The Story Behind Cerebras’ $63 Billion IPO with Founder and CEO Andrew Feldman

0:00 Netflix used to deliver DVDs in envelopes. And [music] when the internet got fast, they became a movie studio, right? It opened up an entirely new business, something fundamentally different. That's what happens with speed. [music] And I think that's what fast AI does. Right now, we're replacing things that everybody can see, coding, design, the SaaS tools. But once we start sort of fundamentally reorganizing around this, you're going to see this sort of new business [music] models and fundamental jumps in productivity. And I'm eager for that. That's so cool.

0:37 Today in O'Reilly's, we have Andrew Feldman, the co-founder and CEO of Cerebras. Cerebras was founded in the mid-2010s to focus on new workloads for AI, particularly the machine learning world, and then has made the transition into a very fast inference for the foundation model world that we live in today. Cerebras recently went public and is currently worth about $63 billion on the stock market. So, Andrew, thank you for joining us in O'Reilly's. Oh, what a pleasure. It's good to see you guys again. Yeah, so first of all, congratulations. So, um your company Cerebras just went public.

1:08 Um as of today, it's a $60 billion market cap, which is pretty amazing. Pretty amazing. Yeah, and you were I think you were with us a year or two ago on the show in one of the earlier episodes, and it was a pleasure to talk to you then, and obviously we're very excited to have you on today. Could you tell us a bit how the business evolved since that time and what you folks just a reminder for our audience what you do, what you're focused on, how you're moving forward.

1:28 >> We we build AI computers, right? Computers computers designed to and optimized to accelerate AI workloads. And right now, we're the the fastest at inference, not by a little, but but by a lot, 15, 18, 20x faster than GPUs. And so, what happened was um starting in about 2025, AI models got smart enough to be useful. People began using them, and you know, we make AI with training, and we we use it with inference. So, as people began to to use it, it began to to sort of be integrated into their day-to-day work.

2:03 Um speed became fundamentally important, and we were just crushed with demand. Is it Is this faster across the board, or is it specific use cases? Faster across the board. Big model, small models, US models, Chinese models, um trillion parameter models, a 1 billion parameter models, across the board. Mhm. And then what happened was at the end of the year, we signed a a deal with with OpenAI. Sort of one of the biggest deals ever in Silicon Valley. Sort of north of $20 billion.

2:31 And then in March, we signed an agreement with AWS, where we'll be deployed in their data centers going forward. And so, it was just a whirlwind year and a half of to chasing the chasing supply and trying to trying to sort of meet the demand. And what what shifted in the year in in the last year and a half? Was it the ramp in manufacturing? Was it a new chip design? Was it something else? Could you help educate folks on >> what what what happened was we'd built a really really fast machine, and for a long time nobody cared.

3:03 And they right into that's Mhm. because AI >> Actually, forgive me for saying so, but a lot of people objected and said this is just a weird architecture. They They called it wrong. Like Cerebras called it wrong. >> Yeah. Yeah, they they did. I think um to be radically better, right? You you you can't build something that that is a similar architecture, right? You're not going to get 15 or 20 times better than the GPU with a minor modification to their architecture. And that's probably true across the board.

3:34 That if you're going to aspire to a radical improvement, your design has to be different. And from the beginning, you know, we chose Wafer Scale, which means we build a 46,000 square millimeter chip, a chip the size of a dinner plate, whereas everybody else is building chips the size of postage stamps. They told us we were out of our mind, it would never work. They They listed reasons why it was impossible. But in 2019, we we proved it was possible. We began delivering it, and we improved on it, and we improved on it.

4:04 Um but we were fast when I was a novelty. And when it's a novelty, nobody cares if you're fast because it's not being used. And so, from about 2023 to the beginning of '25, sort of people pointed at AI, but nobody used it every day in their work. Mhm. And once you use something every day in your work, you can't be slow. I mean, how how long will you guys wait for a website to resolve? I have no attention >> Right. That's exactly right. That That's exactly the way it is. I mean, how big is the market for slow search? It's zero. How big is the market for dial-up internet? It's zero. That's how big the market for slow inference will be.

