Netflix used to deliver DVDs in envelopes and 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 and I think that's what fast AI does. Right now we're replacing things that everybody can see, like 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 models and fundamental jumps in productivity and I'm eager for that. That's so cool.
Today on No Priors, 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 very fast inference for the foundation model world that we live in today. Cerebras recently went public and is currently worth about $63 billion in the stock market. So Andrew, thank you for joining us in No Priors.
Oh, what a pleasure. It's good to see you guys again.
Yeah. So, first of all, congratulations. So, your company, Cerebras, just went public. As of today, it's a $60 billion market cap, which is pretty amazing.
Yeah. And you, 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 going forward.
We build AI computers, right? Computers designed to and optimized to accelerate AI workloads. And right now we're the fastest at inference, not by little, but by a lot, 15, 18, 20x faster than GPUs. And so what happened was 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 use it with inference. So as people began to use it, it began to sort of be integrated into their day-to-day work. Speed became fundamentally important and we were just crushed with demand.
Is it faster across the board or is this specific use cases?
Faster across the board. Big models, small models, US models, Chinese models, trillion parameter models, one billion parameter models, across the board.
And then what happened was at the end of the year, we signed a deal with OpenAI, sort of one of the biggest deals ever in Silicon Valley, sort of north of 20 billion. And then in March, we signed an agreement with AWS where we will be deployed in their data centers going forward. And so it was just a whirlwind year and a half of chasing the supply and trying to sort of meet the demand.
And what shifted 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 happened?
What happened was we built a really, really fast machine and for a long time nobody cared. Right. That's because, actually forgive me for saying so, but a lot of people objected and said this is just a weird architecture. They called it wrong. Like CUDA called it wrong. Yeah.
I think to be radically better, right? You can't build something 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 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 listed reasons why it was impossible. But in 2019, we proved it was possible. We began delivering it and we improved on it and we improved on it. But we were fast when AI was a novelty and when it's a novelty, nobody cares that 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.
And once you use something every day in your work, it can't be slow. I mean, how long will you guys wait for a website to resolve?
I'll have no attention, right? That's exactly right. 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. 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 demand and companies like Cognition and Cursor and Lovable and just all these others that began ramping extraordinary. Many of the ones you guys have invested in are ramping like crazy. OpenAI and others 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 people weren't even like saying AI, sounded weird, right? We were talking about machine learning and the models of the time were 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 quite a great PhD, is like a cat or a chair, like whoa, look how far we've come. 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 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, to your point, now it's this massive use case. People really care about speed of inference and other things. What gave you the conviction back then to do this?
Combination of vision, the right co-founders and a little bit of arrogance, a little bit of luck. You know, 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 enter in the x86 world, right? Where there's nothing new is happening there and nothing has happened for generations. But you know when graphics emerged, you got the discrete GPU and you got Nvidia and when mobile compute hit, you got ARM. And it was interesting that not Intel, not AMD, not all sorts of people who you would have thought really well positioned to win in that business, they all got no share. And so we knew that this new workload would eat a lot of compute. It would require a new architecture, a dedicated architecture, and it 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.
And they turned out to be dead right.
Were there moments where you just doubted whether this would work given the different...
We had a period, you know, we're solving a problem that had never been solved before. Mhm. I mean, there'd been efforts across the entire 70-year history of the compute industry to build a wafer scale product. In fact, Gene Amdahl, sort of one of the fathers of our field, one of 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're having board meetings every six weeks saying, 'I can't build it. No, it's still not working.' And right, 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 it. We got a little bit better at it. And then in the summer of 19, we yielded it and it began to work. And the first time we were sitting in a little makeshift office in downtown Los Altos in a building that was not designed for hardware guys and we're staring at a computer which is about as exciting as watching paint dry and it's working and we just we couldn't speak for half an hour, right? It's like nobody had been able to do this and it's working and we did this and it was amazing.
Because 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 were really important. So were there moments where you doubted whether the market existed?
You know, 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 still going to sell tens of thousands. And the third gen, we had a 2 or 3 year period where we were ahead of the market and absolutely nobody cared that we were blisteringly fast.
And you found some pioneering customers that were like atypical in terms of a starting point, right? 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 think there's a path that has been laid down by new computer architectures and often you begin in the supercomputing world because those guys love speed and they don't care if your software is immature. And so we sort of ran the table there. We won at Argonne 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 won some guys 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. And we won a sovereign, G42, and they became a strategic partner and close friends and they placed a billion dollar order on us and with that we were able to sort of transform the company. We're able to change our supply chain. We're able to deploy equipment in big enough clusters 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.
Right. I mean, you can't put $100 million in your QA lab worth of your own gear. And they worked with us and we began training models for them. We began doing inference for them. They've been an extraordinary partner. This is Peng, who's CEO of G42 and his chairman, Shiekh Tahnoon. We couldn't ask for better partners. And so we were able to, when OpenAI came along, when AWS came along, we had the capacity. We were ready, right? We'd battle tested. We'd sort of gotten over the chasm. We'd had a bridge and so we could meet the demand.
Yeah. I think that kind of path dependence is sometimes undervalued in this field because the ability for you to go from a, you know, like tens or hundred million dollar order to 20 billion of backlog, like there's got to be something in the middle as years of work.
Yeah, it's years of work and you know, I think often and I'm sure many of your listeners are in the software world and you guys can scale so fast, right? But when you're building things, right, 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. You have to make test fixtures. Right. Each step takes real time and effort to grow. We're going to try to increase manufacturing 10x this year.
Right. That's about as fast as anybody in the history of hardware.
And it's also maturity of the software stack for you guys at more scale. Right. You know, when we started the company, Sarah, our one of my co-founders, I know we presented to you, one of my co-founders said, 'Andrew, it's going to take about 10 years to build a compiler.' I said, 'No, that's crazy. That's big company talk. We can do it in five.'
Takes about 10 years, right? It takes a long time to build a compiler. It's an extraordinarily difficult piece of software. And now we've got a good software stack.
Can I ask you as an aside actually just because you have for more than a decade believed that this revolution is going to happen. How much is all of this AI generated coding relevant for Cerebras internally?
Hugely. I would say that 8 months ago we weren't spending $1,000 an engineer on tokens and we're probably at 25 or 30,000 right now and it's ripping. I think it's not useful for everybody. I think that's the truth. I think there are some people who have sort of the perfect mindset for it, right? And they are running eight or 10 agents 7 by 24. They've moved their coding style to being one in which they govern agents. 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? They're often verbose, 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're trying to figure out how we can make it work for our different jobs, for being the CEO, for being the CFO, for being accountants, for being in marketing. But for a small number, it is such a tool. And then the rest we try and try and show them what others are doing, what best practices are.
You're about 800 people now.
