Modal CEO: Has Open-Source LLM Inference Peaked? | What Happens to the Cost of Compute

1. 0:00 – 1:42

   Intro

   1. EBErik Bernhardsson0:00

      I think there is an opportunity to create a $100 billion company with a lot of specialized models in inference. There's gonna be a pullback. I'm okay with that, 'cause I think the long-term opportunity is so fucking big. It's, like, so large.

   2. HSHarry Stebbings0:11

      Today, I'm thrilled to welcome Erik, Co-Founder and CEO of Modal. Now, Modal just hit a billion in ARR. They just raised $750 million led by Accel, and they're priced at $15.75 billion today. But today, we dig into the inference market, what to come, what no one sees that everyone should see, and so much more.

   3. EBErik Bernhardsson0:30

      We've lost new grads going to, you know, some of the labs for, like, 20-plus million. I mean, these are exceptional people. Exceptional. I don't wanna sell this company, 'cause I would just go and start Modal2.

   4. HSHarry Stebbings0:40

      Ready to go? [upbeat music] Erik, listen, dude, it is so great to have you on the show. You know, I was sitting downstairs with you chatting before, and I was just like, you can feel when it's gonna be a good show, because people are themselves, and that was an amazing conversation, and we're gonna have a lot of fun. So thank you for joining me.

   5. EBErik Bernhardsson1:08

      It's so exciting to be here. I mean, I've listened to so many episodes. It's, it's kind of surreal to be inside of it right now.

   6. HSHarry Stebbings1:13

      That's really awesome.

   7. EBErik Bernhardsson1:14

      Big fan.

   8. HSHarry Stebbings1:14

      As I said to you, you, you could've texted me when you started the company, and maybe I would've been a little bit richer.

   9. EBErik Bernhardsson1:20

      Yeah, but you probably didn't know about me then.

   10. HSHarry Stebbings1:22

       No, I didn't, but I would've welcomed investing in another Swede.

   11. EBErik Bernhardsson1:24

       [laughs]

   12. HSHarry Stebbings1:25

       Um, I wanted to start with, I was talking to Zach Frankel last night, and he said about your childhood and, like, early indicators of success and, like, exceptional ability. How did you fall in love with technology? And for me as an investor, what would've made me excited in seeing your childhood?

2. 1:42 – 3:43

   How Erik Fell in Love With Coding at Eight Years Old

   1. EBErik Bernhardsson1:42

      [smacks lips] I started coding 1992 or something like that. I was eight. Uh, and just loved it. You know? It's like one of those things, like, I realized you can control computers. I had a Mac Plus. You know, one megabyte of RAM or whatever.

   2. HSHarry Stebbings1:56

      [laughs]

   3. EBErik Bernhardsson1:57

      Um, and yeah, started making games and, like, I mean, they were terrible, but, you know, it's just ... I, I think back at, like, in a way I'm, like, so grateful. I look back at, like, my childhood. My, my friends, they were, like, collecting hockey cards or whatever. That was, like, their hobby. Like, my hobby was programming, you know? And, and now I make a lot of money from it. Isn't that kinda cra- like you know, you know, when you think about it? Like, the fact that I can make a lot of money, like, doing what, what was my hobby since I was eight. It's kinda nuts. So I, I, I'm just so grateful about, you know, the fact that I fell in love with technology and coding, and then, you know, I started doing programming competitions when I was 16, 17, and so much fun.

   4. HSHarry Stebbings2:32

      Modal was your first company, or is your first company. Why did it take so long?

   5. EBErik Bernhardsson2:37

      I think the consistent sh- pattern in my life is, like, I never had quite enough confidence in myself. Um, [smacks lips] and a lot of it was, that was missing was just the confidence. A little bit the money, too, but, uh, I- so it wasn't until I was, like, 36, 37 I started Modal. And I think, I, I don't think I've ever wanted to necessarily start a company. It was, like, one of those things where, like, I just saw this big opportunity in the market. I almost thought to myself, like, "Why is no one else doing this? Such an obvious idea." And, and that led me to think, like, "Maybe I should start a company." And then, you know, five years later, here we are. Uh, it's gone pretty well. I think in hindsight maybe I should've started much earlier.

   6. HSHarry Stebbings3:16

      Do you wish you had?

   7. EBErik Bernhardsson3:17

      I look at some of these, like, Stanford k- for kids, and they're, like, 21, and, like, I kinda feel like they maybe have too much [laughs] confidence. You know what I mean? Like, they're like, you know, think they're gonna take over the world and they're 21. Uh, I think, I think the ideal age is maybe, like, I don't know, 32 or something like that. You have a little bit more experience. You know how to navigate people. You know how to hire people. You know how to fire people. Like, all these things. Um, I, I don't, I don't know if I would've been a good founder at, like, in my 20s.

3. 3:43 – 5:32

   Why Erik Rejects the “Eating Glass” Founder Narrative

   1. HSHarry Stebbings3:43

      Before we get into the market and everything beneath that, we were chatting downstairs. Um, we often hear that founding companies is like, like eating glass.

   2. EBErik Bernhardsson3:51

      [laughs]

   3. HSHarry Stebbings3:51

      It's hard. It's painful. How do you feel about those two-

   4. EBErik Bernhardsson3:54

      Dude, I, I hate that stuff. It's like, I, I don't get it. Like, you're eating glass, like, every day. Like, why are you doing it then? Like, I don't know. You expect some, like, magic reward, like some big cookie in the end? Like, I never understood that mindset. I, I think it's, like, this mindset that a lot of people have, "I'm gonna suffer pain, and then there's gonna be this big reward in the end." I don't know. I don't think about it that way. I enjoy every day I go to work. Of course there's, like, hard times. Like, there's, you know, you have to have hard conversations, maybe fire people. Maybe you, like, do hard things. But, like, in the grand scheme of things, like, I mean, such is life, and I enjoy life. I, I enjoy building a company, and I don't, I don't think of it as, like, in, one, one day we're gonna flip this company and be o-... I, I hope to work on this for 20, 30, 40, 50 years. I'm having so much fun.

   5. HSHarry Stebbings4:33

      [laughs]

   6. EBErik Bernhardsson4:34

      I don't wanna sell this company, 'cause I would just go and start Modal2.

   7. HSHarry Stebbings4:37

      You really wouldn't?

   8. EBErik Bernhardsson4:38

      No, I don't.

   9. HSHarry Stebbings4:40

      It's funny, I had the chance to sell 20VC, uh, the show's always successful. We said earlier I would just share everything. For, like, 80 million, and I own 100% of it.

   10. EBErik Bernhardsson4:48

       Yeah.

   11. HSHarry Stebbings4:48

       And that for me is a lot.

   12. EBErik Bernhardsson4:49

       Yeah.

   13. HSHarry Stebbings4:49

       And my mother went for a walk with me, and she said, "What would you do tomorrow?" I said, "I'd start The 19 Minute VC."

   14. EBErik Bernhardsson4:54

       [laughs]

   15. HSHarry Stebbings4:55

       And she was like-

   16. EBErik Bernhardsson4:55

       Same thing, yeah

   17. HSHarry Stebbings4:56

       ... she was like, "Don't sell."

   18. EBErik Bernhardsson4:57

       Yeah.

   19. HSHarry Stebbings4:57

       Yeah.

   20. EBErik Bernhardsson4:57

       Yeah, what would you do? I, I don't think money necessarily makes you happy. I mean, a little bit. Like, it's good to have some money, but, you know, beyond a point, like, what, you know. I like building stuff.

   21. HSHarry Stebbings5:06

       You're such a Sw-

   22. EBErik Bernhardsson5:06

       I love technology. I love coding. So much fun.

   23. HSHarry Stebbings5:08

       You're such a Swede. [laughs]

   24. EBErik Bernhardsson5:10

       Is that, is that Swedish, is that Swedish mentality? Maybe it is.

   25. HSHarry Stebbings5:12

       Yeah.

   26. EBErik Bernhardsson5:13

       And actually, you know what?

   27. HSHarry Stebbings5:13

       Dude, never has an American come on before and said it's not about money. [laughs]

   28. EBErik Bernhardsson5:16

       I mean, I do care about money a little bit. But maybe you're right, 'cause I do think in Swedish mentality it's like, it's like self-actualiza- like, doing what you enjoy is, like, kind of the highest form of prestige in a certain way. And it could be music. It could be, like, poetry. Whatever. Like, it almost doesn't matter.

   29. HSHarry Stebbings5:30

       What do you do that you don't

4. 5:32 – 5:56

   Why Erik Hates Fundraising

   1. HSHarry Stebbings5:32

      enjoy?

   2. EBErik Bernhardsson5:33

      Yeah.

   3. HSHarry Stebbings5:33

      Ugh.

   4. EBErik Bernhardsson5:34

      I kinda hate fundraising, honestly. I mean, it's fine. I don't know. Like, I enjoy it when it's, like, good people, but it's, like, kinda challenging. You have to say no to a lot of people. I guess that's a good problem to have. But, you know, in, in general it's, it's-

   5. HSHarry Stebbings5:45

      Oh, you have to say no to a lot of people? Oh, Erik, I'm sorry.

   6. EBErik Bernhardsson5:48

      I know, I know. It's, it's, that's a horrible thing to live with.

   7. HSHarry Stebbings5:50

      Oh, Harry, you can only have-

   8. EBErik Bernhardsson5:51

      I know. It is a tiny violin

   9. HSHarry Stebbings5:51

      ... 10 million, not 50 million.

   10. EBErik Bernhardsson5:53

       Yeah.

   11. HSHarry Stebbings5:53

       Oh. You've never done a deck, have you?

5. 5:56 – 6:25

   Why Erik Has Never Made a Pitch Deck

   1. EBErik Bernhardsson5:56

      I, I've never made a pitch deck, yeah. E- every, every, every round we've done so far it's, I, I like to get to know people well in advance of, of raising money. So, so every round we've done, I mean, most recently we announced the D round yesterday with Accel. I've been speaking, uh, to Rich and Gonzalo at Accel for I think at least a year. And at the, at some point, you know, we're like, "Actually, you know what? We could take some money. Like, what are the terms? Like, let's discuss terms." And then it happened very quickly 'cause we already had that relationship.

6. 6:25 – 8:03

   Why Founders Shouldn’t Always Maximise Valuation

   1. HSHarry Stebbings6:26

      How do you think about price optimization? Is it a game of getting the highest price?

   2. EBErik Bernhardsson6:30

      No, I don't think so. I, I think maybe in later stages it becomes a little bit more of a commodity, like the money itself. But, but I think in the early days, like, uh, like, I mean, for, for several reasons. Like, one is that th- there's a misaligned incentive. Like, VCs always wanna put in way more money than a company needs. Uh, and, and, and that, you know, you just take on a lot of dilution. It's not good as a shareholder. Um, but the other thing is, like, you actually wanna hold back, I think, and, and it's often better to raise a lot less money at a lower valuation and then keep the VCs hungry for the next round.

