Cerebras CEO on the Future of Data Centres, Token Costs & Memory | Should US Companies Sell to China

本站添加: 2026-05-26

[00:00:00]We can't build data centers fast enough to keep up with demand. We have a $25 billion backlog. If demand stays high, we are going to continue to see memory shortages for at least the next several years.

[00:00:12]>> I'm so excited to welcome a dear friend, Andrew Feldman, founder and CEO of Cerebrus. Last week, Cerebrris went public. The largest semiconductor IPO ever. The price went from $185 to $311.

[00:00:25]They got over $5.5 billion. Today we deep dive on the future of chips, the future of USChina relations, the future of data centers and energy. This was an incredibly wide-ranging conversation. I think it has been Nvidia's strategy to try and create competitors for the traditional hyperscalers. They have funded and backstopped and overallocated to the Neo clouds. They have created a dependence which is probably not healthy. So over time, the history of our industry is a massive reduction in the cost per unit compute. for hard problems. There is no upper bound to how much faster you want to be. Ready to go.

[00:01:13]Andrew, dude, it is so lovely to have you on the show. I have to say I was I was quite emotional last week when I saw the IPO cuz you are one of the kindest, greatest people. I I love getting to know you. I so appreciate our relationship and so to see that culminate last week was the IPO. It was really special. So congratulations for last week, dude.

[00:01:31]>> Well, thank you so much. Those were really kind words and it was really exciting day for for the company and the team and uh the people who who' believed in us and and who backed us for for a decade and uh that it was great. Thank you for for for saying those nice things.

[00:01:50]>> Not at all. And I was I was thinking like in terms of this conversation how I wanted to structure it. And I I always go back when I have amazing people like you on the show which is Alan Roosevelt's kind of statement of you know not very intelligent people discuss other people uh mediocre people kind of discuss current events uh and then intelligent people discuss the future and ideas. And so I thought I'll grapple with my own ideas and wrestle with your incredible brain to help me understand where we're at and where we're going. I want to start with on the one hand we look at the the landscapes there and it's like oh my gosh an AI infrastructure bubble and then on the other hand we look at you know you Jansen hang last night who comes out and says we're going to be spending 3 to 4 trillion on AI infrastructure by 2030.

[00:02:35]How should I balance the oh there's an AI infrastructure bubble with this appreciation of 3 to4 trillion spent by 2030.

[00:02:44]>> I've been thinking a lot about this. I I I think when you look at other bubbles and you look at bubbles in the past, uh, and I I was in one in in the late '9s when we built out an enormous amount of fiber optics. And you sometimes have economists who who maybe think it's relevant to look at 1880s, the building out of of rail. I'm not not sure that's relevant, but um what I see is that there was a pension to believe that if we built it, they would come. The infrastructure buildout was way ahead of demand. And that was true in railroads.

[00:03:26]That was true in uh fiber optic cabling.

[00:03:30]And in a strange way, that is the exact opposite of where we are with AI. The infrastructure buildout is behind demand. We can't build data centers fast enough to keep up with demand.

[00:03:43]We have a $25 billion backlog. Nvidia has a backlog. Uh AMD has a backlog.

[00:03:50]Others have backlogs. Um they have backlogs because we can't get data centers built fast enough. And it's not that we're building on the come. We're not building ahead of demand. We're building behind demand. And that is a very different observation than uh those who say there's a bubble. I don't think they've really gotten their head around the fact that we are trying to keep up with demand, not the other way around. And I I don't think that's a characteristic of a bubble. When you are trying with your infrastructure to keep up with what people want today, not in the future, today, and their demands are growing over time.

[00:04:30]Is it ultimately a good thing that we are metered in our ability to build out data centers because it almost tempers the demand if we were able Gavin Baker said actually kind of the delays and the permitting and the challenges that are incurred today actually help because if you were able to have it all today all of demand would be met with all of supply and that would actually be a challenge. Well, I I look, I think uh sometimes the world is like I was in my 20s the first time I went to Vegas and went to the buffet, right? You you eat so much you feel sick for days, right?

[00:05:07]It's all in front of you and you just gorge yourself. I I think the market can sometimes be that and I think Gavin is an extraordinarily thoughtful think sort of guy about this. I think we are being metered.

[00:05:21]We also know that the reason you put meters on a freeway is because it makes the the freeway traffic smoother and it avoids hiccups.