4:44 But we had to wait until it was smart enough to be useful, and that happened in 2025. And that's why you got this sort of explosion of of demand and companies like Cognition and Cursorial and Lovable and and just all these others that began ramping extraordinary. Many of the ones you guys have invested in are are ramping like crazy, OpenAI and and and and others. And and we were right there with the right product. I think I first met you back in 2016 or something like that, and at the time, uh people weren't even weren't even like saying AI sounded weird, right? You were talking about machine learning.

5:20 And the models of the time were uh convolutional neural networks and RNNs and you know, just the emergence of GANs and things like that. We were trying to tell the difference between a chair and a cat, right? That was That was great. [laughter] So, his PhD is like a cat and or a chair. It's like, "Woah, look how far we've come." I mean, it's unbelievable. Yeah, yeah. What do you think gave you the foresight to build against the market? Because to your point, I think a lot of us believed in that this market would be really important, and you more than others, right, since you actually started a company in it. But then it took some time for the market to really expand to the point where uh to your point now it's it's this massive use case, people really care about speed of inference and other things.

5:57 Um what gave you the conviction back then to do this? Combination of of vision, um the right co-founders, and a little bit of arrogance, and a little bit of luck. You know, we we saw AI on the horizon as a new workload. And as computer architects, new workloads are opportunity. Right? It's very, very hard to to to enter in the x86 world, right? Where there's not nothing new is happening there, and nothing has happened for generations.

6:25 But you know, when graphics emerged, you got the discrete GPU, and you you you got uh Nvidia, and and when uh when the mobile uh compute hit, you you got ARM. And it was interesting that that not Intel, not AMD, not all sorts of people who you would have thought have been really well positioned to win in that business, they all got no share. And so we knew that that that this new workload would eat a lot of compute.

6:53 It would require uh a new architecture, a dedicated architecture, and that ought to be very different. The architecture could not be a derivative of what's existing. Those were our big bets, and they were 100% contrarian. Mhm. And they turned out to be dead right. >> Were there moments where you just doubted whether this would work, given that it took time for it to work? >> Yeah. We had a period uh you know, we're solving a problem that that had never been solved before. Mhm.

7:20 >> they'd been efforts across the entire 70-year history of the computer industry to build a wafer-scale product. In fact, Gene Amdahl, sort of one of the fathers of our field, one of the the the guys on Mount Rushmore of compute, failed miserably to do it. We had a period between about 2017, middle of 2017, and middle of 2019 where we couldn't build it. We were spending about 8 million a month. Mhm. You having board meetings every 6 weeks saying, "I can't build it. No, still not working."

7:51 And right, "Oof" is right. I mean, that's a huge amount of money. And a huge amount of conviction your investors have. And each time we did a failure analysis, we got a little bit better at at it. We got a little bit better at it. Um and then in the summer of of '19, we we yielded it. And it began to work. And the first time we were sitting in a a little makeshift shift office in downtown Los Altos in a building that was not designed for hardware guys.

8:19 And we're staring at a computer, which is about as exciting as watching paint dry. Mhm. And it's working. And we just we couldn't speak for half an hour. Right? It's like, "Nobody's been able to do this and it's working. And we did this." >> And it was amazing. Cuz that's the technical side of it. And then there's a market side, right? And also on the market side, to your point, it took time to get to the point where these workloads are really important.

8:39 Um so, were there moments where you doubted whether the market existed? You know, we we we we solved it. And we solved this sort of the hardest problem in the computer industry and nobody cared. Nobody. It was like, you know, the first gen we might have sold a dozen. The second gen we probably sold 300. And now we're selling going to sell tens of thousands in the third gen. We had a two or three-year period where we were ahead of the market.

9:05 And absolutely nobody cared that we were blisteringly fast. >> Mhm. And you found some pioneering customers that were like atypical in terms of of starting point, right? There was a there were some sovereigns who really bought ahead. Like, how did you think about being resilient to this period of being ahead of demand? I I think there's a there's a a a path that has been laid down by new computer architectures. And often you begin in the in the supercomputer world because those guys love speed and they don't care if your software is immature. Mhm.