   3. HSHarry Stebbings6:59

      Can I ask you, is this not the generation, though, where actually capital is the moat and you need so much more money than you think? You said that VCs wanna give you more money than the company needs. I thought AI was finally the output exhaust for venture dollars that was needed.

   4. EBErik Bernhardsson7:14

      I think there are times in the market where you are capital constrained. We have not been capital constrained for most of the per- time of the company. We've been extremely capital efficient. We burn only a few m- million a month. I mean, only, like, you know, it's a lot of money. I have respect for the money, but-

   5. HSHarry Stebbings7:27

      You burn a few million a month?

   6. EBErik Bernhardsson7:28

      No, I mean, up until, like, last year. Now we have to spend a lot of money on capacity. Uh-

   7. HSHarry Stebbings7:32

      How much do you burn a month now?

   8. EBErik Bernhardsson7:34

      A lot. I mean, we, we actually are profitable on a P&L basis, but, like, cash flow is quite negative 'cause we have to spend a lot of money on, on, on getting GPUs and building out capacity. Uh, and so that is a different period in, you know, in the company and how we think about capital. Capital is ... I don't wanna call it a moat 'cause, you know, like, I, I feel like people s- you know, throw around that world, word like to mean, like, things that, you know, ca- M- Moat to me has, like, a very strict definition. Uh, but, but it, you know, certainly helps. Capital helps at this point.

7. 8:03 – 8:52

   Why Early-Stage Startups Don’t Really Have Moats

   1. HSHarry Stebbings8:03

      What's the moat definition for you?

   2. EBErik Bernhardsson8:05

      I mean, moat to me is, like, you go read, like, Seven Powers or, like, one of those things and-

   3. HSHarry Stebbings8:09

      I love Seven Powers.

   4. EBErik Bernhardsson8:09

      It's a good book. But, but I, I think, you know, then people read that and they get delusional. Like, and you know, they're like, "Yeah, we're starting this company. It has all these moats." You never have moats in the early days. Like, you know, like-

   5. HSHarry Stebbings8:20

      Google could do anything they wanted to.

   6. EBErik Bernhardsson8:21

      Google could do ... Yeah, but it's also, like, you know, like, what are the seven? I don't even remember. It's, like, cornering resources, like, market power, like, whatever. It's like none of those things happen until you're, like, 10 years in, you know? I... There's an old blog post I, I always come back to by Ben Thompson. It talks about Snap and the, like, the gingerbread man strategy. Do you know this thing? And it's like, the whole point is, like, Snap doesn't have a moat. They just run faster. Like, they're the gingerbread man. And I actually always come back to that. I, I think that's so true, like, for almost every startup these days. It's like you just have to, like, move faster than

8. 8:52 – 11:43

   Why Brand Matters More Than Ever in AI Infrastructure

   1. EBErik Bernhardsson8:52

      everyone else.

   2. HSHarry Stebbings8:52

      I actually think, like, the three things that matter today when I invest is, like, one, team. Are they a top 1% team?

   3. EBErik Bernhardsson8:59

      Yeah.

   4. HSHarry Stebbings8:59

      Two, speed of product decision-making very specifically. You look at, like, Instinct today. The speed of feature shipping is-

   5. EBErik Bernhardsson9:05

      Yeah

   6. HSHarry Stebbings9:06

      ... really very impressive, and they're very... And then three is brand.

   7. EBErik Bernhardsson9:09

      Yeah.

   8. HSHarry Stebbings9:10

      And I think those are the three determinants to me of defensibility moat, but it's a kind of package.

   9. EBErik Bernhardsson9:15

      Yeah. Yeah. I mean, a- again, I don't know if I would... But, but I, I kind of agree to your point. Like, I think brand, I don't wanna call it a moat, but it, it's, it's actually, I think the value of brand has gone up a lot.

   10. HSHarry Stebbings9:24

       Can I be a dick?

   11. EBErik Bernhardsson9:26

       Yeah, please.

   12. HSHarry Stebbings9:27

       Do you think Modal have done well with brand?

   13. EBErik Bernhardsson9:29

       I think pretty well. I think we have... It's coupled with the developer experience and the product and the website and the visual identity and all these things.

   14. HSHarry Stebbings9:36

       Yeah.

   15. EBErik Bernhardsson9:36

       We just did an event in, in San Francisco, and I thought a lot about that, sort of how we package it, the event and the visual look and feel and the vibes.

   16. HSHarry Stebbings9:43

       But, like, do you think you have a bigger brand than Baseten and Fireworks?

   17. EBErik Bernhardsson9:46

       I think we have that. I think we have, like, really, really good vibes. Like, I'm, I'm terminally online. Like, people tweet, like, way more about Modal. I think there's, like, a little bit more of a community. I think there's, like, much better vibes around us. Like... But, like, I am paranoid about those companies. Like, I am paranoid about competition. Of course, I track it. Like, I q- of course, like, I, I, I see everything that gets tweeted and about our competitors and, you know, it's... But, but this is important. Like, g- everyone wants to be a part of the community. Everyone... You know what I think is, like, the, the ma- the, like, people misunderstand is I, I love Typesafe with Jev.

   18. HSHarry Stebbings10:18

       Yeah.

   19. EBErik Bernhardsson10:18

       'Cause, like, everyone wants to clone Jev. No one wants to get up on stage and wear a fucking pink fur coat and just, like, talk about the Jevolution in an unhinged way. That is what brand to me is. Like, it's... Like, that guy is, like, he's playing the game on a whole different level. And, and I think it's just, like, so, like, true about, like, companies today. Like, you have to think about that.

   20. HSHarry Stebbings10:36

       What, so can you jump out why? 'Cause he's authentic and different? Like-

   21. EBErik Bernhardsson10:39

       I think so. I, I think you just have to create a personality. You kinda have to, like, figure out, like, what is the personality of your company and just be consistent and create, like, almost, like, a sense of movement, a, a sense of, like, community. Like, this is something cool, and, and people wanna use you because you're one of the cool products out there.

   22. HSHarry Stebbings10:55

       Well, like-

   23. EBErik Bernhardsson10:55

       Like, have you tried Modal? It's pretty cool.

   24. HSHarry Stebbings10:57

       What, what has Modal done then that Baseten and Fireworks haven't to build that?

   25. EBErik Bernhardsson11:01

       I think we started with the developer experience, so that was... And we had a very, like, ser- or, like, very clear sort of PLG product market fit, product-led growth, in the sense that people came bottoms up. They just sign up on a product. I think the consumer background of, of coming from Spotify is very important in how I thought about building what's now a B2B company selling tech effectively. You know, it's, like, a very boring category, but I think, you know, like, having personality and, and having, you know, a little bit of an edge is a big part of how you create that sort of movement and that sort of, like, sense of excitement and the feeling that you're different, and I want it to start the moment you go to our website and the moment you try the product and you just feel that it's different, that magic moment.

9. 11:43 – 12:32

   Why Inference Is Becoming the New Cloud

   1. HSHarry Stebbings11:43

      Can I ask you, we've spoken about the three different providers there. Inference as a market has many different speculators commenting on it. When you look at the next five years, how does the inference market play out?

   2. EBErik Bernhardsson11:57

      First of all, I, I think inference is gonna be an enormously large market. It's like the new cloud, basically. And, but I don't think inference in itself is maybe, like, the only thing you need. One of the things that I feel very strongly about is, like, it has to come with a whole set of different, like, a lot of other products. It has to come with post-training sandboxes, agentic stuff, all kinds of, you know, other types of inference, like low-latency inference, batch inference. Uh, so, so, so building that platform, and we-- by the way, we've been working on this for the last five years. I think that to me is, you know, not a mo- but it's a very strong to... It's a very hard product to build-

10. 12:32 – 13:25

    Why Pure Inference Becomes a Commodity Business

    1. HSHarry Stebbings12:32

       And why-

    2. EBErik Bernhardsson12:32

       -that infrastructure

    3. HSHarry Stebbings12:32

       ... why do you need to have the breadth of that product set? Otherwise, you're just in a compression layer.

    4. EBErik Bernhardsson12:38

       Yeah, 'cause otherwise you're just selling commodity product, I think. And like, if you're just selling like tokens for dollars, there's no switching cost, you know? People can just switch to a different provider, right? It's-- That is-- You just-- It's just a race to the bottom. But if you couple it with all these other things around it, sandboxes and post-training and, and, and, and reinforcement learning and all these things, like you, you can actually create a, a whole set of tools. Like, no one wants to build a, you know, buy a single-trick pony. People wanna buy a portfolio of different things. Like, they wanna buy, you know, a whole suite of tools that basically lets them build amazing AI applications on top of it, right? And, and that to me, and the, and the infrastructure underneath is a lot of, uh, uh, uh, differentiation, a lot of, you know, staying power in the market. So it's very hard to copy, like the infrastructure required to build all of these products.

11. 13:25 – 15:00

    Where the Gross Margins in Inference Will Come From

    1. HSHarry Stebbings13:25

       What does no one see about the inference market, or you don't see commented on much, that is very obvious to you?

    2. EBErik Bernhardsson13:32

       One thing that people don't see about the inference market is, I actually think in the long term, there is so much gross margin opportunity in this space because you can build all these things around it. All the sandboxes, all the agentic stuff, workflows, uh, post-training, reinforcement learning. If you do all those things, there's actually so much opportunity to make money.

    3. HSHarry Stebbings13:51

       Is it a market that has software margins? 'Cause we're looking at margins now in the twenty-five to forty percent range. Um, that's kind of traditional players. I don't know if that's yours, but that's kind of what inference providers are generally. Um, do, do they go to seventy as you own more and more of the stack?

    4. EBErik Bernhardsson14:07

       I, I think basically almost no company today has software margins a-anymore. Like, what, what you used to see in like Silicon Valley in the, you know, ten years ago, with these like ninety percent margins plus, I think that's basically gone. And the reason is actually almost like topological. It's like you have this like stack of vendors, right? Like today you have applications, you have models, you have infrastructure, you have cloud, you have Nvidia. And so when a customer pays a dollar at the top of the stack, it has to flow through a lot of different layers. And at every layer, there's gonna be, you know, twenty percent, thirty percent gross margins. So I think there's a lot of margin opportunity, and there's a lot of opportunity to, to, to make a lot of money. I mean, the space is so big, but you're also splitting it across a lot of layers in the stack. So, I mean, you know, like the, the ninety percent software margin businesses are gone. I, I think Modal has an opportunity long term to get to forty, fifty, sixty percent margins, but we're never gonna get to software margins.

12. 15:00 – 16:33

    Why the Inference Market Will Fragment Into Specialized Products

    1. HSHarry Stebbings15:00

       What's a wild prediction you have on the future of the inference market that you think many people would disagree with?

