Andrew Feldman on Building Cerebras and the Future of Chips | Ep. 57

1. 0:00 – 1:07

   Intro

   1. AFAndrew Feldman0:00

      You have board meetings every six weeks. All you've got to say is, "Still can't make it." [laughs]

   2. JAJack Altman0:05

      [laughs]

   3. AFAndrew Feldman0:05

      Right? Ex- that, that's the board meeting. I mean, what else we gonna talk about? Still can't make it again and again.

   4. JAJack Altman0:10

      [laughs] And Eric's like, "Should I get in? C- can I do anything to help?" [laughs]

   5. AFAndrew Feldman0:13

      That's right. [laughs] They're like, "Can we help?" It's like, "Nope." But we always believed that if we could make it, there would be huge demand for it.

   6. JAJack Altman0:21

      [upbeat music] All right. Super excited to be doing this today. Andrew, thanks a bunch for, for being here. So Andrew, you're the, you know, CEO and founder of Cerebras, which is obviously one of the most important chips companies, and also really happy to have my partner Eric here, who's gonna be kinda half host, half guest, um, who, you know, has been working with Andrew since the beginning. But thank you both for, for making time. What I wanna start with is it's 2026 now, and obviously Cerebras is very important in the sort of AI landscape. But you started the company in 2016, and AI was not what it is, you know, now back then. So I guess what was the kind of head space then? What was the idea, the insight? Like, what were you, what were you building, um, when you started the company, and what was the thinking?

2. 1:07 – 2:54

   Why Andrew started Cerebras in 2016

   1. AFAndrew Feldman1:07

      Well, Jack, thanks for, for having me. It's always fun to hang out with Eric, and so appreciate you having me. Um, I think as, as a computer architect, when, when you see new, a new workload on the horizon, you get excited, right? It, it's very, very difficult to win share in a mature market in compute. And so when, when something new emerges, um, you get excited, and you lean forward, and you say, "Well, can I make it faster? Can I build a, a chip that, that, that is better at this work?" And then you ask the other question is, is there enough of this work to justify building a chip that's better? So you have two questions. Can I, and should I? [laughs]

   2. JAJack Altman1:49

      [laughs]

   3. AFAndrew Feldman1:50

      A- a- and what we saw with AI was, uh, an extremely computationally intensive workload, unlike, for example, the r- the rise of Arm, uh, processors for cell phones, that they weren't computationally intensive. They were power intensive, right? We, we saw a problem that was gonna be hard on compute. And, uh, we, we, we saw that, that what was being used w- was an architecture that was, that was being re- remodeled for it, and that, that we could build something better. A- and that was sort of the, the insight, and then it was a, a long, hard slog. [laughs]

   4. JAJack Altman2:30

      [laughs] Before we go to the long, hard slog, Eric, obviously for you, like as an investor, you know, you were not, you know, also in this AI era. But you saw something that made you invest, and you were, I think, you know, previously not a chips investor, and you'd done more like-

   5. EVEric Vishria2:44

      And probably never will be again. [laughs]

   6. JAJack Altman2:46

      Probably never will be again. Like, would you do that every time again? Was there something that made it clear to you, or was it... Like, what, what happened for you when you made the investment?

3. 2:54 – 4:10

   Eric on why he invested despite having no chip experience

   1. EVEric Vishria2:55

      I, I, I think one of the scariest things about venture, and I think about this all the time, is the more you work on companies and you see the challenges and you see how hard it is, you build up scar tissue. And then the thing is like, well, you should do it again, but like, will you do it again? And I ask myself that all the time. Like, there, you actually need some naivete. Um, like, you need to have this like, oh, can-do attitude and this like naivete around it to attempt it. Like, otherwise y- you just don't do it again, and y- I've definitely built up a, a bunch of scar tissue. Um, and I don't know had I actually had any idea how hard it was gonna be for them, like how much science and technology-

   2. JAJack Altman3:44

      [laughs]

   3. EVEric Vishria3:44

      ... had to be built, how many innovations, how many times, like, the company was gonna come to the brink-

   4. AFAndrew Feldman3:50

      Don't tell your VCs how hard your shit is.

   5. JAJack Altman3:52

      [laughs]

   6. AFAndrew Feldman3:52

      Man, the moral of the story is tell them it's no problem.

   7. JAJack Altman3:55

      Keep it inside.

   8. AFAndrew Feldman3:55

      Ke- ke- keep inside. Push it down.

   9. JAJack Altman3:57

      Push it down. [laughs]

   10. AFAndrew Feldman3:58

       Push it down. Just don't, don't tell them-

   11. JAJack Altman4:00

       I'm doing great

   12. AFAndrew Feldman4:00

       ... this is a really hard problem. All's good, man. Shh.

   13. JAJack Altman4:02

       Yeah. So what happens? So you basically, you know, you s- you start the company, you get to the tape out, and then you're done, right? It's like that?

   14. AFAndrew Feldman4:08

       N- n- no. Never.

   15. JAJack Altman4:09

       Okay.

   16. AFAndrew Feldman4:09

       Um-

   17. JAJack Altman4:09

       So how does it go?

4. 4:10 – 9:44

   Attacking Goliath

   1. JAJack Altman4:10

      So what, what happened? Yeah.

   2. AFAndrew Feldman4:10

      I, I think, uh, we, we had a... You know, this, this isn't our first chip company. Th- this, uh, this is my fifth. And the team had been together, uh, the founding team had been together at the last one. And we, we had some really c- clear sort of philosophies on, on, on what you should do. And I, our view is that if y- if you're gonna attack Goliath, if, if there's a, a, a giant standing in the market, l- like Nvidia was even at that time, that being a little bit better or a little bit cheaper is, is not an available strategy, right? They had high margins, so if you come in, and, and even if you're two-thirds the price, uh, they can just cut costs. They, they can just charge less. They can bundle it. They can do 100 other things. What that means is you have to go out with something way better. 10, 100, 500 times faster. You have to come up with a product that has the value proposition that, that even if the other guys give it away, all right, they can't compete. And so that, that's sort of one direction of our thinking. And the other, building off that, is, is w- we want to do hard things that produce that advantage, right? That we want all that hard stuff within the building because that's under our control. All the other stuff's not under our control, but if we can build something that is so fast nobody else can do it, you can't give away your parts to achieve what we can achieve, um, then you're onto something, and the only way to do that, in our opinion, is radical innovation.

   3. JAJack Altman5:46

      Hmm.

   4. AFAndrew Feldman5:47

      You, you can't incremental your way to vastly better.

   5. JAJack Altman5:51

      So what has to be in your control?

   6. AFAndrew Feldman5:53

      You, you have to have all aspects of the design, the implementation. You have to understand the manufacturing. You have to understand every part. And for us, that meant chip Board, system, software, all the way up to the API. And, uh, that was not easy. That makes it more expensive. It makes it take longer. It means there are more, more ways to fail there. When you do radical innovation, there are no vendors waiting for you, right? When you build a, a chip the size of a dinner plate, you, you can't go to a catalog and find a heat sink, right? Because nobody'd ever built one bigger than, than a postage stamp. So nobody has stuff ready for you, and so you, you end up investing an enormous amount of time in, in building things that support your thing, right? The, the surrounding components. But the result of that is, is you develop this extraordinary expertise. We didn't start world leaders in packaging. We're right now the best in the world at packaging. Um, we earned it, failure after failure, year after year, until we got it. And so what you wanna do is you, you wanna think about sort of h- how you can innovate a- across the, the, the various elements of, of a full solution, and push as hard as you can.