[00:05:32]Um that's exactly what metering is designed to do. And so I think he used that that analogy extremely thoughtfully. One of the advantages that OpenAI has was that uh and I think one of Sam's brilliances was that he saw an exponential growth and he saw what that would mean in a year or two to the demand for compute and he wasn't afraid by it. He went out and took action and perhaps others couldn't believe it or they were looking at the same demand sort of steep exponential growth and were like well we can't need that much. you can't need tens of gig that that my mind hurts if you do that. Whereas what Open AI did is they went out and they're like, we're going to contract for it here and here.

[00:06:16]We're going to get power. We're going to get data centers. We're going to sign up for hardware and an an ability to to believe your data in a exponential growth environment out a year or two or three is is a superpower. Can I do you get rewarded for that insight if you can just buy it from Elon now on demand?

[00:06:38]>> I don't think they can buy the same thing from Elon on demand. They bought down rev gear.

[00:06:42]>> I'm sorry. I've learned from doing this show for a long time. I can ask stupid questions. They they bought down rev gear. They bought they got H100s. They didn't get the B200s. They didn't get the most current. They didn't get the They are a generation and a half, maybe two generations behind. So that the this was not a great deal. It was a good deal for Elon. He had them sitting around, but they were forced to take action in a deal that I think was was not the ideal deal they wanted. It was a deal that was available.

[00:07:14]Going back to what we said about kind of the delay in data centers and data centers being a constraint. I just hear everyone say well memory memory is the shortage too Harry and that's why we're seeing an increase in cost you know four 5x in certain cases >> is memory is that true how should we think about memory being the shortage as well. Well, look when what what what's happening here is there is such extraordinary growth in demand that it is putting pressure on all parts of the supply chain. Memory after TSMC which is right after fabace memory is number two item in this uh number two item that's needed. And what's happened is there are only three companies that make the memories GPU use. We don't use that memory. But that HBM is made by Samsung and Micron and Highix and they couldn't keep up and so the prices shot through the roof. I mean Micron producing numbers where they have 80 85% gross margins. I mean they're getting software gross margins on making memory.

[00:08:25]Um yeah I I think it's extraordinary. Um that is a limitation for all GPUs but not us. we don't use it.

[00:08:34]>> Again, going back to my idea of what the future looks like, what should one expect from that? Does it ease? Does it ease over time? What happens to the cost? Well, I I think the the challenge here is that that these are extremely uh lumpy items, right? You you can't just add a little bit of manufacturing capacity at a fab. You you have to build a fab for $40 billion and takes five years to build. So if you see demand explode, you cannot respond quickly. All you can do is fill your factory. Once your factories filled, you got to build another factory, right? It's a a step function in your ability to meet that demand. And the step is huge and takes years. And so if demand stays high, uh they are we are going to continue to see memory shortages for at least the next several years. Do you think we will see a peaking of demand? You've seen so many different >> not if AI continues to improve in usefulness.

[00:09:38]I mean what what's happened here is and and this is something that that I have I haven't heard others sort of talk about is somewhere in 2025 the models got smart enough to be really useful. Before that Harry these were sort of sort of a novelty. AI was like cool and then nobody used it. Remember we we make AI with training and we use it with inference. And so once the the AI we made 2025ish first half got smart we began using it and this explosion in demand that Jensen described all right and that we very much agree with is is happening. That's because people are using it every day and they're using it on more and more problems. They're using it on harder problems and it is sweeping through different demographic groups. It's not just 28-year-olds in Silicon Valley.

[00:10:36]It's my 85year-old father. It's right.

[00:10:39]It's 11year-old my 11-year-old niece. It is right. It is sweeping through demographic groups and they're using it all the time. And that is what's driving this demand. And so um if we continue to find ways to make the AI the frontier models smarter and more useful, we'll keep using it the demand will continue to to on this sort of exponential curve.

[00:11:08]>> You you've compared past cycles before in this conversation. And Sarah Fry said about um kind of cloud providers can be uh similar in some perspective to what we're seeing today in terms of frontier models and she said that last night to what extent do you think you see the commoditization there and they essentially become utilities versus differentiated providers with meaningful modes.

[00:11:33]>> I think it has been Nvidia's strategy to try and create competitors for the traditional hyperscalers.

[00:11:40]I I think that has been a strategy of theirs. I think they have funded and backstopped and overallocated to the Neoclouds.

[00:11:48]I think um I I think that they're they have created a dependence which is probably not healthy. But I I think the truth is is that what what AWS and Azure offer is extremely useful for most enterprises.

[00:12:07]They offer credibility and legitimacy.

[00:12:09]They offer security. They offer layers of different software for different parts of your organization. If you'd like to enter in the AWS world, you can enter with Bedrock. You can use tools like Sage Maker. You you have a collection of different ways to to enter and you can store your data there. You you have your S3 instance. I mean you you can have an entire offering and I I think that is really valuable to a segment of the comm of the of the of the market. I think there might be other segments of the market they're like give me cheap compute. I don't care about anything else.