    2. EBErik Bernhardsson15:06

       I think one prediction about the inference market is gonna split into a lot of different things, 'cause the market got so big in the last-- I, I don't think anyone expected, me included, I don't think anyone expected it to be, you know, so big. It's like ten X or a hundred X bigger than I ever thought. And that to me means there's gonna be so much room for specialization, low latency, high latency, like various types of fine-tuning and post-training. I think there's gonna be some wacky new technologies that let you do it with way less resources, new hardware, like all kinds of different things.

    3. HSHarry Stebbings15:33

       I'm really sorry. Does that not go against what you just said, though, about the need to own it all and have that breadth of product? If it's so much larger that you need real specialization-

    4. EBErik Bernhardsson15:41

       Specialization by product. I think it's gonna fragment into a lot of different products. Now, we can build a lot of those products and ho-host it all under Modal, and that to me is very exciting, 'cause, 'cause a lot of different, you know, there's gonna be a lot of different products that companies are gonna need. For one product, they're gonna need low latency inference. For one product, it's gonna... they're gonna need batch inference. For one product, they're gonna need sandboxes, which we have. For one product, they're g- they're gonna do pre-training their own models. You know, maybe it's, uh, you know, certain specialized, you know, audio models or whatever. Like, we do a lot of stuff besides LLM inference. Uh, and, and so that to me, th-that creates a lot of opportunity to basically kind of rethink what a cloud looks like in the era of AI and, and kind of deconstruct the cloud and rebuild it for this new era of software.

    5. HSHarry Stebbings16:23

       Can I ask you, earlier you said about building capacity with GPUs.

    6. EBErik Bernhardsson16:27

       Yeah.

    7. HSHarry Stebbings16:28

       What's it like doing that today? Is it as crazy as reports suggest?

13. 16:33 – 17:30

    Inside the “Open-Air Drug Market” for GPUs

    1. EBErik Bernhardsson16:33

       So it, it is pretty crazy. I spent some time in March, uh, at GTC, Nvidia's big conference in San Jose. I, I would describe that as like an open-air drug market, but GPUs instead of, uh, drugs. Uh, you're running around, you know, working, talking to five hundred different cloud vendors. You're like, "Do you have GPUs?" Like, "Please get..." And, and, like, you know, in the end, you come out, like you do a lot of deals and, and, and that's something we've had to develop the muscle. I think part of the value prop of Modal is like our customers don't have to think about that. We do the dirty work of doing all the GPU deals under the hood. We work with twenty-five different Neo-Clouds or, you know, right now we control like, you know, zero point one percent or whatever, I don't even know the number, of like the world's compute. But like my like, you know, crazy, like, you know, how I think about the, the long-term goal is like as we control more and more of the world's compute, you can aggregate that capacity and aggregate demand on the other side and, and manage that capacity and demand extremely efficiently. So that to me is like something we actually think we can do. We can fill

14. 17:30 – 19:05

    Why Modal Wants to Aggregate the World’s Compute

    1. EBErik Bernhardsson17:30

       a big gap.

    2. HSHarry Stebbings17:30

       What does that allow you to do, being that aggregation layer?

    3. EBErik Bernhardsson17:33

       Well, so we built a lot of our infrastructure ourselves. Like we, we, we don't use Kubernetes and Docker and things like that. We built our own container files. We built our own file system. We built our own container runtime. We built our own scheduler. We built our own, uh, uh, basically how we manage the, the capacity. And that, that enables us to, to basically treat like all of the world's compute as one big pool of capacity. And so if you c- if you think about the fact that there's, you know, we, we already have many thousand paying customers and, and over time it just grows and grows. Each of those customers have like very sort of variable requirements or like their, their, uh, their workloads kind of vary over the day, and they have maybe bursts and, and, and, and each individually the, they're like very sort of unpredictable. But if you pull all of those customers together, you end up with a very s- sort of smooth, uh, overall demand, and that is something we can manage extremely efficiently. Uh, and, and so one of the value props of Modal, besides like we, we can scale up with you, is also You don't even have to do capacity planning. Just come to us, you- we give you 100 GPUs. I, I think it's so amazing going back to Typesafe, for instance, like, they-- we, we, we didn't even-- We had talked to them, uh, I think a long time ago before they launched, but, you know, they launched overnight on Modal, and we, we weren't even aware of it. They just, like, scale up, and we gave them a ton of GPUs, uh, and it just happened automatically. So, so, so that magic, I think, is, is something that is people really value with Modal. It's like we can give them a lot of GPUs on very short notice, and we can handle that sort of variable. Like, if you go to us, you don't have to lock up, you know, like, a rectangular block of GPUs over three years, and you have to know the exact number. Like, it's, it's pay as you go. You scale up and down, you pay per GPU second,

15. 19:05 – 20:31

    How Modal Forecasts GPU Demand Growing 34% Month-on-Month

    1. EBErik Bernhardsson19:05

       per CPU second.

    2. HSHarry Stebbings19:06

       How do you, how do you do GPU buying in terms of forecasting then? 'Cause you don't wanna have an insane excess of supply just sitting on your books. [chuckles]

    3. EBErik Bernhardsson19:15

       Totally. Yeah. The, the GPU f- uh, lead times right now is about three to four months. So, so we basically do a straight line or exponential line extrapolation, and we look at an aggregate across all our customers. We have a very diversified set of customers. You can actually look at the demand in aggregate pretty well and predict based on monthly growth rates, uh, which have been about 34% Month-on-Month, uh, for the last 12 months, and you just project that, you know, three to four months into the future. You end up with, like, two to three X. You go out and, you know, buy that capacity and make sure to ramp up. And, and sometimes-

    4. HSHarry Stebbings19:44

       Can, can you just buy that capacity?

    5. EBErik Bernhardsson19:46

       You can, yeah. I mean, I, I think, I think there's a little bit of, like, a misconception that, like, we're out of capacity. I mean, NVIDIA's, like, shipping, like, an insane amount of GPUs, like, every month. I don't know how many hundred thousand Blackwells are getting installed. Prices are up for sure. You have to pay more to get the same amount of GPUs. And I, I don't wanna say it's, like, an efficient market, but if you're willing to pay more than other people, you can get a lot of GPUs.

    6. HSHarry Stebbings20:07

       Is there a price at which it doesn't make sense for you? Of course there is, but, like, at what point?

    7. EBErik Bernhardsson20:12

       I don't think so. I think, in fact, like, we're actually more long than we're short GPUs, uh, 'cause we, um, you know, we, we can charge a higher price to our customers. Like, there's effectively a market price and, you know, w- we-- it's, like, different dynamics 'cause we lock up longer term reservations and, and rent them out on slightly shorter duration. Um, but, you know, if, if the underlying pri- price goes up, people are also willing

16. 20:31 – 21:14

    What Happens to GPU Prices Over Time?

    1. EBErik Bernhardsson20:31

       to pay more.

    2. HSHarry Stebbings20:31

       Totally get you. What happens to the price over time? You know, H100s were meant to depreciate and go down-

    3. EBErik Bernhardsson20:37

       Dude, I have no idea. Yeah. I mean, and, like, I, I mean, actually, like, it's funny-

    4. HSHarry Stebbings20:40

       Well, dude, if you have no idea, what am I meant to know? [laughing]

    5. EBErik Bernhardsson20:44

       [laughing]

    6. HSHarry Stebbings20:44

       Like-

    7. EBErik Bernhardsson20:44

       Did you see Jensen yesterday? He was in London, actually. He, someone asked him this question and he's like, "Well, obviously, like, demand is gonna catch up with supply-- or sorry, supply is gonna catch up with demand eventually, and then we'll see what happens with pricing. It might happen in a year, it might happen within five years." He said exactly the same thing I always tell people. Like, hardware prices, like, it's like gravity. Hardware prices always go down in the very long run. Right now, we're in a very, you know, there, there's a supply shortage, and, you know, that means that the prices have gone up. They may continue to go up for another year, two years, five years. I don't know. But eventually, it will come down.

17. 21:14 – 22:43

    Why AI Compute Demand Is Still Extremely Early

    1. HSHarry Stebbings21:14

       How early are we? 'Cause, like, you know, I run Instinct for, for everything. I have over 100 automations on Instinct doing, uh, things that are just absolutely wild, from buying concert tickets to planning shows to, uh, helping with, like, medication for my mother. Um, the amount of compute I am using is Jevons Paradox times 100.

    2. EBErik Bernhardsson21:34

       Yeah.

    3. HSHarry Stebbings21:35

       I- if that is equivalent for actually many people in society-

    4. EBErik Bernhardsson21:39

       Yeah

    5. HSHarry Stebbings21:39

       ... because it's accessible within WhatsApp and what, h- how early are we there?

    6. EBErik Bernhardsson21:44

       I think we're very early. It's-- I think you're saying the same thing. I mean, you know, Jevons power, whatever. Uh, like, I think there's just so much latent demand for these things and, you know, I, I think, you know, eventually we're gonna start to run up against these capital constraints, right? Like, there's interest rates are going up, like, we just can't build data centers fast enough. Maybe that's good. I don't know. Like, I think in a, in a way, like, you know, it, it's a guarantee against, like, overbuild, which-- But I also don't know if we can overbuild these things 'cause there's just so much demand for it.

    7. HSHarry Stebbings22:10

       Yeah.

    8. EBErik Bernhardsson22:11

       Yeah.

    9. HSHarry Stebbings22:11

       Like, I don't think you can overbuild in the next five years.

    10. EBErik Bernhardsson22:14

        Probably not, no. No. [laughs] I, I think, I think what could happen is, like, people realize, you know, Instinct is a good example. Like, allegedly, they don't, you know, uh, charge or they don't make revenue yet. So I think it's a good example. But, like, clearly Instinct can make a ton of money in the future, but so, you know, that's the dynamic in the market. There's a lot of companies that are not yet making money yet, and they're using a lot of GPUs. They just have to eventually start to make money. But, like, that, that lag could take a while, right? And, and that we might run up against

18. 22:43 – 23:41

    Does Specialized AI Hardware Threaten Nvidia?

    1. EBErik Bernhardsson22:43

       that.

    2. HSHarry Stebbings22:43

       You mentioned the, the specialization that comes within the inference market, and we, we spoke about NVIDIA there. We're seeing more and more specialization within the chip market.

    3. EBErik Bernhardsson22:50

       Yeah.

    4. HSHarry Stebbings22:51

       Is NVIDIA's business today stronger than ever, or actually weaker than ever as you see the fragmentation of the market into very specialized players from Cerebras to Etched to Olix to Fractal that weren't there before?

    5. EBErik Bernhardsson23:06

       I, I, I think you're, uh, uh, uh, trying to phrase it as, like, one versus the other. M- my, my view is, like, I think, uh, Eric Vishio was actually talk- It's funny, like, the same thing. Like, all of these things are gonna grow. Like, I think NVIDIA is gonna keep growing. They're gonna keep shipping amazing chips. I think there's also gonna be, like, Cerebras and all these Etched and all these other ones, like, shipping specialized chips, and they're also gonna grow. So I, I think there's gonna be markets for all of it.

    6. HSHarry Stebbings23:28

       So why is Cerebras down 30%?