9:34 And so, we we sort of ran the table there. We we wanted National Labs and at Lawrence Livermore and at Sandia and in Europe at European parallel computing center at LRZ. So, we ran the table there and then we we won some guys in in the oil and gas space and we won some guys in pharma, all of whom have long histories of using extraordinary amounts of compute. But then historically there's this giant chasm because none of them provide the volume to get to mainstream.

10:01 And we won out a sovereign G42. Um and they became a strategic partner and and close friends. Um and they placed a billion-dollar order on us. And with that we were able to sort of transform the company. We were able to change our supply chain. We were able to deploy equipment in big enough clusters that that we could battle test at scale. You know, one of the challenges in hardware is your QA lab can't be as big as some of the customers you want to deploy to. Mhm. Right? But you can't put a hundred million dollars in your QA lab worth your own gear.

10:37 And they worked with us and we we began training models for them. We began doing inference with uh for them. They've been an extraordinary partner. This is Peng who's CEO of G42 and his chairman Sheikh Tahnoun. We couldn't ask for better partners. And so we we were able to when OpenAI came along, when AWS came along, we we had the capacity. We were ready. Right? We'd battle tested. We'd sort of gotten over the the chasm. We We'd had a bridge and so we could could meet the the demand. Yeah, I think that kind of path dependence is sometimes undervalued in in this field because the ability for you to go from a, you know, like tens, hundred million-dollar order to twenty billion of backlog, like there's got to be there's got to be something in the middle. It's years of work.

11:23 Yeah. It's years of work and and you know, it's I I think often and I'm sure many of your listeners um are in the software world and and you guys can scale so fast. Mhm. Right? But but when you're building things, right? You you have to you want to double, you got to call your manufacturing partner, your CM. You got to they have to find power. They have to rent a building. They have to add more lines. They have to make test fixtures.

11:48 Right? Each step takes real time and effort to to grow. We're going to try to increase manufacturing 10x this year. Mhm. Right? That's about as fast as anybody in the history of hardware. And so the maturity of the software stack for you guys that's more scale, right? >> You know, when when we started the company, Sarah, our one of my co-founders, >> I do remember. I know. We we presented to you one of my co-founders said, Andres, it's going to take about 10 years to to build a compiler. I said, "No, that's crazy.

12:18 That's big company talk. We can do it in five." Takes about 10 years. >> [laughter] >> It takes a long time to build a compiler. It's an extraordinarily difficult piece of software. And um now we've got good a good software stack. Can I ask you as an aside actually just because you you have for more than a decade believed that this revolution's going to happen, uh how much is all of this um AI-generated coding relevant for Cerebras internally? Uh hugely. I would say that that, you know, eight months ago we weren't spending a thousand dollars an engineer on tokens, and we're probably at 25 or 30,000 right now, and it's ripping.

12:58 I I think it's not useful for everybody. I I think that's truth. I I think there are some some people who have sort of the perfect mindset for it. Right? And you you they are running eight or 10 agents 7 by 24. They've moved their coding style to being one in which they govern agents. Mhm. Whether they think about how to QA, so they've got a QA agent running. They think about how to sort of remedy some of the weaknesses in the coding models, right?

13:25 They're often verbose. They they often cut out comments. So, they've really thought about and it's a type of puzzle that the perfect fit for their mind. And they've gone from being sort of 10x guys to being 100x guys. I think the rest of us, myself included, we're sort of limping along. We we we're trying to figure out how we can make it work for for our different jobs, for being the CEO, for being the CFO, for being accountants, for being in marketing. Um but for a a small number it is such a tool. And then the rest we try and try and show them what what what others are doing, what best practices are. You're about 800 people now? 800 850, yeah. Um it's a lot of market cap per person. I like that, yeah.