[00:12:46]And in that case your strength as a as a hyperscaler becomes your weakness. You have the security you have the other layers of software and you have some of the costs that are associated with that.

[00:12:59]And if people don't want that, if you don't care about leather seats, right, and there leather seats in the truck, there's extra cost in the truck. And when you buy the truck, right, you you find somebody who's got a truck that's got Ngahide seats, right? I mean, it's exactly the our business just cuz it's wrapped up in technology is no different than any other business. Is it's segmented there. There's value. That value comes at a cost. You have to make that value. The hyperscalers make the value. They make the value through software, through security, through having rules about their data centers, about the security, physical security, the various security checks they put in.

[00:13:40]Those are enormously valuable um to most parts of the market, but not all. You said about the cost there. When we look forward, how do the costs of your business change significantly over time?

[00:13:53]We we spoke about the cost of memory going up five. If we look at the cogs in five years time, how do you think they will look most significantly different?

[00:14:02]Well, the cost of the increase in memory has been very good for us because we don't suffer it, right? This has given us opportunity, right? We we use SRAMM and there's no shortage of SRAMM. Uh the cost of SRAMM hasn't changed. Um, and you know, no SRAMM maker cuz TSMC etches it into your chip while they're making the logic. Um, uh, there are no extra margins to to pay for the pay the HBM maker. And so, uh, we have been advantaged in this environment. We have been advantaged by the fact that there are constraints on co-ass at TSMC. We don't use co-ass. we are advantaged by the fact that we're at 5 nanometer and the 3nanmter node is the most overs subscribed uh and so we we have our supply chain is advantaged on these dimensions um but they uh and others are paying the price the price of of GPUs has gone through the roof >> and so to my question on cogs do we see like a plateauing of cogs in terms of it can't get cheaper and this is the stable state. Do we see a meaningful reduction?

[00:15:18]How do >> I think what happens over time, Harry, is is that we all of us. We we improve our designs. The designs deliver more tokens per unit time. They deliver faster tokens. Now, we we are 15x faster because of architectural reasons. We will continue to to improve over time.

[00:15:40]Nvidia, they will continue to improve over time. I believe the gap will widen between our performance and their performance. But all of us, the whole industry, us, Nvidia, uh AMD, Qualcomm, ARM, everybody's chips will be better in three or four years than they are today.

[00:15:59]They will produce more per unit power and they will produce more per per dollar cost. So over time the history of our industry is a massive reduction in the cost per unit compute.

[00:16:14]>> I was chatting to a friend who is is phenomenal mind and he said that Google will become the lowest cost producer of tokens because they own the full stack from TPUs to data centers networking power procurement.

[00:16:27]Do you think that's right that that full stack ownership will lead to their highest margin lowest cost ability?

[00:16:34]There are uh pros and cons of that strategy, right? Uh the pro is you have everything from the ground, right? Land all the way up to tokens. The downside is you can only sell your TPU to yourself.

[00:16:56]And historically, volume mattered a lot.

[00:16:58]And so your market is constrained by your own demand.

[00:17:03]Whereas if you were able to sell to the whole market, you might have more more demand and be able to drive down the cost.

[00:17:13]It's an open question. Google is threatening that argument. I think your friend's argument is reasonable, but there has historically been a challenge if you only have one customer yourself for your hardware.

[00:17:27]And uh that has historically limited the size of the opportunity landscape for you.

[00:17:33]>> Do you think they should sell to external customers?

[00:17:37]>> I think you are already seeing them step outside of their own data centers for this exact reason.

[00:17:43]Right. What it says in your friend's construction is our ability to sell hardware is constrained by our ability to build data centers. Now one can imagine a world where you don't want that constraint. you would like to be able to sell hardware to anybody's data center.

[00:18:01]A and so I I think these arguments are extremely complicated. Um rarely unfold in in a simple form. Um but it is true that when Google or when Cerebrris puts our equipment in our own data center, right, we have a a significant advantage over a Neocloud because Neoclouds are buying hardware with gross margins of 70 80% for Nvidia. So the hardware in those data centers and then they have to make their margin. That's not what Google's doing. That's not what we're doing. Does that mean they're dramatically overvalued when you look at a Nebus or a Core Weave or any of the others? I think Cororeweave has been an extraordinarily innovative company. I think they've solved uh a a series of of financial challenges with really innovative sort of financial engineering. They were the first to to to use debt very innovative way. They get enormous credit for that.