    7. EBErik Bernhardsson23:30

       I don't know. I don't follow really the stock market.

    8. HSHarry Stebbings23:32

       You don't follow the stock market?

    9. EBErik Bernhardsson23:34

       No. I, I'm in, like, a very passive, you know, buy the index. Mostly because of my own time, like, I don't feel like I have the energy to follow the stock market.

19. 23:41 – 24:38

    Why CUDA Is Not Nvidia’s Real Moat

    1. HSHarry Stebbings23:41

       To what extent is CUDA the moat that everyone says it is for NVIDIA?

    2. EBErik Bernhardsson23:44

       I actually-- I think it's a little overstate-- I don't think CUDA itself is the moat. CUDA is actually [chuckles] kind of annoying to work with. It's, like, a pretty low, like, programming model. Uh, but I do think all the NVIDIA, um, the ecosystem is. Like, I, I think, you know, all the kernel optimizations and, and all the stuff that, you know, every, every-- Like, Torch always supports NVIDIA, you know, much better than, than anything else. And, and, and the open source engines like VLM and SGLang, b- because of that, they always support it much better. So, like, everything is NVIDIA first today. Uh, and that's probably gonna continue. It's, like, funny, for the last, like, 15 years I've been saying, "You know, like, next year is the, you know, is the year of non-NVIDIA accelerators." Uh, it never keeps happening. I'm always wrong. Uh, NVIDIA has just, you know, shown to, to be incredibly good at having that staying power and just dominating the market. We'll see. I mean, there's, like, Trainium, there's TPUs, there's AMD. There's a lot of other stuff that does look quite good, and they're catching up. But, you know, I still think they're

20. 24:38 – 26:47

    Can Neo-Clouds Successfully Move Up the Stack Into Inference?

    1. EBErik Bernhardsson24:38

       quite behind.

    2. HSHarry Stebbings24:39

       To what extent are the Neo-Clouds a friend, 'cause they're partnered with you in a lot of cases, I'm sure, as you mentioned, I think 25, you said, or, or a foe? You know, I had the founder of Nabius on, and he was like, "We're coming for the inference market." Like, we need to own full stack. And I, I don't mean that against him, I totally understand it. But are they friends or are they, they foes?

    3. EBErik Bernhardsson25:00

       I feel like everyone in startup land is like a friend and a foe at the same time. They have this like weird sort of, you know, relationship with everyone. I, I think some of the top Neo-Clouds, Nebulas included, uh, they are trying to move up the stack. I'm personally somewhat skeptical. I think it's like very hard to be good at like your core business and also trying to move into different market at the same time. Um-

    4. HSHarry Stebbings25:20

       Well, why would it be hard for them? Like, what's the bear case on them being able to move into the inference layer efficiently?

    5. EBErik Bernhardsson25:25

       I mean, I'm biased, but like we spent the last five years building, you know, an incredibly complex infrastructure stack. I think it's very, very hard to replicate what we've built.

    6. HSHarry Stebbings25:32

       What's the hardest thing?

    7. EBErik Bernhardsson25:34

       Just all the code and, you know, everything around the file system, the container runtime, the scheduler. Like building the system that can, you know, place like millions of sandboxes every minute, you know, scale up and down many, many thousands of GPUs, you know, power hundreds of thousands of tasks running concurrently. Uh, th- these are, these are very hard engineering problems. I, I mean, I'm, I'm an infra geek. Like, I love that shit. Uh, it's so much fun to work on it, but like these are hard problems.

    8. HSHarry Stebbings25:59

       Is Nebius a better business than CoreWeave?

    9. EBErik Bernhardsson26:03

       I actually don't know. Like I, I don't know. I, I think they're both. Like they, they're executed extremely well, both of them. I have a tremendous respect for what we've done. I think at the, at the end of the day, like what res- what the, the, the thing they have the most respect for is actually more on the financial engineering side. They're very good at that kind of stuff.

    10. HSHarry Stebbings26:17

        Why do you say that?

    11. EBErik Bernhardsson26:18

        'Cause I think a lot of the constraint for building a Neo-Cloud very quickly is on the financial side. You need to be very good at, you go lock in like a long-term customer contract, then you go raise a bunch of money against it, then you go buy the chips, then you deliver the chips, and then you pay off the debt, like o- you know, uh, o- over the next few years. Doing that sort of dance with the financials, like i- it's kind of an interesting game. Uh, and CoreWeave has done it exceptionally well. Uh, it, it's a capital game. It's a financial game.

21. 26:47 – 28:01

    What Happens to Open Models as Frontier Models Get Cheaper?

    1. HSHarry Stebbings26:47

       When you move a layer beneath them, we spoke about Jev, who obviously worked with Modal. You [clears throat] we've seen incredible movement from the open source community, which has been exciting.

    2. EBErik Bernhardsson26:57

       Yeah.

    3. HSHarry Stebbings26:58

       But then you also see Haiku and, and Anthropic's newest models being incredibly performant and incredibly cheap. What's the benefit of open models if frontier models are almost as cheap and as efficient?

    4. EBErik Bernhardsson27:13

       I, I think open source is always gonna be-- there's always gonna be great models coming out. Let, let's also remember, like the open source, the, the open source models, they're all moving into licensing models. So like they're getting better at capturing the value of those models, and I think that means they're gonna have more funding to train better models. And I really hope there's gonna be European and US models, 'cause I don't think it's a good situation that we have so many Chinese models. Uh, I mean, they're, they're obviously, I have tremendous respect for all the Chinese labs. They're extremely good at what they're doing, uh, but I think it's good for the world if there's, you know, more power spread out. Um, but, but anyway, going back to the question, I, I do think that open source, uh, there was, there's always gonna be people who wanna fine-tune, wanna have more control over their models and, and sort of tweak the parameters, like go low latency, high throughput, like whatever it is, and, and open source is gonna play a big role in that.

    5. HSHarry Stebbings28:00

       Does it-

    6. EBErik Bernhardsson28:00

       It's like Windows versus

22. 28:01 – 28:39

    Can Europe Produce a Competitive Open Model?

    1. EBErik Bernhardsson28:01

       Linux.

    2. HSHarry Stebbings28:02

       With the, I, um... Sorry. With the greater service base, do you think Europe will produce a competitive open source model?

    3. EBErik Bernhardsson28:08

       Yeah. I mean, I, I'm, I, I love Europe. I, I grew up in Europe. I, I grew up in Sweden, but you know, I, I do think that like Europe does struggle with, you know, accumulating large amounts of capital. I think both of those things are things that Europe struggles with, and, and that is something that, you know, China's done really well. They've accumulated massive amounts of human capitals. Uh, human capital and, and US obviously has very g- you know, very good at like raising enormous amounts of money and like training very large models. Uh, I, I don't know if Europe has the,

23. 28:39 – 29:52

    Why Nvidia Is Backing the US Open-Model Ecosystem

    1. EBErik Bernhardsson28:39

       either of those.

    2. HSHarry Stebbings28:40

       Does it matter if Nvidia owns the open source model layer in the US [chuckles] with first poolside and now reflection?

    3. EBErik Bernhardsson28:49

       I, I think Nvidia is very smart to focus on the ecosystem, building up an ecosystem. Uh, and I think that includes like propping up a lot of open source models, and I think they probably saw an opportunity to pick up open source models. Like, like, like these, these companies are very resource intensive, right? Very CapEx intensive. And, and, and I think it's smart of Nvidia to like buy some of those and sponsor them effectively in order to create like American-built models. By the way, I wanna take back like one thing I said. I, Europe does have an insane amount of human capital. I do think what Europe is very bad at is like aggregating them and putting all of them at the same company. That is something I think US is very good at.

    4. HSHarry Stebbings29:29

       And I think inspired-

    5. EBErik Bernhardsson29:30

       And China

    6. HSHarry Stebbings29:30

       ... a mindset of ambition-

    7. EBErik Bernhardsson29:32

       Yes. Yeah, yeah

    8. HSHarry Stebbings29:32

       ... and huge, huge scale. I think to the point that you said earlier, which is like, "Oh, I didn't have the belief in myself," I respectfully think if you're in the US, you might have been surrounded by influences that told you to start a company earlier.

    9. EBErik Bernhardsson29:43

       Yeah, yeah. I mean, it's almost like too much on the other side to that point. But, you know-

    10. HSHarry Stebbings29:47

        Oh, when you're four, you're encouraged-

    11. EBErik Bernhardsson29:48

        Yeah, yeah

    12. HSHarry Stebbings29:48

        ... to start.

    13. EBErik Bernhardsson29:48

        Yeah, exactly. Yeah, yeah.

    14. HSHarry Stebbings29:49

        Leave school. Drop out.

    15. EBErik Bernhardsson29:50

        Yeah, yeah.

    16. HSHarry Stebbings29:50

        Kindergarten wasn't worth it.

    17. EBErik Bernhardsson29:51

        Yeah, exactly.

24. 29:52 – 30:25

    Why Specialized Models Can Become Huge Businesses

    1. EBErik Bernhardsson29:52

       Yeah.

    2. HSHarry Stebbings29:52

       I, I totally agree with you. I do wanna talk about, you know, we said about mo- models and Jev in particular. Are we seeing like the unbundling of models on a like per function basis, where Jev is obviously very good for its certain tasks?

    3. EBErik Bernhardsson30:04

       I think so, and I think that's a function of what I talked about earlier, where like the market for inference is so large and, and that, you know, when a market is like, you know, a billion dollars, like you're not gonna have too many sub-markets. But when a market is starting to get ten billion, a hundred billion, whatever, like you're gonna have a lot of different sub-markets. And, and Jev does classification and, and things like that, regression type things very, very well, and, and there's a large enough market for that.

25. 30:25 – 32:57

    The Bull Case for Jev at $100BN

    1. HSHarry Stebbings30:25

       Okay. Welcome. It's your first day as a partner at TwentyVC fund, and we're in the investment committee. And very kindly, we have the opportunity to invest-

    2. EBErik Bernhardsson30:33

       All in

    3. HSHarry Stebbings30:33

       ... in Jev. We're all in on Jev's new round?

    4. EBErik Bernhardsson30:36

       All in, all in. I-

    5. HSHarry Stebbings30:37

       What's, what's the bull case for Jev being worth 100 billion?

    6. EBErik Bernhardsson30:40

       I, I told you before, but like I, I think the brand and the community and the movement around, you know, this, just this sort of everything they're doing is, is so incredible. And that to me is like a massive, massive asset in this new age.

    7. HSHarry Stebbings30:51

       That's a $100 billion company

    8. EBErik Bernhardsson30:53

       I think there is an opportunity to create a $100 billion company with a lot of specialized models in inference

    9. HSHarry Stebbings30:59

       You would've done it though at 8.5 billion?

    10. EBErik Bernhardsson31:01

        I would've done it at 8.5 billion. I, I think there's a pretty big outcome here. Could be an exit, could be, you know, like a, a public company, who knows? But, but I do think there's a lot of upside here.