   7. EVEric Vishria7:20

      I, I think, like, Andrew taught me this in, in, in semis and everything else, just to, like, make this, like, what you need to be shooting for concrete. So if, if you think of whoever the incumbent is in whatever market, you know, they're getting, say, twice as good every year in a market like this. It's very dynamic. For a new company to get to scale, to start to get to scale, it's gonna take five years at least. Let's say with everyone moving as fast as they can, AI tools, you nail everything, first tape out works, first bring up works, you know, everything goes right, five years, okay, to, to, like, to get to scale. And so you're basically at two to the fifth, so that puts you at 32x. And then you need at least a multiple advantage of that, so say 3x. So you're basically at 100x. So one of the things that happens when you hear a lot about these new ideas is they're like, "Hey, we're gonna be 50% better. We're gonna be two times as better. We're..." But what they are... They're like, "Once we have this design, it's gonna be even 10 times better than what exists today." But what exists today isn't the target because if you're looking at the companies that are out there right now, these big companies with a lot of resources are actually doing, they're, they're doing great work. Everyone's... Nvidia's advancing the ball materially every year, like, very significantly.

   8. AFAndrew Feldman8:42

      Nvidia and Google and Trainum, they're, they're moving, eight-

   9. EVEric Vishria8:45

      The TPU team

   10. AFAndrew Feldman8:46

       ... AMD, I mean, they're moving the ball.

   11. EVEric Vishria8:48

       And-

   12. AFAndrew Feldman8:48

       So you gotta aim at, at 100, 500, 1,000 times better if you're gonna arrive-

   13. EVEric Vishria8:54

       Yeah

   14. AFAndrew Feldman8:54

       ... if you're gonna intersect the market ahead of them.

   15. EVEric Vishria8:57

       And so that, that, just, like, when you realize that, you're like, "Oh, wait, it can't be small changes in something. There has to be a very big underlying, like, architectural change, a big thought." Like, there has to be something materially different, whether it's Wafer Scale, SRAM, whatever it is, like, there, those things matter.

   16. AFAndrew Feldman9:19

       You, you can't get there with a collection of modest improvements.

   17. EVEric Vishria9:22

       Mm-hmm.

   18. AFAndrew Feldman9:22

       Right? Your, your, your biggest competitor buys silicon for less than you. They buy manufacturing capacity for less than you, right? [laughs] They, they have, they probably pay less for their EDA tools. And so you can't go at them, you know, you can't run at-

   19. EVEric Vishria9:37

       Can't go straight up

   20. AFAndrew Feldman9:37

       ... Goliath straight on. I mean, you gotta think about how you can deliver something profoundly different.

5. 9:44 – 10:44

   Near-death experiences and the Valley of Death

   1. JAJack Altman9:44

      What were the, like, along the way, after some of these, like, early moments, I know that the company had, like, you know, some near-death experiences or whatever. Like, what were the, what were sort of the hard hurdles to get through?

   2. AFAndrew Feldman9:55

      I had some near-death experiences.

   3. JAJack Altman9:57

      [laughs]

   4. AFAndrew Feldman9:57

      Pain in my chest when, when, when, when you can't build the thing you're supposed to build. Yeah.

   5. JAJack Altman10:02

      But, like, what are those per- Like, how would you frame what those periods are where you're like, "We just need," I don't know, was it a certain amount of time and a certain amount of capital, and-

   6. AFAndrew Feldman10:10

      No

   7. JAJack Altman10:10

      ... it's just gonna be Valley of Death? Is it that you just don't know whether you're gonna make a technical breakthrough? Like, what are the things that lead to those-

   8. AFAndrew Feldman10:17

      I, I think i- in our space, we were always confident if we could make it, we'd sell it, right? That, that there are sort of two axes in our life. There is can you make it, and the other axis is can you sell it, right? And w- we always believed that if we could make it, it, there would be huge demand for it. I think nobody had ever done Wafer Scale.

6. 10:44 – 12:16

   18 months of "still can't make it"

   1. AFAndrew Feldman10:45

      There was a, a whole sort of collection of people who were saying it could never work. And, uh, we were unsure, too. I mean, we believed we could do it, but there was no evidence. And there was a period of time, about 18 months, and we couldn't make it. And we're spending about $8 million a month. You have board meetings every six weeks, and all you've got to say is, "Still can't make it." [laughs] Right? That, that's the board meeting. I mean, what else are we gonna talk about? I mean, still can't make it, again and again and again. And, and I, I give a hu- huge, right, right.

   2. JAJack Altman11:21

      And Eric's like, "Should I get in? Can, can I do anything to help you?" [laughs]

   3. AFAndrew Feldman11:24

      That's right. And they're like, "Can we help?" I was like, "No." [laughs]

   4. JAJack Altman11:27

      You can't.

   5. AFAndrew Feldman11:28

      Um, there's nothing, nothing that can be done.

   6. JAJack Altman11:30

      Nothing to be done.

   7. AFAndrew Feldman11:30

      Nothing to be done.

   8. JAJack Altman11:31

      No.

   9. AFAndrew Feldman11:31

      But we believed because we, we had ideas still. We hadn't run out of ideas. And each time we, we built it and it failed, we'd go through sort of good engineering practice. We'd do a full failure analysis. We'd understand it. We would, uh... And we wouldn't fail that way again. So we-

   10. EVEric Vishria11:53

       All new mistakes.

   11. AFAndrew Feldman11:54

       All new mistakes. So we, we sort of had a, th- that, that was sort of our mantra, only new mistakes, right? Only new failures. Um, and over time, we, we could see progress. I mean, in the beginning, we were shattering wafers, right, in seconds. And then it took minutes, and then we had one run for, for an hour, and then we

7. 12:16 – 13:26

   Solving a 75-year-old compute problem

   1. AFAndrew Feldman12:16

      shattered some in minutes again. [laughs] And then we, we, we built back up to the point, we had a day in sort of July of, of 2019 where w- we were running and temperature was, was flat, and we just stood there in a tiny little office that had been converted into a lab, and we'd drilled a hole in the wall to suck the air out, and, um, stared at a server, which is about as exciting as looking at, at paint dry, and said, "Holy crap, we've solved this problem that nobody in 75 years of compute had ever solved." And, um, that was one of the great minutes of my life.

   2. JAJack Altman12:52

      Hmm. That's cool. When you look forward, like if you... So, you know, there's the wafer scale, and then there everything around that, the packaging and, and, and, and cooling and powering it, all these things. But, you know, when you look at the R&D envelope that you have going forward over the next five years-

   3. AFAndrew Feldman13:14

      Right

   4. JAJack Altman13:15

      ... I, I know there's, like, things you're excited about, but, like, where do you ta- like, uh, where do you take this, like, giant thing-

   5. AFAndrew Feldman13:22

      So-

   6. JAJack Altman13:23

      And, like, what could be the next wafer

8. 13:26 – 16:19

   What comes after Wafer Scale

   1. JAJack Altman13:26

      scale-

   2. AFAndrew Feldman13:26

      Yeah

   3. JAJack Altman13:26

      ... innovation?