[00:19:04]they have been extremely good at sort of rapid deployment which itself is a really important skill in in this environment. Um I I don't know about the others um but I think all of us have challenges as our business grows and I I think they have produced really interesting things through creativity.

[00:19:26]Now, it's different creativity than than than what I have. Um, but uh they've gotten paid for real innovation in in in financial uh thinking.

[00:19:39]>> Speaking of real innovation, I I saw the post about you running Kimmy K2.6, 6.7x faster than the next fastest GPU cloud, >> right? We we posted it while one bozo in a at an analyst firm was on TV saying we couldn't do it. I mean, if ever there was if ever there there there was a a a sort of example of being empirically proven dead wrong to have these numbers posted while you are on TV saying they can never do it. It was perfect. I was I enjoyed that. I'm a collector of examples of people being dead wrong.

[00:20:23]>> My wife has a list. My wife has a list of when I'm dead wrong. So, I I've sort of embraced this and collected.

[00:20:29]>> You should be a venture investor, my friend. With a portfolio of 30, you'll be dead wrong a lot.

[00:20:34]>> You have, if you're lucky, 80% of your portfolio where you were dead wrong.

[00:20:39]>> You should do if you're doing it right.

[00:20:41]Yeah, >> that's right. That's right.

[00:20:43]>> I agree with that. How important was that for you? And is there a stage where actually it doesn't matter being that increment more important like 6.7 times.

[00:20:54]This is so much more important. It's not 20% more important.

[00:20:58]>> That's right. You know I I think for hard problems there is no upper bound to how much faster you want to be nor the value of speed. If in three minutes we can solve problems that that take others 20 minutes, then think of all the extra problems we get solved. And think of if I'm your competitor and I'm solving your hard problems in 3 minutes and you're taking 20. Imagine over a day or a week, you get you get smoked. You will be smoked in this in this example. And that is the way this is going. Um, speed is of the essence and it's true in coding. It's true in egentic flows. It's true in every part of the uh of the AI landscape. I mean, let me just ask you this question. How big's the market for slow search?

[00:21:52]Really, how it's zero. How big is the market for dialup for slow internet?

[00:21:57]Right? How much would I have to pay you?

[00:21:59]Let's try turn it around and say there's a negative market here. If I gave you $1,000 a month to have slow internet in your home, right, you wouldn't take it.

[00:22:10]$1,000 a month, you wouldn't take it.

[00:22:13]That's how impossible it is to engage with an important technology slowly. Why do we believe that inference will be any different? There will be zero market for slow.

[00:22:24]>> I am kind of pushing them. When you power codeex and you're able to be so much faster, if you're claw code, are you not like ah bugger? They are. I think you have to be.

[00:22:34]>> Are you able to sell to them? Also, again, please tell me to sort off.

[00:22:38]>> Oh, no, no, no. Look, I I think uh right now we are digesting one of the largest deals in the history of Silicon Valley.

[00:22:46]>> Right.

[00:22:47]>> You're like, for [ __ ] sake, Harry, give me a break. I've just signed a 20 billion deal. You want more?

[00:22:52]>> You know, while we were on the road, some investors would ask, they say, "Oh, you're heavily concentrated. You have a big portion of your business with OpenAI." And and we say I talked to you a year ago when I had a billion dollar deal with G42 and you said you're heavily concentrated. You have a billion dollar deal with I said I come back to you in a year with a 20 plus billion dollar deal. And you tell me the same thing >> but with a different customer as well.

[00:23:18]>> With a different customer with a different customer. And I I tell everybody that the the way you get good at and the way you have succeed with many customers of size is first you win one. The the the way to catch right big customers is first catch one and learn build the muscle, change your supply chain, learn how to work with a large customer. Then you're in a position when the next one comes to to have a chance to win. And what's more, chance to keep them happy once you won.

[00:23:50]And then once you have that muscle, you're in a position to go out and win the next one.

[00:23:54]>> It is a huge deal.

[00:23:56]>> It's a huge deal.

[00:23:56]>> What are the biggest challenges in fulfilling it? And with the greatest of respect, do you go to sleep at night going that is quite a lot.

[00:24:09]>> Look, I I think uh what has happened and Sam said this. He said he said the first time people used G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GPT I think it was what four something they said oh this is amazing and the next day they're like how come it's not faster it the the rate at which you get accustomed to something and then want better is amazing in our industry and it used to be the case that 20 megawws was a lot and then 100 megawws was a lot and then a gigawatt was a lot and now we're running around looking for multi-gawatt facilities and that's in any other time 750 megawatts would have been a mind-boggling amount and now we're like yeah we got that right I it is the the the change in mentality over the last year or two for everybody in our industry has been sort of extraordinary five years ago if you'd have said we're engaged in a in a multi- gigawatt buildout or you think about what Crusoe is doing or think about some of the other cool companies, what uh SoftBank Power is doing, what some of these groups are doing, and you say, "That would be delusional 5 years ago." And right now it's like, "Oh, another one.