    11. HSHarry Stebbings31:12

        Dude, I'd love to see him ringing the bell [laughs] on the New York Stock Exchange.

    12. EBErik Bernhardsson31:15

        Yeah. Yeah.

    13. HSHarry Stebbings31:16

        Do you know what I mean?

    14. EBErik Bernhardsson31:16

        Yeah. Fantastic guy.

    15. HSHarry Stebbings31:18

        Uh-

    16. EBErik Bernhardsson31:18

        Love that guy.

    17. HSHarry Stebbings31:19

        Is he great?

    18. EBErik Bernhardsson31:21

        I mean, you know, his charisma, his like-- You have to create like a personality, like around your brand. You have to create this like visual identity. It's obviously not the visual identity I would [laughs] have picked. But-

    19. HSHarry Stebbings31:31

        But I think, I think then I hear you, but I think then you should be more loud, if I'm just honest, Erik.

    20. EBErik Bernhardsson31:35

        Modal?

    21. HSHarry Stebbings31:36

        No. We- well, yeah, but you.

    22. EBErik Bernhardsson31:37

        Yeah. I mean, that's possible, right?

    23. HSHarry Stebbings31:39

        Like I, I don't think there's like a personality-led cult from Modal from your perspective.

    24. EBErik Bernhardsson31:43

        I think that's a fair point of criticism. I think I could be out there even more. I am tweet- I try to tweet, I try to-- Uh, blogging, I've been like kind of bad at that recently. You know.

    25. HSHarry Stebbings31:51

        No one gives a shit about blogging.

    26. EBErik Bernhardsson31:52

        Yeah, maybe you're right. Maybe I just need to like go more unhinged on Twitter. It's like founder-led marketing, basically. Like in the early days of, of Modal, like that was a big part of our growth, right? Like me just like tweeting all the time about what we're building, and I think that's really important. If-- I mean, every, every company has to find its own sort of, you know, go-to-market strategy. But in the early days, like we spent a lot of time just talking to people on Twitter, just like hitting people up. "Try the product. What do you think?" Half of them ended up angel investing. You know, they're now supporters of Modal. Uh, and, and, you know, you have to lean into what you're good at. And, and in the early days, I, I did a lot of shit posting.

    27. HSHarry Stebbings32:22

        [laughs] Do you know for every follower that followed me before 50,000, I would send them a personal like thank you message.

    28. EBErik Bernhardsson32:30

        That's amazing. You typed it out yourself?

    29. HSHarry Stebbings32:31

        Every night from 1:00 till 2:30 AM-

    30. EBErik Bernhardsson32:33

        That's amazing
26. 32:57 – 34:41

    How Modal Plans to Expand From a Product Into a Platform

    1. HSHarry Stebbings32:57

       Neil Mehta has a great thing, which is like, how can I 10X the customer experience? And actually, like let's, let's do it. Fuck it. Let's play a game. How do I 10X the current Modal customer experience?

    2. EBErik Bernhardsson33:09

       I think the biggest thing we need is we need way more products in, in Modal. Like we wanna create a ba- basically a comprehensive experience of anything you need to do in order to build an AI application should be a product inside Modal. The other thing we can do is hire way more FDs and way more sales reps. I, I will say we've been behind on that, but we should hire like way more people that reach out to you personally and say, "Hey, I saw you get started on Modal. I saw you working in the bio space. Uh, how can we help you? Here's some bio examples, you know, protein folding," whatever, right? Like we work with a bunch of other similar companies. You seem skeptical.

    3. HSHarry Stebbings33:42

       Well, no, I don't at all. I, I, I'm actually just thinking about like I like the product ambition, but I go as a European who goes, "Wow, they're-"

    4. EBErik Bernhardsson33:51

       We're in the B2B space, so I think it's like, you know, I-

    5. HSHarry Stebbings33:53

       But do you worry about being stretched too thin?

    6. EBErik Bernhardsson33:55

       Yeah.

    7. HSHarry Stebbings33:56

       To your point on like, you know, why can't Nabis do what you do so efficiently? Because you're so good and you've spent so much time. As you expand into all the specialties, there are players in each of them who've been doing that for years, where I'm like, "Oof, Erik."

    8. EBErik Bernhardsson34:08

       I don't think that's true. I, I don't think they're very good, a lot of them. But, but I mean, I, I sort of agree with your higher level premise. Like if there's anything that's like super critical and never goes away, it, it's, it's prioritization and resource allocation, right? Like you can't boil the water and try to do everything at the same time. You need to be extremely ruthless. And, and having a multi-product strategy I think is a very hard balance. Like when do you branch out into second product or third product, right? Like it took us four years to see success with our second product-

    9. HSHarry Stebbings34:34

       What?

    10. EBErik Bernhardsson34:34

        ... which was Sandboxes.

    11. HSHarry Stebbings34:36

        Four years to see success with your second product?

    12. EBErik Bernhardsson34:38

        We, we launched Sandboxes-

    13. HSHarry Stebbings34:39

        But you're five years to a billion in revenue.

27. 34:41 – 35:46

    How Modal’s Sandbox Business Reached a $300M Run Rate

    1. EBErik Bernhardsson34:41

       Yeah. So, so we, we launched Sandboxes in 2023. Uh, didn't see a lot of success in the early days. I, I mean, there was like some success. You know, it wasn't horrible, but it, you know, it was a relatively insignificant part of our revenue, uh, up until, uh, summer of 2025 when, uh, models I think like Sonnet 4.5, one of those, like they, they just got really good at generating code and building whole apps, and then it s- it just exploded. I mean, Sandboxes today is a 300 million run rate business. Uh, it's about a third of our revenue. And, and so that's been a kind of a crazy journey. And now we're thinking about, you know, what's the third and the fourth. We, we just launched a training product. I'm super excited about that. Working on batch products. Working on reinforcement learning. We're working on all kinds of other stuff. And so that evolution into be- becoming a platform, it, it's actually there. Like if you go back to the very early decks of Modal, uh, you know, I put together some just for my own, uh, you know, thinking. Uh, it, it was all there. It's like we, we had this like market map and all these, you know... It ended up being very different products than we thought. But it's so important for me to create like, you know, to focus on a category and focus on a set of solutions, not just like be a single, single widget

28. 35:46 – 37:41

    Why Infrastructure Products Can Take Years to Find Their Killer Use Case

    1. EBErik Bernhardsson35:46

       v-vendor.

    2. HSHarry Stebbings35:46

       What do you know about the Sandboxes market today that you wish you had known when you started and released the first product?

    3. EBErik Bernhardsson35:53

       It, it, it actually is like one of those products. I, I think one of the benefits of being in infrastructure, but also one of the drawbacks in a weird way, is like you kinda just have to trust your gut on like what are like important primitives and just build like amazing building blocks. I, I think we're in the space of like providing building blocks. We try, try to look at like what do, what do engineers need? And, and not like listen too carefully, 'cause like they'll ask for like all kinds of random stuff. We kind of distill it down into like what are some of the building blocks? And we put out a serverless function product in, in 2021. Uh, it, it saw success in 2023 with, with GenAI inference. We put out a Sandbox product in 2023 because we saw a lot of people using LLMs to generate code. And then it saw success in, in 2025, like I said. So a, a lot of like how we think about building products has always been kind of like, let's just put out like good building blocks. And then it exploded. And, and but like to, to answer your question, it was reinforcement learning, it was vibe coding, uh, background agents, and maybe some other stuff too in the future. Uh, and those were things we actually didn't anticipate. So, so I th- I think, I think infrastructure is often like that. Like we didn't anticipate stable diffusion when we started building our serverless function product.

    4. HSHarry Stebbings37:01

       What's-

    5. EBErik Bernhardsson37:01

       It just happened to be the killer app.

    6. HSHarry Stebbings37:03

       What's your takeaway from that? The only thing is, that matters is you've gotta stay in the game long enough for your timing to be right?

    7. EBErik Bernhardsson37:11

       Sometimes I think that's right. I mean, we, we've definitely been extremely lucky in certain ways with timing. Like, we put out products, like, right before. We started the company in 2021, and then, you know, Stable Diffusion and ChatGPT came out in 2022. We had really good timing with that, right? Just created so much demand, and we're very well-positioned. But some of it is also, like, with infrastructure, like, again, like, I think it's about, like, building products that, you know, you anticipate will be useful in the future and, and from your own experience are things that you always wanted to see. Um, I, I, I think you have to kind of trust your gut in that.

29. 37:41 – 38:19

    Why Modal Underinvested in Open-Model Inference

    1. HSHarry Stebbings37:41

       Dude, you said that about, yeah, resource allocation and that-

    2. EBErik Bernhardsson37:44

       Yeah

    3. HSHarry Stebbings37:44

       ... being very, very important as you think about this breadth of products. Where did you allocate resources where with the benefit of hindsight you wish you hadn't, and how do you reflect on that?

    4. EBErik Bernhardsson37:52

       I think we should have allocated more resources towards LLM inference, like, earlier. I think we're a little bit, uh, on, on the, the sidelines for a while. We didn't realize how big open source inference would be. Um, and we thought it would be kind of a race to the bottom, just a commodity space. And then DeepSeek came out, and we're like, "Fuck." Like, we gotta move very fast. And so we staffed up a team immediately. You know, went from, like, you know, one or two people working on this to, like, 30 people 12 months later. Uh, and, and now we're one of the top three providers.

30. 38:19 – 39:11

    Has Open-Model Inference Already Peaked?

    1. HSHarry Stebbings38:19

       Jason Lamkin, a dear friend of mine who I do a show with every Thursday, he always says that now we've reached peak actually, uh, open source LLM inference usage. We're gonna see frontier providers bluntly provide incredibly cheap products that they are now. US enterprises don't wanna use Chinese models. That will continue. And so actually, with those two core prevailing winds, frontier will actually research, and we will see a reduction/plateauing of the OS LLM inference model.

    2. EBErik Bernhardsson38:50

       I think that's a little bit of a fixed pie size mentality. Like, I, I, I think, you know, with, with all due respect, I do think both spaces will grow 5X, 10X, 20X over the next 12 months and, and beyond. Like, I... And, and so, like, it's hard to argue, like, you know, which one's gonna be bigger than the other, but I think both will grow, like, in, you know, orders of magnitude over the next few years.

    3. HSHarry Stebbings39:08

       5X, 10X, 20X over the next 12 months.

    4. EBErik Bernhardsson39:10

       Yeah, for

31. 39:11 – 39:38

    Modal Hits a $1BN Revenue Run Rate

    1. EBErik Bernhardsson39:11

       sure.

    2. HSHarry Stebbings39:11

       You just hit a billion in revenue.

    3. EBErik Bernhardsson39:13

       Yeah.

    4. HSHarry Stebbings39:13

       Well done.

    5. EBErik Bernhardsson39:14

       Yeah. Thank you.

    6. HSHarry Stebbings39:15

       Does that feel pretty cool?