   4. AFAndrew Feldman13:27

      I, I think i- if you simplify, uh, what we build down to its most fundamental elements, um, a computer is built of, of, of three things. Um, we do calculations, we store the results, and then we move the results to where they're useful. So we, we build a core that does the calculation. Um, we use memory, and we move data to and from memory, where we store the results. And then we have IO, and that, that, that's how we ship the results to somewhere where it's useful. Um, I, I think if, if you're in the computer business like we are for AI, you better be working on all three. You gotta be thinking about how to make your cores faster, how to make them tuned for, uh, AI, but general enough to, to withstand the innovation happening in AI. You better be thinking about memory and both capacity, how much you can store, and how fast you can get data on and off it. And then you gotta be thinking about how you get the results off your chip and somewhere where they're useful, and that's your IO. And so w- we're working on all three of those, and we have programs w- with the, with the US government already, big programs where we're thinking about how to stack memory, right? How, how to put, uh, HBM o- o- onto an SRAM-based wafer. And what this does is it, it enables HBM to behave like SRAM, which is exactly what everybody wants. You get the, the capacity of HBM and the speed of SRAM, right? We're, we're working on, on, uh, optical wafer stacking. So you'd put a, an optical switch onto a wafer. This would change the world. Um, you know, Jack Dongarra said, and he, he's sort of one of the, the pioneers in, in big computing, he said, "We've, we've been better at making flops than moving flops," right? [laughs] A- and that, that, that's, that's the IO part. And so thinking about how to, how to solve that problem by bringing optical switching smack up against compute i- is something that, that we also have large government contracts for and we're enormously excited about. So, you know, you gotta get faster, you gotta find ways to store more and get to memory faster, and you gotta find ways to move that, move those results a- at 10, 100, 1,000x faster to where they're useful.

   5. JAJack Altman15:59

      I wanna go to a bit broader in the supply chain. And, you know, I, I think it's kind of, like, well understood right now that AI is very supply constrained. But, um, I, like probably many other people, I wouldn't say I have, like, a perfect understanding of the supply chain.

   6. AFAndrew Feldman16:16

      Oh.

   7. JAJack Altman16:16

      And one of my favorite things to do on this podcast is you get a, you know,

9. 16:19 – 22:30

   The chip supply chain explained

   1. JAJack Altman16:19

      successful person like you, and because there's cameras on, I get to ask you a really simple question, and I get to, you know, be humored with it. So, uh, could you explain sort of in a somewhat simple way, how do we go from, like, sand to a ChatGPT answer?

   2. AFAndrew Feldman16:34

      Dude, it's just TSMC.

   3. JAJack Altman16:35

      [laughs]

   4. AFAndrew Feldman16:36

      That's it. There's nothing to it.

   5. JAJack Altman16:38

      We're all talking about-

   6. AFAndrew Feldman16:38

      There's a black box. We call that TSMC.

   7. JAJack Altman16:40

      The supply chain.

   8. AFAndrew Feldman16:41

      Fed by another black box-

   9. JAJack Altman16:42

      This-

   10. AFAndrew Feldman16:42

       ... called ASML.

   11. JAJack Altman16:43

       [laughs]

   12. AFAndrew Feldman16:43

       And out the other end comes-

   13. JAJack Altman16:45

       You know-

   14. AFAndrew Feldman16:45

       ... comes great chips

   15. JAJack Altman16:46

       ... we're supply constrained, but I don't know what the supply... [laughs] Can you teach me what the supply, how it works?

   16. AFAndrew Feldman16:50

       Yeah. I, I, I think the first thing to, to, to think about is that, that, that a fab, and especially a, a fab that, that builds at cutting-edge geometries, is a, is a modern pyramid. It's one of the greatest things humans make.

   17. JAJack Altman17:05

       What is it?

   18. AFAndrew Feldman17:06

       It is a collection of machines that, that take, uh, a chunk of silicon and use a photolithographic process in which they etch transistors into that piece of silicon, such that when you deliver power to it, they do calculations. And it, it is a collection of... I mean, this is a factory. It's just a reasonable way, a real factory. Um, it costs $40 or $50 billion to make. It has a four or five-year lifetime.

   19. JAJack Altman17:36

       How big is it?

   20. AFAndrew Feldman17:38

       Football fields.

   21. JAJack Altman17:39

       Okay.

   22. AFAndrew Feldman17:39

       When they, when they, when, uh, Samsung was building a factory in Texas, you, they began building a power plant. The power plant was used to make concrete. They ran concrete trucks, hundreds of concrete trucks, 7 by 24, for years to pour enough concrete to build the foundation on which to put the-

   23. JAJack Altman17:57

       Mm-hmm

   24. AFAndrew Feldman17:58

       ... the fab, [laughs] right? The, the, these are unbelievably complicated things.

   25. JAJack Altman18:03

       So ASML makes the machines?

   26. AFAndrew Feldman18:05

       ASML makes, makes, uh, the, the machine that does the photo lithography. Um-

   27. JAJack Altman18:11

       What's that machine like?

   28. AFAndrew Feldman18:13

       It's the size of, uh... Each machine is, what is it? 50 or 60 feet long and 20 feet high

   29. JAJack Altman18:20

       Costs what? Half a billion?

   30. AFAndrew Feldman18:22

       Yeah, they're, they're expensive. Um, what's interesting is that they sell the same machines to different fabs, and fabs use them in different ways and are able to do different things with TSMC able to, to, to, to achieve things that others can't.
10. 22:30 – 25:51

    Why the US punted a strategic industry

    1. AFAndrew Feldman22:30

       you can't just knock them out. It's not a cookie cutter. A- and that's why we, we don't have enough of them right now. And the fact, in the US, sort of three decades of bad policy that pushed the fabs away. When the fabs left, the tool vendors, the, the collection of, of vendors who, who, who provided services to them, they all left. The next step in the process, called packaging, and those are companies like Amkor and ASE, they left.

    2. JAJack Altman22:59

       Mm-hmm.

    3. AFAndrew Feldman22:59

       And we just punted a, a strategic industry, and we, we gotta do better. That, that was... That's not smart.

    4. JAJack Altman23:06

       Yeah. So what do you think should happen there? Like, like, c-

    5. AFAndrew Feldman23:08

       Well, I think, I think-

    6. JAJack Altman23:09

       Yeah

    7. AFAndrew Feldman23:09

       ... we should sit down with GlobalFoundries and TSMC and Samsung and have a 20-year period where we waive all local ordinances to allow them to build fabs. I, I, I think, uh, this is about building US domestic fab capacity.

    8. JAJack Altman23:27

       Yeah.

    9. AFAndrew Feldman23:28

       Um, I, I think... I mean, look, w- when, when a, a big ship, right, tried to parallel park in the Suez Canal, right, we were delayed in chips, and we couldn't buy washing machines.

    10. JAJack Altman23:41

        Yeah.

    11. AFAndrew Feldman23:42

        Right? I mean, if we lost our chip capacity, it would be catastrophic for our, for our industry, and not just for our industry, but for-

    12. JAJack Altman23:51

        For the country

    13. AFAndrew Feldman23:51

        ... for the country-

    14. JAJack Altman23:52

        Yeah

    15. AFAndrew Feldman23:52

        ... for the economy.

    16. JAJack Altman23:53

        Uh, Eric, how did you as a board member, when you were, like, learning this stuff as you went, you still obviously... You know, I don't think Andrew's been lying to me. I think you've managed to be very helpful. Like, what-

    17. AFAndrew Feldman24:03

        [laughs] Eric was extremely helpful.

    18. JAJack Altman24:04

        So, like, how did you approach this?

    19. EVEric Vishria24:08

        Um, I have a very small circle of competence. It is not any of the stuff that he talked about. If he, if he, if you asked me to explain what packaging is right now, I could not do it. I certainly could not do it in an adequate way, and Andrew explains things to me all the time Um, and you know, so I think it's important just to be like, "Hey, this is where you, what you can do and this is what you can't do," you know? And there are periods of time in any company, particularly if you're doing technological innovation, where there's just like, there's just, you just have to let the engineers engineer and like, and the scientists do their thing and, and, and stay out of the way and keep them financed. Maybe keeping them financed was very important I guess for a huge part of it.