[00:25:24]Yeah, that makes sense. I mean, we should try and get our UAE Stargate at 5 gig." Oh, oh yeah, yeah, no problem.

[00:25:30]That seems reasonable.

[00:25:33]That's what's happened. It's this extraordinary extraordinary sort of change in thinking. If we are nonchalant to multi- gigawatt build hours today, >> what are we in five years time?

[00:25:49]>> It's difficult to imagine.

[00:25:50]>> And by the way, that that is exactly where where I think Sam is the best in the world. Maybe Elon is where everybody else's brain shuts down, right? when you're trying to think about 100 gigawatts or 500 gawatts, those guys, they have sort of this ability to not be constrained by sort of the way the world has always been. And that is such an extraordinary power. when you have such scale as the multi- gigawatt you mentioned there the 500 gawatt um uh does energy not just become the core crux and bottleneck that enables winners and losers I certainly think that uh people like Sam and Elon and others have said that's what they believe that at the end the we're in the business of turning electricity into intelligence and Therefore, the limiting factor is electricity. I don't know if I agree with that, but that is certainly a very reasonable view from from where we are.

[00:26:56]>> Why? What's the the bare case against that? What's the alternative argument?

[00:27:01]It doesn't have to be yours, bl that you bump into something else that that what happens in fact is our models can't keep getting smarter. That you hit something that has an assumption built in that the models keep getting smarter and you keep feeding them more energy. And at the end of the day, the models are either smart enough or keep getting smarter so that it makes sense to keep feeding them energy. That might be true. Um I don't know.

[00:27:25]>> Do you think we'll be able to build out data centers in the way that we need to and build out capacity in the way that we need to when we see that 40 out of 100 data centers are now not being built out even post approval because of local municipalities permitting dis like disruption. I AI is not popular. Harry, I think it's hilarious. People say, "Oh my god, the data centers are late. Oh my god, they're delays." Like, "Have you built a kitchen? Your contractor was late."

[00:27:55]Right? Pick a little tiny project in your home. Right? Was it built on time and on budget? No.

[00:28:03]Now imagine building, right? something the size of 50 football fields and requiring interaction with local municipalities and and power companies and regulated industries.

[00:28:16]These things aren't going to be delivered on time historically. It people's mind explodes and they never think about their own experience in their own homes. Right. Does your contractor show up every day? No. Right.

[00:28:28]Right. Does he do exactly what he says he's going to do? Rarely. Right. Do the materials, the tiles or whatever you you selected for your home, right? Do they sometimes delayed? Yes. All that same thing happens when you build a data center, right? The the generators, sometimes they're late. Sometimes they're fall off a truck. Literally, they fall off a truck and damage is done to them. Are the transformers late?

[00:28:54]Sometimes the transformers I mean that that everybody suddenly sort of throws their arms up and says, "Oh, everything's late." or they have to deal with localities.

[00:29:04]Anybody who's built anything big knows this is part for the course. This is what building is.

[00:29:09]>> I totally get that. So, you're not concerned then about bluntly local neighborhoods seeing data centers as >> Yeah. I I think our industry did a shitty job of engaging the community properly. I think Brad Smith put out a post a while ago that should have been the way we all work from the get-go, right? And it was the these can be clean. They can make jobs. They can be good for communities. We can do this thoughtfully, right? These create thousands of local jobs.

[00:29:45]Thousands. And thousands of local jobs means restaurants and lunches and hotels and a whole Right.

[00:29:52]the way they were done uh was I don't know if sneaky is the right word but sort of shielded and and wasn't open and they weren't good neighbors. Now there is no reason why we can't be good neighbors. There's no reason why the we we we can't add these to to communities and and have the the community benefit from it, right? And we have to do some thinking, right? We have all the the heavy equipment out there. Build a football field for the for the local school. Build a school at a church or synagogue to the community, right? We can be good neighbors at very very low cost. We can pay our own way, right?

[00:30:34]don't try and use sort of uh loopholes in the way power companies have historically amortized the cost of new power lines over 30 years uh and push that on the community. That's BS. We we ought to pay our own way. We ought to look after our neighbors. And when we do that, I think that the neighborhoods that that embrace this will will benefit enormously. But we we >> we didn't do a great job. I mean, we didn't do a great job as an industry at all.