    7. EBErik Bernhardsson39:16

       It feels pretty cool. Yeah. I mean, I, I, I... It's, you know... Like, I, I don't know. Like, I'm this, like, you know, Swedish guy that, you know, grew up in a random... I, I, I... For me, it's like these numbers are absurd, but, like, yeah, like, it feels great. [laughs]

    8. HSHarry Stebbings39:29

       [laughs] I love you.

    9. EBErik Bernhardsson39:29

       It's so much money.

    10. HSHarry Stebbings39:30

        Erik, you're so... Do you know what? I wasn't expecting you to be so, like, likable and cool. [laughs]

    11. EBErik Bernhardsson39:36

        I, I don't know. Like, it's, like, a lot of zeros. Like, how many

32. 39:38 – 40:55

    How Big Can Modal Be in 12 Months?

    1. EBErik Bernhardsson39:38

       zeros are in a billion?

    2. HSHarry Stebbings39:39

       It, it's, it's a lot. Um-

    3. EBErik Bernhardsson39:40

       Yeah

    4. HSHarry Stebbings39:40

       ... your wealth manager will tell you. [laughs] When you look at that, where will you be in 12 months?

    5. EBErik Bernhardsson39:48

       I don't know. What I can say is we've been growing. We grew 15X in the last 12 months, so you can extrapolate that. I mean, I don't know. Like, at some point, we start, you know... Like, markets just get very big, and they start to slow down a little bit. But, um, I, I... With, with the growth rate right now, like, the growth rates are not slowing down. Like, we're growing 30, 40% every month.

    6. HSHarry Stebbings40:05

       Do you think you'll be at 4 billion?

    7. EBErik Bernhardsson40:07

       I mean, I definitely think we should... I, I hope to be way beyond that. We'll see.

    8. HSHarry Stebbings40:11

       With the greatest of respect, 15 billion then feels like quite a cheap price. Like, if I'm an investor, so this is now my ballpark. Like, if you, if you think you're gonna be at 4 billion, say, dropping, I don't know, uh, 1 billion to the bottom line, yeah, then they actually only paid, what, 14X end of year next year bottom line, 3X top line.

    9. EBErik Bernhardsson40:34

       Yeah, it's kinda cheap.

    10. HSHarry Stebbings40:35

        That feels pretty cheap.

    11. EBErik Bernhardsson40:36

        I think it's cheap. I mean, like, I wouldn't do this if I didn't think we could grow and become, like, a generational company. I think there's so much upside. I think this is one of the largest markets opportunities that's ever existed. I, I think the, the demand for compute, the demand for AI applications, it's just gonna, it's gonna be enormous, and they're all gonna need a place to run, and I hope a lot of them are gonna run

33. 40:55 – 42:28

    Modal’s Next Big Revenue Opportunities

    1. EBErik Bernhardsson40:55

       a Modal.

    2. HSHarry Stebbings40:55

       What revenue line do you not have today, or is tiny, that you think will be very significant in three years' time, five years' time?

    3. EBErik Bernhardsson41:04

       That is a good question. I think training is, is something we just launched. I think that could be very large. I, I hope that could get to 100 million run rate in the next six to 12 months. Um, post-training similarly, uh, working on batch jobs. Um, I'm very excited also about a lot of, like, sort of traditional, like, data warehousing stuff, databases, um, workflow orchestration, uh, feature stores, like, vector database. I mean, there's just so many parts of this, and they all kind of fit together very nicely, and that, that to me is, like, what I'm so excited about. It's like, once you start building in this space, there's so many natural adjacencies. Like, this is, like, a problem with me is, like, I get excited about so... I have, like, no lack of ideas. Like, at Modal, like, we have so many ideas of things we wanna go. CI, I would love to get into continuous integration. That's, like, a very interesting space. Agent harnesses. Maybe we should train models. Maybe... You know, there's just 5,000 directions, you know, and, and that's like my ADD brain, like, wants to run in all of those directions. You have to be obviously ruthless with prioritization and only take on, like, one or two things at, at the right time, but, but, but, but there's so many different directions we can go.

    4. HSHarry Stebbings42:04

       It's amazing that your wife calls you for decisions. [laughs] She must get a very incoherent answer.

    5. EBErik Bernhardsson42:10

       I know, I know.

    6. HSHarry Stebbings42:11

       What is this? Is this, is this? I just wanted some curtains done.

    7. EBErik Bernhardsson42:13

       Yeah. You know, with the curtains, I just like pick one of them.

    8. HSHarry Stebbings42:16

       [laughs]

    9. EBErik Bernhardsson42:16

       Have you seen Simpsons and, like, Rainier Wolfcastle becomes the president-

    10. HSHarry Stebbings42:19

        Yeah

    11. EBErik Bernhardsson42:19

        ... and they're like, "Which one do you wanna pick?" You know, "One to seven?" He's like, "I was born to lead, not to read."

    12. HSHarry Stebbings42:26

        [laughs]

    13. EBErik Bernhardsson42:26

        "Number four."

34. 42:28 – 44:13

    How Important Will Post-Training Become?

    1. HSHarry Stebbings42:28

       Yeah, I, I, I totally, uh, get that. Can I ask, we've seen this real shift, I think, in, in emphasis but also in resource allocation from pre-training to post-training. Um, how do you feel about that, and do you think that is the new normal?

    2. EBErik Bernhardsson42:45

       I, I, I don't know. I, I think post-training works because there is a lot of tasks that people do in real life that have fairly finite sort of, uh, feedback mechanisms. So coding is, like, a great example, right? Like, you can do coding with reinforcement learning pretty well, uh, because, you know, you can have a... It c- you can run code in sandboxes, and a lot of labs are running, uh, uh, uh, modal sandboxes for that exact purpose. Uh, you start to look at cybersecurity is another area. Math obviously with Lean. I mean, we've seen some fantastic pro- I think it's kind of incredible all these, like, OpenAI, uh, proofs the other day. So law maybe to some extent you can, you know, h- do, like, eval. So I, I think, I think post-training right now, like, there's, like, s- such incredible focus on it because you, you, you, you have all these tasks that traditionally only humans can do but are actually well within reach for, for agents and, and for LLMs. Um- But, but, but like I, I don't know if it extends beyond that. Like, I don't know. Like, there's a lot of tasks out there that are like have way longer feedback loops that I don't know if like post-training quite yet are, is, is within reach. A lot of people ask me like, "What's gonna happen with software engineers in the long term?" And I think, you know, we're probably gonna automate everything that a, that an agent can do within 24 hours or, or a week or something like that. But I kinda feel like, you know, engineers, like, they, they work on a longer timescale, and they have feedback loops that are longer. So I, I, I have a feeling like reinforcement learning, we're gonna start to run into some limitations around that. But I've been wrong for so long, so maybe I'll be wrong, and just keeps going, and eventually we can re- do reinforcement learning on things that take six months for a human. I don't know.

35. 44:13 – 45:26

    Why AI Agents Won’t Kill Software Engineers

    1. HSHarry Stebbings44:13

       What does the day look like in three years for an engineer?

    2. EBErik Bernhardsson44:16

       I mean, I, I'm extremely bullish on software engineering. I've been coding for 30 years, and I've seen so many different changes with like programming languages and the cloud and, you know, ed- so many different frameworks. And it, it, it's, you know, every single time we make pro- product, pr- programmers more efficient, it just increases demand for software. There's so many parts of the world that desperately need software, and so I, I'm very, very bullish on software, uh, and, and the need for soft- software engineering. And I think one of the bi- biggest misconceptions from the outside is that developer is about like translating from English to code. Like, that has never been in, you know, the job of software engineers. It's like to understand the requirements in the first place and be kind of an interface between the humans, you're building for humans at the end of the day, and like trying to figure out, like what does that even mean in terms of processes and workflows? And, and, and that continuously moves up in the stack. You know, back in the '60s, we used to write assembly and punch cards, and then we moved to COBOL, and then to C++, and then to Python, and, and now we're using agents. And that's just like one... And, and like it's definitely like the biggest thing that I've seen in the last 30 years, but I do think we have many, many analogs and many parallels throughout the, the history of, of, of software and computers. Uh, to me, that makes me feel like, you know, software engineers are gonna be fine for the next 10, 20 years.

36. 45:26 – 46:06

    Modal Spends ~$500K Per Month on Coding Agents

    1. HSHarry Stebbings45:26

       How much money do you spend on coding agents a month at Modal?

    2. EBErik Bernhardsson45:29

       I think a half a million or something like that.

    3. HSHarry Stebbings45:31

       How many engineers do you have?

    4. EBErik Bernhardsson45:32

       150.

    5. HSHarry Stebbings45:33

       Dude, that's so, that's so worrying, you know.

    6. EBErik Bernhardsson45:35

       You wanna-- Should be token maxing more?

    7. HSHarry Stebbings45:38

       They, they should be 10X that.

    8. EBErik Bernhardsson45:39

       Yeah, maybe. I don't know.

    9. HSHarry Stebbings45:41

       Uh, Brandon said, uh, McCool, um, that they spend 3X their human salaries on inference.

    10. EBErik Bernhardsson45:47

        Yeah. Yeah. Yeah, we're not quite there yet. Uh, but I will say like our software engineers are extremely agent-focused, including myself. I mean, I, I write code with agents like all the time. Um-

    11. HSHarry Stebbings45:56

        But I think like the big question if you're looking at like an Anthropic IPO, and don't worry, I'm not doing a Bloomberg and like pulling you in to be like, "What do you think of the Anthropic IPO?" [laughs]

    12. EBErik Bernhardsson46:03

        I hate those ones.

    13. HSHarry Stebbings46:04

        Yeah, you know what it's like.

    14. EBErik Bernhardsson46:05

        I think it's great.

    15. HSHarry Stebbings46:05

        Yeah,

37. 46:06 – 47:11

    Will Coding-Agent Spend Keep Growing as Models Get Cheaper?

    1. HSHarry Stebbings46:06

       it's great. It's, uh-

    2. EBErik Bernhardsson46:07

       Wonderful.

    3. HSHarry Stebbings46:07

       No one in AI is like, "Oh, not a fan." [laughs]

    4. EBErik Bernhardsson46:10

       Yeah, yeah, yeah.

    5. HSHarry Stebbings46:10

       Everyone is-

    6. EBErik Bernhardsson46:11

       I'm against it.

    7. HSHarry Stebbings46:12

       Go. Go. [laughs] Um, but I do think a big, uh... The motherload, so to speak, is Claw Code. It's $60 billion of revenue. It largely kind of, um, dictates the $2 trillion price. Um, and you look at Marc Benioff, who says that they spend 5% of engineers' salaries, 300 million, on Claw Code today. Now, if that comes way down as efficiencies improve, you name it, that's a challenge. Or will that go up as it just eats more and more of engineers' workflows?