    20. AFAndrew Feldman24:48

        I, I think there are a couple things that, that made, uh, Eric a good board member, and I think that are foundational in being a good member right here. You don't know about everything, and, and share and make us better in those domains where you're a real expert. And don't talk about those other domains [laughs] , right? They, sometimes there's a lot of words and, and we were lucky. We had a really good board-

    21. EVEric Vishria25:10

        Right

    22. AFAndrew Feldman25:10

        ... and everybody knew what they were good at, and they, they helped us in those domains in which they had real expertise, right? I mean, one of the, the advantages of, of, of the venture world is you, you can see across an industry, right? When, when we're deep in it, we're going deep. They can see wide, right? But th- there weren't efforts by the board to try and solve technical problems. They, they didn't have... That, that wasn't their, their expertise. Um, talking about how we might finance the company, thinking about, uh, all sorts of other things, they were enormously helpful and they were patient. And I, I think they asked thoughtful questions, and our board member... Our, our, our board meetings

11. 25:51 – 27:00

    How to be a good hardware board member

    1. AFAndrew Feldman25:51

       made us better.

    2. JAJack Altman25:52

       It's like, what, what, what, what should a board do in a hardware company in general when, you know, you expect that there's gonna be much less to contribute on the product? Is it financing and recruiting and...?

    3. AFAndrew Feldman26:03

       I, I, I think the, the question when you've got a long, hard project, right? And this is the opposite of SaaS, right? You, you raise money, and then you spend two or three years to build one before you have any real idea what the customer's gonna say. You go and you talk to customers, and they say, "Yeah, yeah, that sounds great," because w- who's not gonna say, "Yeah, yeah, that sounds great," right [laughs] ? And, and so the product management is unbelievably difficult. You survey customers, and because it's no cost and it's easy to say, "Yeah, yeah, this is great," nobody wants to, to, to sort of put their foot on the throat of, of somebody else's idea [laughs] , r- right? They say, "Yeah, yeah, it's great." It takes you three years to, to, to, two and a half years to get to the point where you can bring it to a customer. I think what you can do as investors is understand that trajectory, understand that your first chip is very rarely a good one, and it's a second or third one. I mean, even really strong teams, like Google's TPU team, the fourth one was good.

    4. EVEric Vishria26:59

       Mm.

12. 27:00 – 27:50

    Specialization vs. flexibility

    1. AFAndrew Feldman27:00

       The first two were, "Oh, eh." The, the fourth or fifth were really good parts. It takes years. You have to, to know that going in. You have to know it's a long game, right? Um, ask questions related to the long game [laughs] , right? Are you hiring the right people, right? Are you thinking about this in the... Is it, have you made the right decision between, uh, specialization and flexibility, right? Th- these are questions that, that can really help sharpen our thinking. But, you know, w- w- whether to u- use this one technique or this other technique in, in, in design, whether to use w- who, one tool vendor or the other, you, you gotta let that, you gotta let the team-

    2. EVEric Vishria27:45

       Yeah

    3. AFAndrew Feldman27:46

       ... team pick that.

    4. JAJack Altman27:48

       Did it feel really different for you versus, like, an

13. 27:50 – 29:05

    Hardware vs. software investing

    1. JAJack Altman27:50

       infrastructure-

    2. EVEric Vishria27:51

       Totally

    3. JAJack Altman27:51

       ... role?

    4. EVEric Vishria27:51

       Yeah, totally. It's just very different. I mean, your time to revenue's so much longer. The revenue comes in, like, giant chunks. It, it, um... The initial customers, you know, we had US government customers, then we had, you know, huge partners-

    5. AFAndrew Feldman28:06

       Sovereign cloud

    6. EVEric Vishria28:07

       ... sovereign cloud and G42. You know, then you have now today obviously OpenAI, and so you have these, like, things. Your, so your customer concentration's more... It, it just, like, every dynamic's different.

    7. JAJack Altman28:19

       Yeah.

    8. EVEric Vishria28:19

       The, the investment and go to market's so much less. So I think it, it, it feels really, really different. The things that's, are the same, by the way, are you need really good people from a whole bunch of domains coming together who stay motivated over a really long period of time. That part's actually the same. And, and then the market is very dynamic. Like, I mean, think about it. Pre-transformer, you started the company pre-transformer.

    9. AFAndrew Feldman28:45

       We did.

    10. EVEric Vishria28:46

        Right? At that time, um, TensorFlow was dominant. So the first so- software-

    11. AFAndrew Feldman28:50

        TensorFlow was dominant. ResNet was, was still a thing.

    12. EVEric Vishria28:53

        Right.

    13. AFAndrew Feldman28:54

        Um, th- these were very small, very simple networks compared to where we are today. Um, yeah, the, the actual ML, the AI

14. 29:05 – 35:22

    The pivot from training to inference

    1. AFAndrew Feldman29:05

       looked nothing like it does today.

    2. EVEric Vishria29:07

       Well, to- actually, I think that's really, you know, for entrepreneurs for, who, who listen to this, like, it was started as a training system. Like, part of the vision that you pitched-

    3. AFAndrew Feldman29:17

       Yeah

    4. EVEric Vishria29:17

       ... and part of what you explained was like, training's a harder problem because of back PROV. Inference we thought was gonna be on devices and computers and really distributed, which may still end up happening. And, like, fast-forward to, to 2022, 2023 maybe-

    5. AFAndrew Feldman29:33

       Yeah

    6. EVEric Vishria29:33

       ... '24, we, you pivoted the company, or I don't know if pivot's the right word, but, like, to, to inference. So where, h- how and why and where did we get lucky and where did, where did you have real foresight on that?

    7. AFAndrew Feldman29:45

       I, I think there's sort of, uh... There, there are two ways to, to, to make decisions and to, to think about it. One, one I think about like an old telephone circuit, right? You, you, you, they, they set up a, a dedicated view all the way to your brother in New York, and tha- that's one way to do vision, right? That, that, that's one way, that, that you've got this idea of the way the future's gonna look way out there, and th- that's a circuit we, we think of.

    8. EVEric Vishria30:19

       Right.

    9. AFAndrew Feldman30:20

       Um, the other way to do it is the way routers work, the way the internet works, where you go hop to hop You go to Cleveland and then there's a decision whether it's best to take the next hop or, and that's sort of the way we worked, where we, we knew there was a pot of gold out there, but the path to it we knew was, was unknowable. And so each time you get over a new mountain, you look around and you re-decide. So it's, it's this sort of hop-to-hop thinking. So can we, can we build it? Yes. Can we make it work in routing? Yes. Is routing now where everybody's focusing? That's where they're focusing. Can we be a fast router? Yes. But now we're, we're seeing the lay of the land. We're seeing the sort of unfolding of AI, the rapid growth of intelligence. Well, who's gonna use it? We use AI through inference. So we, we're in a position, we're engaged in conversations, we had product being used where it allowed us to see something new, and that new thing was, holy cow, the trajectory of AI will make it smart enough that everybody will wanna use it. If everybody wants to use it, inference is gonna crush-

    10. JAJack Altman31:33

        Mm-hmm

    11. AFAndrew Feldman31:34

        ... [chuckles] right, the compute infrastructure. We better be there. And so each time we achieved something, it, it sort of moved us up a, a mountain or a hill. It gave us a new view of the landscape. We could make some new decisions, and we were always sort of moving in the same direction, but we didn't know how to, how to get there. And so decision, hop, you, you get to a, a point of view, think carefully, earn the next viewpoint.

    12. JAJack Altman32:04

        Yeah.

    13. AFAndrew Feldman32:05

        Earn the next viewpoint.