[00:31:05]>> Feels like kind of the cartels in Colombia where it's like, you know, they they built the churches and they built the schools and they had such good businesses and such high margins that they could kind of get away with it because of that. And hey, great. I don't think that's right. I I I think these localities have have a resource that that isn't being used. They have power.

[00:31:26]And uh many of these land is cheap because nobody wants it, right? You're not going to parts of Metro New York.

[00:31:34]Everybody wants that chunk of land.

[00:31:35]You're going to places where the the the price of land is depressed. You're near uh power resources.

[00:31:43]And my position is that we ought to be good neighbors and we ought to be transparent. We ought to pay our own way. Now, this seems not to be very controversial in my mind. Um, and I I think the best sort of and most concise description is is what Microsoft has put forward. And we should have been doing that from the get-go. Everybody is or most communities are comfortable when their neighbors pay all their own way.

[00:32:12]And that that it's only when groups tried to to shift costs or not pay for the full resources they're using. Our data centers don't need to use a ton of water. they can recycle it. You can have a closed loop. We we can pay our own way. We can upgrade substations. We can upgrade sort of grids uh and pay for it in its entirety. Uh we shouldn't be pawning that off on local communities.

[00:32:42]Do you worry about like AI as a brand?

[00:32:44]You you see matter lay off a huge amount of people today and it's challenging to see at 4:00 a.m. emails from Zuck and you know jobs being lost. I I do worry about it. Those are those are people and they have families. I think there are sort of two views, Harry. I I I think to date most of the layoffs were AI washed.

[00:33:05]They were because we did boneheaded hiring dur during co it is actually because a great deal of productivity gains has been have occurred over the years that we're just now harvesting. The ability to gather information from across the organization to synthesize it and put it in one place is now changing what it means to be middle management, right? The role of information gatherers and presenters is being eliminated. The ability for us to automate roles and none of this is AI yet, right? That is really 90% 95% of what the in my view what the the terminations have been about. It's easy to to to put them under the umbrella of AI. Now AI is starting just now to have meaningful enterprise impact. But if if you are an engineering organization that can't see how to take advantage of vastly more productive engineers, I don't think you're long for this world.

[00:34:07]I mean, the list of things I want our engineers to do is 50 times as much as we have engineers, right? If if we get as we get more productive, we do more things, we're going to hire more engineers. We're not going to hire less engineers.

[00:34:23]Um, and can can I ask you, we we we saw Benov say that he spends 300 million a year on anthropic, which equates to about 3.8% of developer salaries on Anthropic. to make it justify the valuations that we're seeing for these companies, it needs to be 20%. Do you have any concern in that movement from 3.8% to 20%.

[00:34:45]>> No, I mean I I think if you look at and I I've never done this in any detail, but if you look at what we pay hardware engineers and you look at what the tools we we the EDA tools they use, I bet you're much closer to 15 or 20% than 2 or 3%. What's happened is historically software engineers used very lowcost tools and hardware engineers used extremely expensive EDA tools. And so that's interesting, isn't it? I mean I I I I think we the cost of bugs in hardware is so high that we became accustomed to using many expensive tools and in software we threw people at the problem rather than tools. And as AI becomes more productive, I certainly don't see a problem where software engineers are using 50 or 100,000 a year each in tokens.

[00:35:40]There are 47 million software engineers in the world. I mean that's5 trillion just in software engineering token use.

[00:35:50]What we we mentioned hardware engineers, we mentioned software engineers. What role does not exist today that you think will be incredibly commonplace in three to five years?

[00:35:59]>> So I I've been a a a part of over the past 25 or 30 years several technical transformations that produced jobs in companies that didn't exist prior to the mid90s. The role of CIO didn't exist.

[00:36:20]CIO arose as a role with Cisco with their sort of rise to dominance is prior to the mid90s the amount of enterprise networking was to minimize. There was a role that was often VP of telco infrastructure. That job is gone. We don't have a phone system. In fact, we don't have phones on people's desk. Call me on my cell phone.

[00:36:48]Right? that job disappeared completely gone and companies that built the PBXs like Rome and all these other that business has shrunk to nothing. Now later what happened in the 90 in and 2000s right with the rise of PaloAlto networks and these other security the role of cso never existed prior to that all right and what you're going to see is the the rise of roles that reflect the governance of AI in companies some companies have chief AI officers I don't know if that's what it is but as these technologies sort of become important in company's life, new jobs emerge, jobs that never existed before, new organizations exist uh where there were none and previous ones disappear. I I think the role of HR changes fundamentally. The the part of AR HR that just that answered questions that provided information about benefits that that disappears. A AIS can can answer all your questions. Uh they can provide better answers, faster answers, more thoughtful answers.