    8. EBErik Bernhardsson46:46

       Yeah, it's just gonna eat more. I mean, like, again, like we've seen this so many times. Like you make something more efficient, it makes something more powerful, it just unlocks like five new use cases, and the demand just goes up in the end. Like, that's, that's just the history of like the last 50 years of computers and software. But like I wouldn't be here if I wasn't such an optimist. That's why I started a company. And like I, I'm just like the biggest, like, optimist that there is in the world about software and computers and like I just love this shit, and I think it's just

38. 47:11 – 48:46

    Why AI Training Data Is Such a Lucrative Market Today

    1. EBErik Bernhardsson47:11

       gonna keep growing.

    2. HSHarry Stebbings47:12

       The data market is an interesting one. We have so many players-

    3. EBErik Bernhardsson47:17

       Yeah

    4. HSHarry Stebbings47:17

       ... that have scaled so fast.

    5. EBErik Bernhardsson47:19

       Yeah. It's incredible. If I was 20, and I just wanted to make a lot of money, I would go out and start an environment company and just sell data to Frontier labs. There is so much money to be made in that space. It's kind of incredible.

    6. HSHarry Stebbings47:30

       Let's go a layer deeper. What would you do next? You specialize on a specific market?

    7. EBErik Bernhardsson47:34

       I mean, the coding space is where you see the most demand right now. Like, if you can sell like code environments to, to the labs, like you can make a lot of money. Like, there's a big space in that.

    8. HSHarry Stebbings47:43

       Still?

    9. EBErik Bernhardsson47:43

       Still, yeah. Yeah. I meet startups all the time. I, I angel invested in a bunch of that. I mean, they're incredible businesses. I, I don't think-- What I will say, like my one piece of advice is like you make a lot of money today, but like is this a venture fund? Like, do you wanna get on the venture treadmill? I'm not sure. Like, I think these are companies that are gonna make a lot of money today. I don't know if they're gonna make a lot of money in the l- very long run, but like who cares? Like, you know, take the money. It's like great money. Like, it, it, it's a fantastic business today, and that's all that matters.

    10. HSHarry Stebbings48:10

        I think there's, there's, there's data marketplaces, and then there's data and services, and when you look at the, the leading providers, they started as the data acquirers and then just sell you the data, and that's kind of it. And I think they've really gone full life cycle now to deployment, implementation, evaluation, monitoring with a real enterprise hand-holding ability-

    11. EBErik Bernhardsson48:29

        Yeah

    12. HSHarry Stebbings48:30

        ... similar to the FDs-

    13. EBErik Bernhardsson48:31

        Yeah

    14. HSHarry Stebbings48:31

        ... where they will really be that partner to ensure full implementation.

    15. EBErik Bernhardsson48:36

        Yeah.

    16. HSHarry Stebbings48:37

        That's how you expand margin and, and have some form of business in the long term.

    17. EBErik Bernhardsson48:40

        For sure. Yeah, and I think it's gonna be a tough business in the very long run, but like today you can have extremely good cash flow selling

39. 48:46 – 51:24

    Should Every Company Own Its Own Intelligence?

    1. EBErik Bernhardsson48:46

       these datasets.

    2. HSHarry Stebbings48:47

       Can I ask, when we think about that kind of implementation, working with partners so closely, you've said, and Lynn has said too from Fireworks, about the importance of owning your own intelligence. Can you unpack how this looks? Because specialized models for companies in this way, kind of jury's still out.

    3. EBErik Bernhardsson49:05

       Yeah.

    4. HSHarry Stebbings49:05

       You know, we have Harvey who says that it's fantastic, and, you know, it's been amazing. And then you have Lagora who say, "Ridiculous, the underlying models move so fast, it's the wrong strategy."

    5. EBErik Bernhardsson49:17

       Yeah. For some stuff it's good to own your own intelligence. For other stuff it's not, like, I don't think it matters. Like, like we have a lot of customers who are on proprietary models on, on Modal too. Like, and, and their customers, you know, they use a proprietary model. Like, I don't, I don't think it makes sense for everyone to own their own intelligence.

    6. HSHarry Stebbings49:29

       Where does it make sense for you to versus not?

    7. EBErik Bernhardsson49:32

       I, I think for certain use cases, you can take an open source model and you can fine-tune it, and you can get really good performance. And the nice thing is you know what the fine-tuned model is based on. It's not gonna change. Like, they're not gonna take away the model from you. It's nice to have that val- like, that consistency and, and ability to, like, continuously, like, tweak it without n- w- without, you know, the underlying model changing, which has happened in some cases with the Frontier Labs. I, I, yeah, that's like, you know, that's nice. Like, you know, I support that, but, like, I, I, you know, I think there's many other things that matter to, to people. You know, owning your intelligence is, like, just one thing out of many. Like, I think latency matters, I think cost matter, performance. I think, you know, hackability. I think, you know, being, you know, having a platform of a lot of adjacent services, like, these things matter, too. So owning o- your own intelligence, like, yeah, I'm supportive, but, you know, there's many other things.

    8. HSHarry Stebbings50:18

       Isn't that kinda going against your business?

    9. EBErik Bernhardsson50:21

       No, I don't think so. I, I, w- the business of-

    10. HSHarry Stebbings50:23

        I mean, I love the honesty, but, like, dude, it would be in your favor to say, "Yes, every company's gonna need it, and we're the partner to help you." I kind of fucking teed you up there, and you've [laughs]

    11. EBErik Bernhardsson50:33

        Well, I mean, not, not really. I mean, if you think about TypeSafe with Jev, right? Like, you know, it's a highly proprietary model. Like, I want people to use Jev. I mean, it runs on Modal [laughs] like... Like, so we have many customers, like, that themselves are proprietary model, like, that are hosting on Modal that are serving their customers, and I'm very [laughs] very in favor of that, too. I, I think the beauty of Modal is always that it's, like, a general purpose platform, and we support so many different things. That, that to me is, like, you know, the beauty. It's like, you know, we can-- if you wanna use open source models, like, that's great. Like, we'll help you, and there's many customers that build incredible AI applications on top of open source models, fine-tuning them, reinforcement learning, have a bunch of, uh, you know, adjacent sandbox use cases or training, whatever. But there's a lot of companies that have built highly successful businesses, you know, Suno, Runway, TypeSafe, and many others, like, running their own proprietary model on Modal, like, and, and their customers don't care. And, and that, that, that I think is, like, you know, puts us

40. 51:24 – 52:40

    Is Model Routing a Defensible Business?

    1. EBErik Bernhardsson51:24

       in a really good spot.

    2. HSHarry Stebbings51:25

       I wanna ask two more questions before I do a quick fire.

    3. EBErik Bernhardsson51:27

       Okay.

    4. HSHarry Stebbings51:28

       Model routing, as the dear Americans say, is that a defensible business?

    5. EBErik Bernhardsson51:33

       What do you say in the UK? Routing?

    6. HSHarry Stebbings51:35

       Routing.

    7. EBErik Bernhardsson51:35

       [laughs]

    8. HSHarry Stebbings51:36

       Routing.

    9. EBErik Bernhardsson51:37

       I love that.

    10. HSHarry Stebbings51:38

        Well, I mean, we are actually. Uh, the original language, America just took ours and bastardized it.

    11. EBErik Bernhardsson51:42

        Yeah.

    12. HSHarry Stebbings51:42

        But yes, um, uh, my question to you, is that a defensible segment of this value chain, or is that a commoditized product that everyone just builds themselves?

    13. EBErik Bernhardsson51:53

        I think it's possible to build a successful business in model routing. Uh, I think you can do that today as a standalone product. I wonder if long-term it has to be part of the more com- comprehensive platform. At the end of the day, I think what, what these companies want, they want evals, they want, you know, ways to fine-tune. They want, you know, more control over, uh, performance, like latency, throughput, like all these things. And I, I, I think it's, like, one of those things that kinda makes sense today. Like, these are very good businesses. I invested in OpenRouter personally, uh, and they had a great outcome. Uh, I actually think they could have done even better on a standalone basis. Um, but, but I think in the very long term, it, it feels like one of those categories, like, you know, and, and but that's my view on everything in AI. It should really be a part of, you know, a larger, more comprehensive set of tools.

41. 52:40 – 53:45

    Why Erik Would Have Told OpenRouter Not to Sell

    1. HSHarry Stebbings52:40

       Congratulations, you're on the OpenRouter board, and a acquisition offer comes in. Please make the case for why they would have done that.

    2. EBErik Bernhardsson52:47

       Reject. Yeah.

    3. HSHarry Stebbings52:47

       Yeah, please make the case for why-

    4. EBErik Bernhardsson52:48

       I think, I think OpenRouter, like the growth rate and the revenue, like there's just... And, and the brand, frankly. Like, it- they have a really strong brand. I underst- I mean, there's a lot of, like... There's 500 other competitors, but, but, but, but the, the space is growing so fast, and there's so much revenue. Uh, I, I would've told them to keep going.

    5. HSHarry Stebbings53:04

       Final one. I think a really telling question is what do you believe that does not go in the favor of your business? And so, for example, I think that seed investing today is kinda dead and bullshit. This idea of, like, two on 10, um, we'll slowly build for 18 months.

    6. EBErik Bernhardsson53:24

       Well, it's kinda bad-- Like, it's kinda bad if you wanna make money, right? But it's great if you wanna develop a relationship with engineers or, or me- I mean, with founders.

    7. HSHarry Stebbings53:30

       Sure. But that goes against what I do. I'm, I- I'm traditionally-

    8. EBErik Bernhardsson53:34

       You're, you're against a relationship with [laughs] with founders or-

    9. HSHarry Stebbings53:37

       No, no, no. I'm saying, like, I think the death of seed investing, the two on 10 and build for 18 months, is dead.

    10. EBErik Bernhardsson53:42

        Yeah. No, I, I understand what you're saying.

    11. HSHarry Stebbings53:44

        So what would

42. 53:45 – 54:18

    Why Erik Expects a 20–30% AI Market Pullback

    1. HSHarry Stebbings53:45

       yours be?

    2. EBErik Bernhardsson53:45

       I, I don't know. I, I mean, I think, I think one, one thought I have is, like, obviously at some point, this AI stuff is growing so fast, and we're here, and we're gonna go to here, and then there's gonna be pullback, you know, 20%, 30% in the stock market. Everyone's gonna freak out, and, like, people are gonna say it's the end of the world. I'm almost certain it's gonna happen, but, like, I'm so bullish 'cause then we're gonna keep going and, like, you know. Like, so, so that, that would be, like, one thing that I'm like, you know, there's gonna be a pullback. I'm okay with that, uh, 'cause I think the long-term opportunity is so fucking big. It's, like, so large.

43. 54:18 – 55:25

    Why Employee Secondaries Matter in the AI Talent War

    1. HSHarry Stebbings54:18

       Totally agree. Do you take secondaries?

    2. EBErik Bernhardsson54:22

       Very little.

    3. HSHarry Stebbings54:24

       Why?