14:08 >> Yeah, it's a good good metric overall. Um when you think about where to go from here, you know, making business bigger, strategic directions, like what do you what do you predict? Where where can you go from here? I I I think we >> Besides delivery. Well, when you've got a backlog that's worth the 20 billion, delivery is pretty important every day. Um I I think we have to uh continue to to to sort of be fearless. I I think one of the malaise of companies as they get to 1,000 to 2,000 3,000 people is they stop taking the type of risks that they were taking before, right? You move from being a fearless engineering culture to to to sort of being what can we get in in the time frame in the next rev. And I I I think that's extraordinarily damaging. And we take such pride in doing fearless work. Um we want to hire people who do fearless work. We want to we're going to sort of guard that culture that that says we would much rather fail in pursuit of the extraordinary than succeed in the ordinary. That that is a horrible thing to do.

15:11 And so those are some of the things that worry me. I think recruiting, right? You you have so many openings and it is so easy to settle. And it's so easy to just try and put a butt to see Yeah, pretty good. Let's get that button to see I mean that that is death. And so um we think really hard and I spend a meaningful part of every day and talking to candidates. Those are things that that sort of I I worry about I think about every day. We have a um a lot of founders and leaders who you know, listen to the podcast who are thinking about maybe they have a successful business and they're managing through the period of waiting for the market or trying to figure out if they're still right. They think about how to hire from 800 to several thousand. Um there we talked about the managing of your own psychology when you're like, am I right for this decade?

15:58 How how do you like keep and motivate employees when there wasn't external feedback for this long period of time? Well first I I I have empathy for them. I mean being CEO is an extraordinarily lonely thing. And uh you you're building a business. You're building a business. You guys know this that that being a leader is is lonely. And it's not easy. And people don't like to say that. Especially for those of us who like to solve problems, specifically the problems everyone else says can't be solved. You sort of you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you you Right.

16:41 [laughter] Right. That's I thought that was just my No, that's right. That's exactly right. You know, um you know, you you were at a top venture firm. You wanted to do it your way. Right? And so you stepped out and doing it your way. And you said to yourself, I can do this. And it's not easy. And um that that's one thing. The The other thing is you have to love the journey, right? The this things we do are too hard if you don't like the building.

17:09 Right? That you do this for the money is is a horrible thing. There are way easier ways to make money than than trying to create something extraordinary and compete with somebody as strong as as Nvidia. That is not the easiest path. You You got to love being a David. Right? I'm a professional David. This is my fifth startup. I compete against Goliath. Um that is what I do for a living. And I I I think to myself that every dollar, every million dollars, every billion dollars we sell, if it wasn't for our brains, their muscle would have taken it in a heartbeat.

17:42 And you got to love that. And And if you don't love that, it's a it's a very long road. When do you think cuz there's sort of two views of the world in terms of when to give up on something? And you know, one argument is just keep going no matter what, and you know, hopefully things work out or eventually Well, the other view of the world is you know, you should be constantly reassessing whether the journey you're on is the right one.

18:05 And there's some moments where actually giving up is the smartest possible thing you can do. Um what's your view on that? Or how do you think about when's the right time to give up on something? I think it is clearly this the the right time to give up when uh you've laid out a set of hypotheses about what it's going to take >> Mhm. to to win, and they all come back negative. Yeah, but I see people kind of do this sequentially, right? They say, "Oh, I just need to test one more thing." And they test and it doesn't work. And they say, "I need to test one more." And so The The The slippery slope is a beast.

18:35 The slippery slope in in all things. In ethical situations, in your life, I mean, the slippery slope is really um something you have to guard against, right? And I I think sometimes having other former CEOs or other really seasoned entrepreneurs who are on your side, and who can share with you, "Remember a year ago you said if you got to this point and you didn't have this." And to remind you, so so they pull you back off that slippery slope, right? They They said, you know, the old frog in the the warm water thing is like you said if it got this hot you were going to get out.

19:16 And it slowly kept getting warmer. >> can other people keep you effectively accountable to your instructions. >> to your own thinking. >> Yeah. Um if you understand why it's not working. Right? If there are some things that that you can articulate that have to change >> Yeah. Um in order for it to work and and you can put some sort of time frame on it. Um but that is an extraordinarily hard question. And I I think uh it's probably the case that that lots of of of efforts ought to be truncated. Mhm. Yeah.