[00:37:55]um it becomes something different. The management of people become something different. I think there are all sorts of other parts of of of organizations that have fundamental changes um because AI can answer the questions that they used to answer.

[00:38:09]>> Do you agree the biggest inhibitor to enterprise adoption of AI is data structure and data cleanliness preventing?

[00:38:15]>> No, the biggest are are are lawyers. I think no really I I think the security apparatus and the lawyers who when they don't understand the technology say no we can't do it right they're in the saying no business and entrepreneurs are in the getting it done business and the the reason that uh and there there's a reason for this that your security apparatus and your lawyers I mean every they're in jobs that everybody just blames them Right. No credit, no credit, failure, blame. No credit, no credit, no failure, blame. I mean, that that's their life and it's brutal. It's brutal.

[00:38:56]>> You're selling it so well as lawyer.

[00:38:59]>> Listening go to being a CISO.

[00:39:02]It's brutally hard. Um, >> we had a year where nothing happened.

[00:39:06]Well done.

[00:39:07]>> That's right. That is their dream. I mean, every day their phone doesn't ring. They're they're like, "Oh, made it to another day." But when confronted with new technology, because their payoff structure is such that they're in the business of trying to avoid risk, they are a drag on adoption of new things. And you see this across the board and lawyers don't know how to contract for it. There's no precedent.

[00:39:32]That's in a business of backward-looking precedent. You want to make a lawyer uncomfortable. I know your girlfriend's a lawyer. Ask her to work in an area with no precedent. They don't know what to do. their whole training is about what has everybody else done before, how do we synthesize this, how do we work within those rules.

[00:39:48]So I I think the the widescale adoption and use of AI um in organizations is today limited by security and uh legal.

[00:39:58]I think once they agree we need to do this. We here here are the rules we will use. Um there's a huge amount of productivity be gained then you are immediately constrained by the way you chose to husband and marshall data the way you chose to organize data over years. And so companies like or organizations like Mayo Clinic that have been on a 30-year quest to organize data and they are at a huge advantage. Same with companies like Galaxos Smith Klein >> and other companies who who haven't perhaps been as disciplined as thoughtful about the organization of their data uh are at a disadvantage. On the security and the provisioning side, do you think we will see industries tip like legal has done where the biggest firms in the world are now going, "Oh [ __ ] we need legal. Our clients are saying we need, sorry, we need AI. Our clients are saying we need AI. Harvey or Agora." And I'm not going to get into which one, but like there's two options.

[00:40:59]Boom. Do you think all industries will follow the tipping or do you think most will follow the slow agreement that it's the new normal?

[00:41:08]>> I I think uh what's happening is the leaders are tipping. I I think even Jensen told a story that that he was battling with his own internal lawyers around the use of I think it was cursor and finally he just decreed we're going to do it and I I I think I got that right but somebody on on will correct me for sure if I got it wrong but um I I think at some point leaders weigh the productivity gains against the unseen boogeyman of brisk and that the problem with unseen boogeyman is sometimes times they're actually real, right? Not often, but sometimes.

[00:41:47]And that that's the problem. What does he call him in in John Wick? Baba Yaga.

[00:41:53]John Wick is the guy you send to kill Baba Yaga.

[00:41:58]>> I'm just, you know, I think for me, yeah, I'm not that young anymore, but I'm definitely uh capable of exuberant.

[00:42:05]>> I know you're in your 60s, but you look good.

[00:42:07]>> Yeah. Listen, it's it's a facial moisturizing routine. We mentioned um security, permissioning, legal, everything in between. They get even more freaking nervous when it's open source. Like they [ __ ] the bed. How do you think about that? I see more and more companies, especially in the valley, really push the boundaries on with Frontier and then try and get as close as possible with open source given the cost advantages. Is that the future?

[00:42:33]And what does that mean? Look, I I think we as a as an ecosystem have made real progress in sort of the the legal uh gunk around open source, but it the result has been a complexity that hurts your head. And if you if you ever want to to dive down a rat hole that has sort of no bottom, begin a discussion with lawyers about open source software.

[00:43:00]And there's no end to the depth and the boredom wi-i which you will suffer a as you head down this hole. This is made doubly worse by some of the best open source models were made by Chinese companies. And uh and they are exceptionally good models. Kimmy K2, Deepseek, Quen, the the GLM. These are extraordinarily good models. They're not quite as good as the closed source models, but they're exceptionally good models. And I I think that is a case of uh people trying to decide uh whether it makes sense to to to to save money. Uh they have been, you know, easy for us to adopt to to demonstrate extraordinary speed on. Um it's a hard problem. I mean, I don't envy the the the legal team and and the security groups that are thinking about these things, but the truth is the title wave is so big and the demand is so high that they they often just get get washed over.