    14. JAJack Altman32:06

        Well, it's interesting 'cause, you know, with software you obviously can both, like, think nimbly and change everything the next day.

    15. AFAndrew Feldman32:12

        Right. No.

    16. JAJack Altman32:13

        With hardware you can't do it like that.

    17. AFAndrew Feldman32:14

        We don't... That, that's right. Our, ours, uh, we, we have bigger discrete decisions, right? A- and, um, that's why, uh, y- the choice in your chip of specialization versus flexibility is so important. We, we made a couple really good decisions in, in our first architecture, where we decided not to sort of embed technology that, that would accelerate convolutional networks. Instead we said, "We don't know how long those will last. If we work underneath that and it accelerate the, the algebra that underpins all AI we knew about," that was a really good decision because when transformers came out, we were the fastest at those too, even though we'd never seen them and never had heard of them, and they hadn't been invented when we set the architecture. And so it, it helps to make a few good decisions.

    18. JAJack Altman33:10

        Yeah. It's interesting, like, on the, you know, on the backs of, you know, companies like SpaceX, Anduril, Cerebras, you know, Palantir, maybe others, but obviously, like, hardware is now hotter than hot and everyone wants to fund hardware, but it's, it does, just talking this through it, it's crazy hard.

    19. AFAndrew Feldman33:28

        It's really hard.

    20. JAJack Altman33:29

        Yeah.

    21. AFAndrew Feldman33:29

        It, it's hard and, uh, it, it requires enormous internal fortitude and it requires, uh, success has historically been predicted by some experience in the field.

    22. JAJack Altman33:42

        Yeah.

    23. AFAndrew Feldman33:42

        Um-

    24. JAJack Altman33:43

        Different than, like, you know, AI where you see the advantage to-

    25. AFAndrew Feldman33:46

        Well, I, I think in, look, in, in AI, in social networking, um, a lot of the, the founders and the leaders were building tools for themselves a- and their classmates and their friends, and there they had unique insight. I mean, if, if you look at cognition, if you look at Cursor, th- these are some of the best engineers, software engineers in the world. They're building tools for themselves, right?

    26. JAJack Altman34:08

        That's true. You know-

    27. AFAndrew Feldman34:09

        And, and that's different from, uh-

    28. JAJack Altman34:12

        I also wonder if this last topic probably plays a big role in it, which is when you can change your opinion the next day and it's fine that you were wrong yesterday, speed and decision-making speed-

    29. AFAndrew Feldman34:21

        Mm-hmm

    30. JAJack Altman34:22

        ... trumps-
15. 35:22 – 39:14

    Young product leaders and seasoned hardware engineers

    1. AFAndrew Feldman35:22

       old, like 27. [laughs]

    2. JAJack Altman35:24

       [laughs] Well, that's... But, but you know what's interesting actually is if I look at the product leaders in your organization, if I look at some of the go-to-market leadership-

    3. AFAndrew Feldman35:32

       Yeah

    4. JAJack Altman35:33

       ... these are, they're actually very young.

    5. AFAndrew Feldman35:35

       They are.

    6. JAJack Altman35:36

       And, um, and I'm just interested in, like, how, was there intentionality around that? Not necessarily ageist obviously, but just, like, how you combined the perspective of, you know, young, at the cutting edge product leaders with seasoned hardware engineers.

    7. AFAndrew Feldman35:55

       I, I think we tried to think really hard about w- where experience mattered, where blistering intelligence matters. Um-

    8. JAJack Altman36:02

       Mm-hmm. That's true

    9. AFAndrew Feldman36:02

       ... you know, if you look at, at our product organization, um, unbelievably smart. Un- I mean, some of the best product people I've, I've ever seen. Um, all young, all promoted from within. Um, you know, we, we, we don't have big company rules. You gotta be in a job for this amount of time before you can get promoted-

    10. JAJack Altman36:22

        Sure

    11. AFAndrew Feldman36:22

        ... right? I mean, if you're extraordinary, we're gonna give you more and more responsibility and more and more, and if you do a great job with it, there's no end to wh- where we will, uh, we will take you. Um- You know, my co-founder, Sean, uh, you know, one of the five of us, I mean, in my last company, we, we hired him as an individual contributor. And when we were acquired four years later, the guy was 25 or 26 when we, when we hired him in 2007, right? When we were acquired by AMD, we made him a, a corporate fellow, right? There, there was just no end. I mean, and now, now he's a founder and he's CEO- CTO of a, of a public company.

    12. EVEric Vishria37:02

        Mm-hmm.

    13. AFAndrew Feldman37:02

        Um, and we, we do that sort of ruthlessly. Um, and, uh, there are some areas where to be exceptional requires a tremendous amount of experience. There are some areas where it doesn't require any experience. We, we, we don't have a long history of, uh, uh, of understanding what customers want in AI.

    14. EVEric Vishria37:22

        Mm.

    15. AFAndrew Feldman37:22

        So their methodology, smarts, insight trump experience. In other areas, in the making of chips, it's been my experience that, that, that, uh, uh, some odd years of previously building chips is the best predictor, uh, of whether you're gonna deliver exceptional chips. And that's also true on the mechanical side, on the system side.

    16. EVEric Vishria37:44

        Mm.

    17. AFAndrew Feldman37:44

        There's just not a lot of chance in college or in graduate school to actually build silicon, to actually build a, a machine. They're so expensive. They're, they're so hard to build. Um, they take so long that even in a doctoral program, you don't get a chance to tape out a chip, to deliver it, to bring it up, to put it on a board, to power it, to write the software for it, and so it takes some time.

    18. EVEric Vishria38:06

        Are the, are the relationships something that are critical, or can those be earned quickly by young people?

    19. AFAndrew Feldman38:12

        Um, I, I think the following. First, five is too many founders, without question.

    20. EVEric Vishria38:17

        [laughs]

    21. AFAndrew Feldman38:17

        Except that we'd worked together before.

    22. EVEric Vishria38:19

        Mm-hmm.

    23. AFAndrew Feldman38:19

        And so the-- everybody knew what the other, other folks were good at, right? And, and so there, there wasn't a lot of head butting at all. I mean, we'd all worked together previously. We all had tremendous respect for what the others could do and, uh, and some humility about what we couldn't do. And so we, we were able to, to, to do that. Um, your specific question is does it take a lot of time to build trust. I, I think within the six or eight weeks of working with someone, you can tell if they're extraordinary.

    24. EVEric Vishria38:48

        Mm-hmm.

    25. AFAndrew Feldman38:49

        Right? I mean, extraordinary people, in the first email they send you, you go, "Whoa, that's exactly what I needed." [laughs]

    26. EVEric Vishria38:55

        Yeah.

    27. AFAndrew Feldman38:55

        Right? Every list is in descending order of importance. There's not a lot of fluff. There's high signal. There's, you go, "Whoa." And then you see that again, and then you watch the way they run a meeting, and you go, "Whoa." And then you watch them deliver something. They motivate a bunch of people around them. They can work across the organization. You say-

    28. EVEric Vishria39:12

        Yeah

    29. AFAndrew Feldman39:12

        ... "Well, that's a person I need on the next

16. 39:14 – 42:20

    External relationships and TSMC

    1. AFAndrew Feldman39:14

       important project."

    2. EVEric Vishria39:14

       Yeah.

    3. AFAndrew Feldman39:15

       And then y- they, they just, they chew through work.

    4. EVEric Vishria39:18

       I always thought with recruiting, it's like if I came away, like, learning real things and I wanted to have another interview or meeting just 'cause I was like, "I'm gonna learn more stuff-"

    5. AFAndrew Feldman39:27

       Yeah.