19:55 >> And those people sort of redeploy their efforts to new and different ideas that they have. Yeah, it's kind of like I have you heard of opportunity cost in life? And for some people it's the best moment of their lives that in terms of productivity or things I could do. And so you know, the cost of time is extremely high. Um you know, in your guys' case obviously it worked out. What made you all decide to go public? Similarly, there's differing opinions on when to go public, why to go public, what's the benefits, what's the drawbacks. What what was that in your mind and what made you decide to go out now?

20:21 First, uh sort of go going public is exchanging some professional investors venture capitalists who specialize in technology investing for a different class of investors. And and in so doing reducing your cost of capital a little bit, right? This is this is really what's happening. Mhm. Um suddenly we go from pros like you to my dad. Right? That that that's sort of the trade-off. Um and in return for that uh you have to agree to to be governed by a set of extraordinarily stringent rules. I think your question is complicated by the fact that there have been for the first time in history four or five companies that can raise huge amounts of money without going public.

21:03 That this was never a thing before OpenAI and Anthropic and uh maybe Databricks. Option package timeline from for Silicon Valley comes from. It's like a four-year timeline. >> Yeah, it used to be how long it would take you to get public. Right, it used to be right, it used to be four years and that was the way you got a valuation in the hundreds of millions, right? But I I think >> have a tender cycle. That's right. And at a certain scale. It took us 10.

21:33 Um and I I I think that changes a lot. Right, what we did is we opened up the secondary market and let people sell, right? If you're going to bet big chunks your career with us, we thought it would be perfectly reasonable for you to to find modest liquidity as you went along. I think you have to think very differently if it's going to take you a decade. But I think for a very small number of companies, those three in particular, they've been able to raise sort of public market money at public market valuations in the private market.

22:05 Um I think for the rest of the world, if you want super high uh valuations, if you want uh the legitimacy that comes with it. Um historically large companies like doing business with other public companies in the US and you get a credibility and a legitimacy from having your books audited, from them being able to see who you are that that is different than when you're private and I I think all of those are are reasonable reasons.

22:34 I also think we could offer the public market something unique, right? We would be the first and only for a period of time AI pure play. We we are the only company that that you can 100% of the revenue this exact market. Um there's no gaming, there's no graphics, there's no PC that this is it. And that was an opportunity, a differentiator that we thought was interesting. I I think there ways around all the other things. You can deliver returns to your investors. I I think both Elon and Ali have been really creative about allowing employees to sell and and allowing investors who who have 10-year funds to to to find some liquidity in the process.

23:18 But I I I think more than anything um for us, it was an opportunity to to graduate from corporate adolescence to corporate adulthood. Can you talk a little bit about I'm so curious like how did the OpenAI deal happen? You know, what were um what do you think was the point at which you you knew that you were a good fit for them? I I think I spoke to Sam in in sort of middle of summer in in 25.

23:48 And he said for the first time he said we're we've been trying so hard just to keep up with demand. We now see the importance of fast inference. That produced a a set of trials and some testing that that was done. Um and we were so much faster than the than the competition. It felt really good. And when what we love talking to super smart customers, right? I mean, I I can't I know you do consumer, too. I I can't do consumer. I have a rule that if my my mother buys it or uses it, I don't want to make it or sell it. Um cuz I >> [laughter] >> I I really want super smart customers who are doing really interesting things with our stuff. And so we got in with um some of their guys and they were like, "Whoa.

24:34 This is We understand now." And at Thanksgiving, the night before Thanksgiving, we signed a term sheet. And you know, 4 weeks later on the 24th of December, we signed a a big master agreement. And so um >> incredibly fast. >> You know what? They can fly. And uh you know, we were working 7 days a week. I mean, they had several law firm I mean it it was a huge for a 20 plus billion dollar deal to do it in 4 and 1/2 weeks was was exceptional. I actually think that's like a crazy characteristic of this market that I've not personally experienced before, which is everybody's trying to keep up with demand. I And and I think, you know, I I talked to the guys at uh at Cognition, right? They bought Windsurf over a weekend.

