join a high-growth startup, learn a bunch of things. But there was something in the back of my mind—Frank is probably the same—in that we felt like we were learning these different things along the way. In our heads, we were figuring out, okay, here’s what it takes to build not just a product, but also do the marketing, understand the positioning, do the whole thing. That journey felt like progress and positive progress. Part of that was obviously us being fortunate to have the ability to do that, which was basically just living with parents for a while, and we ran out of money. But maybe it was a combination of being a little thick-headed and seeing positive progress.

Jared: I’ll admit I’ve been kind of a diehard Claude Code user since Garry Tan became addicted to Claude Code. But recently I’ve been using OpenCode. Very cool. I dropped a bunch of PRs from OpenCode this week and I’ve been really impressed with how well it works with open source models on our existing codebase, which is a very large, very complex codebase. For folks who are watching who have possibly only used Claude Code or Codex or Cursor, how should they think about trying OpenCode and potentially switching to it?

Jay: When you hear about a new model that comes out, especially an open source one, and you want to try it out, you could hack your way into using it with Claude Code or one of the closed source alternatives. But OpenCode is really good for this. You just go in, look at the model picker, and GLM 5.2 or whatever the new open source model is will probably be up there. You can pick it and start using it right away.

Harj: Okay. Well, Jay, I think that’s all we have time for today. I actually learned lots of really interesting stuff about your backstory that I didn’t know in this episode. I think it’s a pretty inspiring story, honestly, for anyone that wants to start a company.

Jay: Or at least entertaining.

Harj: Entertaining and inspiring. It can be. It just hits on so many of the classic startup wisdoms. You should pursue your interests, have eccentric taste, be living the future a little bit. I would argue that trying to order coffee through the terminal is—I’m not sure if that’s living in the future or the past, but it’s not living in the current time. Maybe there’s something in there. Also, just the fact that you kept building your taste and kept building things over a long period of time. Then when lightning strikes, you’re actually in a position to capture it. I think that’s the thing that doesn’t get mentioned: to catch lightning in a bottle, you actually have to position the bottle correctly, be ready for it, and know what to do with it. You guys were well positioned to do that. So congrats on all your success. I know it’s only going to get more explosive from here.

Jay: Thank you. Thank you for having me.