    6. EVEric Vishria39:27

       "... I was like, that's my best predictor."

    7. AFAndrew Feldman39:28

       The best.

    8. EVEric Vishria39:28

       Yeah.

    9. AFAndrew Feldman39:28

       The best.

    10. EVEric Vishria39:29

        It's so good.

    11. AFAndrew Feldman39:30

        The best.

    12. EVEric Vishria39:30

        What about the external relationships? Like, you know, you've gotta work with a lot of people outside your company. And you know, you gotta be in, you gotta be in Taiwan. You gotta work with, you know.

    13. AFAndrew Feldman39:39

        You know, it, it helps to bring those with you a little bit, some of them, right? So, you know, we'd been, been building chips with, with TSMC for decades. We'd been working with our contract manufacturers for decades. And so, um, especially in a time of contention, that, that, that, that those relationships had, had been in place for years, that you'd been good to your word, not once, not twice, not just in good times, but you'd, you'd been good to your word over good and bad times. That was really, really important. I, I think you earn relationships with new, uh, with new partners in exactly the same way. You, uh, you're good to your word. Um, you get them information early. You write. I mean, that, that... I- it's, you know, write a thank you note. [laughs]

    14. EVEric Vishria40:26

        [laughs]

    15. AFAndrew Feldman40:26

        Right? No, really, I mean, be, do, do what your mother said, right? Be a good person. [laughs] Write a thank you note. Do what you say you're gonna do. Um, and-

    16. EVEric Vishria40:33

        I think this, this to me was one of the big learnings, like particularly during the COVID era and everything. If I think of the software companies, even software infrastructure companies we work on, it's really, like, their vendor that matter is AWS, and like maybe now their vendor that matters is AWS and, you know, a foundational model company or something like that. There's like two vendors that matter. This is like, I don't know what the vendor list is, but it, it's, it's insane.

    17. AFAndrew Feldman41:02

        Dozens.

    18. EVEric Vishria41:02

        It's dozens, right? It's, it's, it is the, all of these components that go into the system, all of these specialty manufacturers that build the cooling plate and part of the water system and, like... And so you have all of these things that have to kinda come together and, and we just aren't used to that. Like that, those ma- that many dependencies in order to deliver a product or at least-

    19. AFAndrew Feldman41:25

        Yeah

    20. EVEric Vishria41:25

        ... I wasn't... So that was like a big thing and you have to kind of work with them on-

    21. AFAndrew Feldman41:30

        Software guy coming to grips with the supply chain.

    22. EVEric Vishria41:32

        Yeah. [laughing]

    23. AFAndrew Feldman41:33

        [laughing] Whoa.

    24. EVEric Vishria41:34

        Whoa. Yeah, exactly.

    25. AFAndrew Feldman41:35

        They call it a chain for a reason. [laughing]

    26. EVEric Vishria41:38

        [laughing] And, and, and it really is.

    27. AFAndrew Feldman41:39

        It is.

    28. EVEric Vishria41:39

        And you, and you're like, "Wait a minute. Holy cow, this, all of this stuff." And when you're growing exponentially, like, then everything becomes even more complicated, right? 'Cause it's like you're putting in... I, I n- never thought about this. You know, it's like AWS, we want more capacity. You know, you go online and you add capacity. Like, you want more capacity on, in terms of like number of wafers, like that's a 15-month lead time kind of decision.

    29. AFAndrew Feldman42:06

        Mm-hmm.

    30. EVEric Vishria42:07

        And a huge amount of capital, and your vendor is also making a huge allocation decision, and so they have to buy into it too. And so I, I, that, that was a big learning for me in just terms of like how much

17. 42:20 – 44:12

    The AI infrastructure buildout

    1. EVEric Vishria42:20

       complexity there is in that.

    2. AFAndrew Feldman42:23

       One of the things we've been talking about a lot internally is that it feels like no matter how big you think it is, y- y- it, it is hard to wrap your head around the size of the AI buildout that's happening-

    3. EVEric Vishria42:34

       Yeah

    4. AFAndrew Feldman42:34

       ... and the CapEx going in. You know, I, I actually saw a chart this morning that was like 1% of GDP-

    5. EVEric Vishria42:39

       Yeah

    6. AFAndrew Feldman42:39

       ... was spent per year on

    7. JAJack Altman42:41

       ... highway and telecom.

    8. AFAndrew Feldman42:43

       Right.

    9. JAJack Altman42:43

       And then it was 2% for railroads.

    10. AFAndrew Feldman42:46

        Yeah.

    11. JAJack Altman42:47

        And then this AI build-out is, like, 3.5% of GDP.

    12. AFAndrew Feldman42:50

        Yeah.

    13. JAJack Altman42:50

        It's like-

    14. AFAndrew Feldman42:51

        Only one person got it right

    15. JAJack Altman42:53

        Which was?

    16. AFAndrew Feldman42:53

        Your brother.

    17. JAJack Altman42:54

        Nalla? That's true.

    18. AFAndrew Feldman42:55

        Yeah. I mean, the only person, and everybody thought he was out of his mind. I mean, what, what, w-what, what Sam's really good at, and I, and I think is so hard, is he saw an exponential and wasn't afraid.

    19. JAJack Altman43:09

        Yeah.

    20. AFAndrew Feldman43:09

        Right? [laughs] They... You take that exponential-

    21. JAJack Altman43:12

        It's big

    22. AFAndrew Feldman43:12

        ... out two or three or four or five years, you go, "Holy crap."

    23. JAJack Altman43:15

        Mm-hmm.

    24. AFAndrew Feldman43:15

        Right? He wasn't afraid. Everybody else was afraid. Every- "No, it's gonna slow down. It's gonna..." And he was like, "No."

    25. JAJack Altman43:21

        I remember the, uh, Stargate, like 7 trillion or whatever.

    26. AFAndrew Feldman43:23

        Right. I mean, now, now the initial Stargate is, is, is sadly small. [laughs]

    27. JAJack Altman43:28

        Yeah.

    28. AFAndrew Feldman43:28

        Right?

    29. JAJack Altman43:29

        Yeah, yeah.

    30. AFAndrew Feldman43:29

        And it, it was mind-bogglingly large.
18. 44:12 – 53:14

    The data center supply chain

    1. JAJack Altman44:12

       What, what, what is that like for you?

    2. AFAndrew Feldman44:15

       The data center is one of the, the links in the supply chain to deliver compute via the cloud. And the truth is, it's whether you deliver it by the cloud to, via cloud, or you deliver it what we call on-prem, it impacts both case. In one case, you rent the capacity and you put your, your equipment in it. In the other case, your customer rents the data center. And so i-it's, the, it is a, a limitation in the market right now. Um, it, it exposes a, a whole bunch of weaknesses in the US. Uh, our grid is pathetic and sort of built on 1940s or '50s technology. Uh, we, we stopped doing work in, in very interesting technologies that turn out to be extremely clean, like nuclear. I mean, wouldn't it be ironic if it took AI to get, bring us back to doing nuclear? [laughs] Right? I mean, what-

    3. JAJack Altman45:09

       Yeah.

    4. AFAndrew Feldman45:09

       That's right.

    5. JAJack Altman45:09

       Yeah.