    4. EBErik Bernhardsson54:26

       I, I mean, I personally don't need a lot of money. I, you know, I, I'm pretty happy with life, and I have enough, uh, so, so personally I don't. I, I actually support it for my employees, though. I think, um, it's, it's quite important actually as a founder to, uh, to that, um, you make the equity worth as much as possible. And one of the way to do that and to tell future candidates that, "No, this isn't just pa- paper money. This is real money," is to have a precedent of s- of doing employee secondaries. And you can point to, you know, last round we let employees sell this much or whatever, and, you know, up to 10%. Like, I, I think you have to have guardrails. Like, you don't want employees to sell too much. Uh, you want them to still be very, you know, have a lot of skin in the game and, and be very incentivized to, to make the company successful. But I think it's a w- it's a way to, like, you know, like, m- create a very liquid, hard currency out of the equity.

    5. HSHarry Stebbings55:20

       I totally agree, and as you think about a talent marketplace that's-

    6. EBErik Bernhardsson55:23

       Absolutely

    7. HSHarry Stebbings55:24

       ... ever more competitive,

44. 55:25 – 57:10

    Why Top AI Labs Are Paying Exceptional New Grads $20M+

    1. HSHarry Stebbings55:25

       it's very-

    2. EBErik Bernhardsson55:25

       It's extremely competitive out there.

    3. HSHarry Stebbings55:27

       Yeah. Is it nuts?

    4. EBErik Bernhardsson55:29

       It's nuts. It's nuts. I mean, you know, it, it, like, the, the offers we see for we've lost new grads going to, you know, some of the labs for, like, insa- like, talking, you know, 20-plus million I mean, these are exceptional people. Exceptional

    5. HSHarry Stebbings55:43

       What makes an exceptional, exceptional new grad?

    6. EBErik Bernhardsson55:47

       You know, like, just, just fucking sharp people. You know? Th- these are like... I, I know it when I see it. Like, they're just, you know, they're just extremely hungry, hardworking-

    7. HSHarry Stebbings55:58

       And where do they go?

    8. EBErik Bernhardsson55:58

       They know everything about technology. I mean, they go to some of the big labs.

    9. HSHarry Stebbings56:02

       Is that it? They go to OpenAI and Anthropic.

    10. EBErik Bernhardsson56:03

        Yeah, the big labs. Yeah, they pay a lot.

    11. HSHarry Stebbings56:05

        But is that it? Is that where the talent comes?

    12. EBErik Bernhardsson56:08

        There's, like, a few other ones, but, but, I mean, mainly those people are, are known for paying very, a lot of money.

    13. HSHarry Stebbings56:13

        And you won't pay that much?

    14. EBErik Bernhardsson56:15

        We pay a lot. We can go very aggressive. Uh, we wanna be out there paying 95- fifth percentile for in the market. And, and there's, like, a few companies that I know, like, they're probably gonna pay a little bit more. I think we have way more upside in equity, and going back to what I said about secondaries, I think it's important that for, for that reason also, like, to create, like, a real tangible value in the currency, in the equity that makes us, you know, our offers seem very, very strong. Plus the fact that we have a lot of upside. I mean, you, you, you said it, right? Like, our stock seems cheap. Like, I kind of agree. There's a lot of upside in our stock. So that enables us to not necessarily have to match with the labs, but I think, you know, create a compelling sales pitch for why our comp package is, like, basically better than the labs, and we're willing to go very far to get the best talent.

    15. HSHarry Stebbings56:58

        Okay, we're gonna do a quick fire round 'cause otherwise I could talk to you all day. You have a great angel portfolio. OpenRouter, uh, you mentioned Lovable before. You can put all of your money into only one.

45. 57:10 – 57:39

    Why Erik Would Back TurboPuffer

    1. EBErik Bernhardsson57:10

       I love TurboPuffer. I know Simon really well. He's a Danish guy, so we have the Scandinavian connection. Um, just a fantastic entrepreneur. I just love the business.

    2. HSHarry Stebbings57:20

       How big can that business be?

    3. EBErik Bernhardsson57:21

       That's a good question. I mean, it is, like, somewhat specialized. They're doing vector databases. Um, I, I think it could be a very large category. If I ran that company, I would have a 500 different ideas of how to expand into many adjacent things around databases, data warehouses, other types of query, querying system. Like, there, there's... It could be a large category.

46. 57:39 – 58:04

    What Erik Changed His Mind on About the Sandbox Market

    1. HSHarry Stebbings57:39

       What have you changed your mind on most in the last 12 months?

    2. EBErik Bernhardsson57:43

       Probably the magnitude of the sandbox space. I think it's kind of insane. Like, I look at the revenue and it just keeps going up, and I'm like, "Well, clearly it has to stop right now." And I think I'm, like, doing some sort of reverse psychology on the revenue 'cause it just keeps going up and I'm like, "What is going on?" It just keeps growing and growing and, like, I've been wrong so many times, so like, you know, in a good way. Like, you know, and, and, and so I- I... It's incredible how large that space has gotten.

47. 58:04 – 1:01:07

    Which AI Companies Erik Would Avoid Investing In

    1. HSHarry Stebbings58:04

       What do you think is the most overheated category within AI today?

    2. EBErik Bernhardsson58:10

       Oh, boy. I, I don't know. I, I don't think anything is overheated. I, I will say, like, one of the things that I'm a little nervous about is companies that are dependent on continual infusion of equity capital. I think I would not invest in them. Uh, and, and there's a lot of them that I... I, I would be nervous about anything where, like, if there's a venture capital pullback, they are, you know, s- they will struggle to raise more money. I think it's important that you raise the, the r- every round as if it's your last round, and there's, like, a real path towards commercialization and revenue.

    3. HSHarry Stebbings58:45

       You got two kids.

    4. EBErik Bernhardsson58:46

       Two kids.

    5. HSHarry Stebbings58:46

       Biggest lesson on being a good dad and being a good CEO?

    6. EBErik Bernhardsson58:49

       Uh, I mean, sometimes I feel like it's kind of similar. Like [laughs] uh, you mean, like, in terms of, like, how to balance it? Or do you mean in terms of just, like, being a good dad in general?

    7. HSHarry Stebbings58:57

       Like, what does it take to be a good dad?

    8. EBErik Bernhardsson58:59

       Setting them up for success, uh, i- in, in life. And I always, like, think about, like, the right amount to push the kids. I'm not a tiger dad, but, like, I think s- a little bit you need to push the kids a little bit to, like, get them exposure. I don't wanna create, you know, in the future a situation where my kids come back to me and say, "Hey, Dad, I wish you'd made me do X," or, "I wish you'd made me do Y." But I also don't wanna create a situation where my, my kids come back to me when they're 20 and they say, "Hey, Dad, why did you make us do X?" Or, "Why did you make us do Y?" I, I... So a, a little bit of pushing. I make them play piano. I don't know why, but, like, I think it's, like, good way to, like, learn something that I hope, you know, when they're 20, 30 they will actually appreciate. They don't like it that much, but I think stuff like that. You gotta push them a little bit, but not too much. My, my parents are very lax and, you know, I discovered math and, and programming and they were like, "Erik, why are you doing so much math? Like, you should go out and play like the other kids." Uh, so [laughs] in a way I actually wish they would've pushed me even harder, but I pushed myself so hard anyway, so it doesn't matter. Uh, so, so you... Like, that, like, right amount of pushing.

    9. HSHarry Stebbings59:57

       How did you push yourself so hard at... You're, sorry, you're actually from a nation where you are told that it's, like, bad to stand out. I'm forgetting that word, but I had a Swedish girlfriend. Uh [laughs]

    10. EBErik Bernhardsson1:00:08

        Yeah. The, the Jante Law.

    11. HSHarry Stebbings1:00:10

        Y- yeah. Where it's like it's bad to, to stand out.

    12. EBErik Bernhardsson1:00:13

        Yeah, don't think you're special.

    13. HSHarry Stebbings1:00:14

        Yeah, why-

    14. EBErik Bernhardsson1:00:15

        But I, I don't know if that's true. Like, I think there's a lot of misconceptions to Scandinavian culture. I do think there's a lot of, like, respect for people who follow their dream and, like, do things and, like, whether, like I said, you know, it's music or writing books or, you know, making furniture or art, like, or being an entrepreneur. Like, I actually think there's a lot of respect for s-... I think this, like, you know, self-actualization and following your dreams is, like, kind of highly regarded, and I think, you know, it's like, uh, they... Like, that's the flip side of having a very lax culture is, like, you kind of pursue people to follow their dreams. And I don't know. I... My dream was, like, doing math and programming. I love that shit, so I did that.

    15. HSHarry Stebbings1:00:47

        Unless you're from Norway where a tax structure exists that basically makes it impossible to start a company.

    16. EBErik Bernhardsson1:00:52

        Yeah. I mean, Norway has some issues.

    17. HSHarry Stebbings1:00:53

        [laughs] Um, yeah.

    18. EBErik Bernhardsson1:00:55

        They also barely work.

    19. HSHarry Stebbings1:00:56

        Yeah. This is very true. But we do have a sovereign wealth fund.

    20. EBErik Bernhardsson1:00:59

        Yeah.

    21. HSHarry Stebbings1:00:59

        So thank-

    22. EBErik Bernhardsson1:00:59

        They are rich. Yeah

    23. HSHarry Stebbings1:01:00

        ... thank God for that.

    24. EBErik Bernhardsson1:01:01

        Maybe it's, like, the first post-economic society. I don't know.

    25. HSHarry Stebbings1:01:04

        Uh, uh, literally almost.

    26. EBErik Bernhardsson1:01:04

        In certain... In, in multiple meanings of the word.

48. 1:01:07 – 1:02:01

    Why AI Drug Discovery Excites Erik Most

    1. HSHarry Stebbings1:01:07

       Um, final one for you. What are you most excited for? It can be anything. Like, I, I look at, um, drug discovery for kind of chronic conditions. My mom's got a chronic-

    2. EBErik Bernhardsson1:01:15

       Actually, drug discovery is, like, one of the application I think is so incredibly cool. Like, I, I can't wait to take AI-generated drugs. You know, I'm, like, old and my skin is getting bad and, like, I don't know. Like, you know, every time I drink a beer I feel, like, horrible the day after. Like, I don't know. Give me all the AI-generated drugs I can take. Like, I, I wanna live long. I wanna live a healthy life. I, I, you know, and I, I think it has-

    3. HSHarry Stebbings1:01:35

       Do you think that is actually a realistic future?

    4. EBErik Bernhardsson1:01:37

       I think so. But, but I think with the FDA and other things it is gonna take a while. But I, I think there's just so much incredible progress today in, like, drug... In, in, in AI-aided drug design. I think there's so many cool stuff. So many things.

    5. HSHarry Stebbings1:01:49

       Dude, I've so enjoyed this show in a way that I, I honestly wasn't expecting you to be this fun, open, charismatic. You've been amazing, so thank you so much.

    6. EBErik Bernhardsson1:01:59

       Thank you so much for having me. This was a lot of fun.

Episode duration: 1:02:11