25:19 Right? I I think many of the things that we thought were speed of light weren't. Right? Could be done much faster. And I I think, you know, the rate at which Elon has been able to build data centers. Everyone said, "Oh, you can't do it that way." Except if you're him, in which case you can. Or you can't buy a $300 million company in three Actually, you can. You can't do a deal like this in in 24 days. But if you work on it every day, 8 or 10 hours a day, you can. And I think the the art of the possible has been expanded by by this push, you know, in a way I never would have expected. And it is a huge advantage to have the ambition for speed if you believe it is possible.

26:00 That that's right. I I think we have seen some extraordinary operators in this market build amazing things, right? I mean, uh the guys at Cursor or Cogni- You see sort of growth we've never seen before. You can't grow that fast. Well, actually, you can. Um you can't build data centers. You can't do deals. It just Those were sort of truncated aspirations, which is interesting. Speaking about these companies like Cog and Cursor and such, uh the growth of the open source ecosystem has enabled a generation of companies to do really impressive things. Like >> Super. Super impressive. You know, Devin on Cerebrus is a really magical experience. Coding on Cerebrus is like like high performance at massive speed is really special. Um how do you how do you think about, you know, open source and post-trained workloads and and your perspective on that going forward.

26:56 They have fed this market, right? When close-source was was too expensive, the open-source community had has sort of kept the interest alive and kept the flame going. And I I think that that the and pushed the the close-source guys. I I think that the sort of techniques that we saw by uh some of the Chinese makers, like, "Whoa, we we got to stay ahead of that, right?

27:29 We we can't rest on our laurels. We we can't depend on the fact that we have uh bigger training clusters and more data." Um and I think that's made for an extraordinarily vibrant ecosystem, right? I I think it's uh made for creativity uh or and allowed creativity to to to to take root and and really produce interesting results. And that's fun to be in the mix of, right? It's fun to see other people's ideas do interesting things on your hardware. And that's if you don't love that, your infrastructure's not right for you. You got to love other people's ideas to take flight on on what you built. Uh when you think about um experiences you imagine will be possible only on Cerebras? Is there anything you're excited about in a couple years from now that we should look out for?

28:19 You know what? When I think about what speed does, um it it it doesn't make the existing business models a little better, right? Um No, Netflix used to deliver DVDs in envelopes, and they thought their competition was Blockbuster. And when the internet got fast, they became a movie studio. Right? That's what happens with speed. I mean, it wasn't a They didn't get better incrementally and more efficient to delivering DVDs, right? It opened up an entirely new business.

28:54 Something fundamentally different. And then they sort of became a movie movie studio. They bought existing movie studios. And and I think that's what what fast AI does is it will present entirely new sort of business models that that are available. I think the easy and the obvious is to replace existing. And we we know that that that when the the PC came in it replaced, right, typewriters and general ledger accounting. But the big jump in productivity was when it reorganized so we did work. And you got the cloud and then with the cloud you were able to get SaaS. And with SaaS you were able to get tools that you previously couldn't afford cuz they were so expensive to the individual company and to the small number of seats, right?

29:41 Then you got this massive jump in productivity. And I think AI's in the same way that right now we're we're we're replacing things that that that everybody can see, coding, design, right, some of the the SaaS tools. But once we start sort of fundamentally reorganizing around this you're going to see this sort of new business models and fundamental jumps in productivity. I'm eager for that. That's so cool. Very exciting. Thank you so much for joining us today. Guys, thank you so much for having me on your show. Really appreciate it. Congratulations. Thank you so much.

30:16 Find us on Twitter at No Priors Pod. Subscribe to our YouTube channel if you want to see our faces. Follow the show on Apple Podcasts, Spotify, or wherever you listen. That way you get a new episode every week. And sign up for emails or find transcripts [music] for every episode at no-priors.com.