    6. AFAndrew Feldman45:09

       Pretty... Right. Um, it, it showed that, you know, in these data centers, uh, we use generators as backups, and these are either diesel or, or LNG or one form of gas, right? And these come from GE Vernova, or they come from Caterpillar. There'd been no innovation in, uh, in generator and diesel gensets for, for decades. Now suddenly there's innovation. Um, the guys at Boom want to use what they designed for jets to, to, to power data centers. We're seeing all sorts of interesting things in, in battery backup and, um, uh, like from Bloom Energy, you're seeing interesting fuel cell. I mean, there's just this enormous innovation be-because there's necessity driving it. We don't have enough data centers. They're coming on too slowly. Uh, the grid is old. Uh, and so we're, it's an area of tremendous innovation. I, I think the industry did itself no, no favors by doing some dumb stuff at the beginning, tried to pawn off some costs on local communities-

    7. JAJack Altman46:16

       Mm

    8. AFAndrew Feldman46:16

       ... it tried to take advantage of local municipalities. There, there is no reason a data center shouldn't pay its way. There's no reason why it should use very much water. We use closed loop systems, right? All the data centers in the US use less than the California almond growers, not by 1X or 2X or 4X, but between four and seven times the almond growers use more.

    9. JAJack Altman46:37

       Yeah.

    10. AFAndrew Feldman46:37

        So, I mean-

    11. JAJack Altman46:38

        The water thing's just not a thing

    12. AFAndrew Feldman46:39

        ... the water thing's not just a thing.

    13. JAJack Altman46:40

        Yeah.

    14. AFAndrew Feldman46:40

        But w- as a community, we didn't do a good job of communicating with local communities, getting their buy off, showing them that we're gonna bring thousands and in some cases, ten thousands of high-paying construction jobs. We're gonna pay ongoing jobs, and their tax base ought to go down over time. And we didn't do a good job of that, and now we're paying the price.

    15. JAJack Altman47:01

        I mean, it kind of goes with the whole theme of tech doing a terrible job communicating about AI in general, I would say.

    16. AFAndrew Feldman47:06

        Horrible job. We're doing a horrible job.

    17. JAJack Altman47:07

        Yeah. Um-

    18. AFAndrew Feldman47:09

        I think the data center thing, the other part of the data center thing that's interesting to me is, like, if you a- if I asked you, I don't know, 2018, 2020, whatever, what's the probability that data centers were gonna be a critical factor in our supply chain- Yet another thing I'd have gotten wrong

    19. JAJack Altman47:22

        ... right?

    20. AFAndrew Feldman47:23

        Yeah.

    21. JAJack Altman47:23

        We'd all got it wrong.

    22. AFAndrew Feldman47:24

        We got it wrong.

    23. JAJack Altman47:25

        And it just, it, and that was just like a, it's a new thing. It's like, okay, now you have to build it all the way through and deliver it, but-

    24. AFAndrew Feldman47:31

        How long does it take to go from shovel... I know you, you joked about shovel r- I actually want to ask you about that, but how long does it take to go from start to finish on a data center?

    25. JAJack Altman47:38

        So sh- sh- shu- shovel-ready i- is an expression you hear in the data center world. It means we haven't done shit. [laughs]

    26. AFAndrew Feldman47:45

        But we're ready to.

    27. JAJack Altman47:46

        Right. We wanna sell you a pile of dirt-

    28. AFAndrew Feldman47:49

        [laughs]

    29. JAJack Altman47:49

        ... ready for your shovels to start doing something.

    30. AFAndrew Feldman47:52

        It does sound better than-
19. 53:14 – 54:10

    Speed creates markets

    1. AFAndrew Feldman53:14

       If you wanna punish them, right, d- don't, don't take away their phone. Ratchet it back to dial-up speed.

    2. JAJack Altman53:21

       [laughs]

    3. EVEric Vishria53:21

       [laughs]

    4. AFAndrew Feldman53:22

       Right? Right. This is a real punishment. Let them use it for a week at dial-up speed.

    5. EVEric Vishria53:28

       Holy.

    6. AFAndrew Feldman53:28

       Slowly, right? That why is... We laugh and then we say, "Oh, it's okay for, for AI to be slow." I mean, think about it. I mean, when, when the internet was slow, Netflix delivered DVDs in envelopes.

    7. EVEric Vishria53:40

       Yeah.

    8. AFAndrew Feldman53:41

       And when the internet got fast, they became a movie studio, right? That's not... They didn't get better at their other thing. They became something entirely new, and I, I think what we're seeing with the launch of, uh, of GPT 56 Seoul, right, it's put out in limited, limited availability last week, that people are thinking of whole new applications. You've got frontier intelligence instantly and that, that opens up all sorts of new opportunities. So that's one thing you do with speed.

20. 54:10 – 55:24

    Disaggregation with AMD and AWS

    1. AFAndrew Feldman54:10

       The other thing you do is you try and think about how you can drive up throughput, how you can make more tokens, and one of the ways we're doing that is by partnering in, uh, something called disaggregation, and we're doing it with AMD, we're doing it with AWS, where you think about in the work of inference, is there a part that can be done by somebody else and is there a part that can be done by you such that the, the result is higher throughput? And, you know, with, with AMD we're seeing 5X additional throughput. I mean, five times as much throughput while keeping the speed the same. And we're seeing similar numbers with, with, with AWS. We have an opportunity because of our architecture to do that with, with the entire GPU landscape. We, we, we could do it across the board. And so, um, you know, there, there are four major, uh, chip makers right now in, in our category. Obviously, Nvidia, we'd love to partner with them. There's AMD, there's the Google TPU, and there's AWS with their Trainium parts, and we're already working with two. So that's, uh, something we've thought a great deal about and, and is a vector we are extremely interested in, in chasing down.

    2. JAJack Altman55:23

       What has it been

21. 55:24 – 57:30

    What actually makes Nvidia great

    1. JAJack Altman55:24

       about Nvidia that has made them have the just crazy run that they've had? You know, like I've seen-

    2. AFAndrew Feldman55:28

       I, I think most people are, are wrong about N- N- what makes Nvidia great. First, N- Nvidia's, you know... I- in the first quarter of this century, they're, they're the great company without any question, right? Um, and I, I think people look to a bunch of things. They look to CUDA. They, uh, I, I don't think it's CUDA. They, they look to their chip architecture. I don't think it's their chip architecture. Um, uh, it's this unbelievable grit and intensity that was born of a decade of not having success as a public company. I think if you look at their stock chart between about 2002, what, 2003 and 2013, 2004, 2014, for a decade, right, they traded horribly. And you're a public company, and you're fighting tooth and nail, and no one's listening to you, and you can't sell very much. And the relentlessness and the grit that that takes is awesome. And to come out of that as the most valuable company a decade later, right, the most valuable company in the world, that sort of intensity and grit for a company of their size, in my view, makes them one of the great companies in history.

    3. JAJack Altman56:40

       Yeah.

    4. AFAndrew Feldman56:40

       It's, it's not this, these other things. These are just things people say. But th- th- that, that sort of a decade of being a public company and fighting, and fighting, and fighting, that gets in your DNA, and that, that's awesome.

    5. JAJack Altman56:55

       Yeah, I mean, you like you see Jensen like still at events.

    6. AFAndrew Feldman56:59

       Fighting tooth and nail.

    7. JAJack Altman57:00

       Yeah.

    8. AFAndrew Feldman57:01

       Right? I mean, where-

    9. JAJack Altman57:02

       CEO-

    10. AFAndrew Feldman57:02

        Could you have imagined-

    11. JAJack Altman57:03

        ... of $5 trillion company.

    12. AFAndrew Feldman57:04

        Could you imagine old tech leaders before Jensen doing that?

    13. JAJack Altman57:08

        No.

    14. AFAndrew Feldman57:08

        Right. Right.

    15. EVEric Vishria57:09

        Have that kind of fight.

    16. JAJack Altman57:11

        Yeah.

    17. AFAndrew Feldman57:11

        I mean, th- that in my view is what's awesome.

    18. JAJack Altman57:14

        Yeah.

    19. AFAndrew Feldman57:14

        Um, that level of fight. I, I think the other stuff, cool, good, but, but when, when I look at what, what I can do better, when I look at sort of w- what I ought to be thinking about as being a CEO, th- those are the things I look to.

    20. JAJack Altman57:29

        Do you think, did your

22. 57:30 – 59:58

    Near-death experiences and the DNA of a Goliath fighter

    1. JAJack Altman57:30

       near-death experiences give your company some of that DNA?

    2. AFAndrew Feldman57:33

       You know, we, we, we are, uh... And I, I think this is what's interesting about Jensen, too, is he sees himself still as the underdog.

    3. JAJack Altman57:42

       Hmm.

    4. AFAndrew Feldman57:42

       Now, now he's the big dog. But, you know, this is my, my fifth startup. You know, we sold three and took one public previously, and we've now taken this one public. Um, I'm a, I'm a professional David in the battle with Goliath, and I, I wake up every day with that mentality.

    5. JAJack Altman57:59

       Hmm.

    6. AFAndrew Feldman58:00

       And th- th- we're now bigger. Uh, we, we have sort of bigger competitors, right? We have bigger challenges. We have more people throwing stones at us. Um, and so I, I, I hope we take that. I personally wake up every day with that passion and that drive. I mean, every day I think to myself, when we started, they said it would never work. You can't do wafer scale, and we made wafer scale. And then they said, "All right, you did wafer scale, but y- you can't yield it in volume." And then we yield it in volume. They said, "Okay, you can yield it in volume, but you can't package it in volume production." And then we packaged it in volume production. They said, "Okay, now you've packaged it in volume production. You only have a government customer." And we said, "Okay." Then we won, uh, a sovereign cloud. And then they said, "You don't have a, a hyper- you don't have a, a, a na- a, a frontier lab, and then we won Open Air. And then they said, "Okay, you've got government. You, you've got-

    7. JAJack Altman58:55

       Yeah

    8. AFAndrew Feldman58:55

       ... sovereign clouds. You've got a hyper- you've got a, uh, a, uh, a frontier lab. You don't have a hyperscaler." Then we won a hyperscaler, [laughs] right? E- each time they said, then they said, "You, you've got all these cool customers, but you couldn't, you couldn't do big models." Now we're serving GPT-5.

    9. JAJack Altman59:11

       What are they saying? What, what, what's that now? Like, what are they saying you can't do now?

    10. AFAndrew Feldman59:13

        The CEO's a boomer. [laughs]

    11. EVEric Vishria59:17

        [laughs] That's good.

    12. AFAndrew Feldman59:19

        But, but wh- but each time-

    13. EVEric Vishria59:20

        And then I became a millennial, and I showed them. [laughs]

    14. AFAndrew Feldman59:23

        That's right. [laughs] You know-

    15. EVEric Vishria59:24

        That was, that was-

    16. AFAndrew Feldman59:25

        It, it, right, and so I, I think wh- when you're sort of a professional David, when you, when you are an entrepreneur at heart, right, each one of those fires you up, right? We're only interested in solving problems that other people can't solve. The- we're only interested in doing things that, that other people can't do. That's why we get up every morning. Um, it, it's not the money. Uh, it's not, it's not, I mean, notoriety. It's because we love building cool things, and we really like building cool things that are so hard that other people can't build them.

    17. JAJack Altman59:56

        Eric told me you had a pretty cool and

23. 59:58 – 1:02:42

    Andrew's childhood next to William Shockley

    1. JAJack Altman59:58

       unique childhood. I don't know if that fed into this at all, but, like, how'd you grow up?

    2. AFAndrew Feldman1:00:02

       I grew up on, uh, the Stanford campus. Uh, my parents were faculty, and there's a, a little neighborhood, uh, where all your, all your neighbors are, are other professors. Um, my neighbor was, uh, William Shockley, he invented-

    3. JAJack Altman1:00:17

       That's crazy

    4. AFAndrew Feldman1:00:18

       ... transistor and-

    5. EVEric Vishria1:00:19

       So crazy. [laughs]

    6. AFAndrew Feldman1:00:20

       Right. [laughs] So all we knew about him, I, you know, we're 10 or 8, is that his wife gave out full-size candy bars at Halloween, [laughs] right? He's the inventor. He was at Bell Labs. He invent- he brought Silicon Valley, right? He, he-

    7. JAJack Altman1:00:33

       Yeah

    8. AFAndrew Feldman1:00:33

       ... by the movement of Shockley to the West Coast created the foundation for Silicon Valley, and we're thinking he gives big 3 Musketeer bars, [laughs] right? Um, but I, I think there were a couple things that were glorious. First, the only currency was intellectual horsepower, right? That, that was n- nobody cared if you were rich. Nobody cared if you'd started a company. That, th- this was the '70s. Nobody cared. What they cared about was, "Oh, that dude's really smart. He does good work."

    9. JAJack Altman1:01:02

       Hmm.

    10. AFAndrew Feldman1:01:02

        My other neighbor was Amos Tversky, and f- for, uh, his work with David Kahneman, they got a Nobel Prize in Economics. Um, my dad's tennis match, uh, there were six or eight guys in rotation. They played doubles on Saturday and Sunday. And, um, somewhere in my mid-20s, I realized three had Nobel Prizes and one had a Fields Medal. [laughs]

    11. JAJack Altman1:01:23

        That's-

    12. AFAndrew Feldman1:01:24

        Right? Um, and-

    13. JAJack Altman1:01:26

        Maybe he sucks at doubles, but... [laughs]

    14. AFAndrew Feldman1:01:28

        Let me tell you, it was some old man tennis. I mean, their serves were grim. [laughs]

    15. JAJack Altman1:01:33

        That's a crazy-

    16. AFAndrew Feldman1:01:33

        Their physics was good. [laughs]

    17. JAJack Altman1:01:35

        Cra- it's a crazy way to grow up.

    18. AFAndrew Feldman1:01:36

        That's a crazy way to grow up.

    19. JAJack Altman1:01:37

        Yeah.

    20. AFAndrew Feldman1:01:37

        And you know, we, the neighborhood was safe. Uh, we'd get on our bikes, and we'd just go all summer and come back at, when it was dark. Um-

    21. JAJack Altman1:01:45

        And so did you think you'd be an academic?

    22. AFAndrew Feldman1:01:47

        I did. I, uh, actually, I was working on a PhD. I got a little bored. I went to, to the business school at Stanford while I completed my qualifying exams for my PhD, and I sort of got, got sucked into this by mistake.

    23. JAJack Altman1:02:00

        Hmm.

    24. AFAndrew Feldman1:02:01

        And, uh, you know, my dad still asks me, he's like, "You gonna finish your PhD, Andrew?" [laughs] I'm like, "Dad, all my professors are dead," right? [laughs] Nobody left. Uh, so...

    25. JAJack Altman1:02:11

        Yeah. Amazing. All right. Well, Andrew, this was a blast. Thanks a ton for doing this with us, and obviously, um, you know, you've built something extremely special, and it's been cool to just watch and learn vicariously through Eric, so thanks for everything.

    26. AFAndrew Feldman1:02:22

        Tha- it's a pleasure to, to, to be here and chat with you guys. And you know, Benchmark was, uh, an extraordinary partner. I mean, I, I think if you do hardware, you're gonna be in bed with your, your, your backers for a decade. And pick good ones, and, and I'm proud we did. [upbeat music] Thank you.

Episode duration: 1:02:45