[00:44:00]>> Do you think we should be selling chips to China as a result?

[00:44:04]>> No. I I think um I I think uh let's remove all of us that are self-interested and e even though I'm arguing against my self-interest, right?

[00:44:16]Um if you remove me and you remove Jensen and you remove Lisa and you remove everybody in the chip industry and and you say if we sell to to somebody in the security business and you ask this question, if we sell leading edge technology to China, will their military use it? Everybody says yes. There is no debate on that point.

[00:44:36]Their military will use it. Okay. You ask a second question which is if you sell our leading edge technology, will they will their government use it through their industry to compete with us? All right. In an advantaged way, the answer is also yes. In that and so that's where I stop. There's complete agreement that those two things are true by everybody in the security business and outside of the chip business. And now you can say that keeping them in our ecosystem is the best way to manage that problem. That's one argument. And there's some merit to that. There is keeping them from building their own their own ecosystem is something that's in our interest.

[00:45:23]There's real merit in that. I don't agree with either of those arguments, but but they're real arguments and they have real merit. They are at least today our industrial adversary. And uh as you travel the world and you see sort of the results of some of their industrial policy, the for example the uh driving down the cost of solar, driving down the cost of lithium batteries and the results it's had in their auto industry and the fact that you you travel the world and you see Chinese cars um and fewer and fewer American cars. uh they they're an industrial adversary and I I don't love that. I for years did business with extraordinary entrepreneurs there at at BU and at Tencent and DD and all these companies and they're every bit as good as anybody looking and I would love a world in which they weren't an industrial adversary and instead we were working together to solve real problems but the state of the world is the state of the world. If it's an industry, if as American industry, we sold fewer chips and we didn't sell them to China, I'm just fine with that.

[00:46:30]>> People would argue back and say exactly as you said there, if if we don't sell to them, they'll build their own capabilities and they'll get very good at it and then we won't control it. Why why do you not think that's a credible argument? I think the the the the chip industry requires you to go through TSMC and TSMC requires you to go through ASML or or Samsung. I think there are reasonable choke points to manage those challenges.

[00:46:58]I think the strategy in any case is even those I think who who disagree with me would would suggest that uh you don't want to sell them your cutting edge technology. You want to keep them down rev. I I'd like to keep keep my my industrial adversaries more than down rev. With that, how important is it that we onshore TSMC like capabilities and companies given Taiwan's vulnerability to China? We have problems in the US in long range policy >> policy that endures more than a single administration.

[00:47:37]Right? We have problems building infrastructure that is clearly needed and crosses municipality lines. So uh let's look at things China has done extremely well. Their power infrastructure is extraordinary and in the US we are a patchwork of 1950s technology if we're lucky and that's really bad. What was the the fundamental question? I lost my train of thought.

[00:48:03]I'm sorry.

[00:48:03]>> It was how important is it that we onore TSMC capabilities given the vulnerability? So there there are things we don't do well and uh one of them is thinking about long-term consequences of uh decisions like not investing in fabs in the US. I mean we didn't just lose the fabs, we lost the surrounding ecosystem. We lost the packaging expertise.

[00:48:30]we we we lost a whole set of of surrounding uh strategic jobs and industries and it is extraordinarily important we get it back and I've been saying that for decade and a half that not the chips act not subsidizing Intel not it's important that that we have cutting edge fabs in the US and that we uh surround them with cutting edge packaging technologies and these are a strategic asset.

[00:49:02]>> If I said that you have one policy change that you could usher through with no resistance, what would it be?

[00:49:10]>> Um I I would allow TSMC and uh Samsung uh both to uh a 20-year period free from all local and legal local ordinances all of them to build fabs in their desired location in the US. If that's Arizona, that's great. If that's Texas, that's great. 20 years, no local rules, no allow them to build fabs. And I I would say that we use the same rules you use in Taiwan.

[00:49:41]>> Don't don't build garbage. You use exactly the same construction techniques and rules, etc. that you that you've built fabs successfully elsewhere in the world. But local ordinances are disastrous and not intended to cover pyramids, right? FABs are modern pyramids, Harry.

[00:50:02]I mean, they are the greatest things humans make in manu in the manufacturing world by far. Nothing's close. Can I ask Andrew, I sit here in London. Should I be worried? And you have the best frontier labs in the US. You have amazing open source and amazing manufacturing capabilities in China.

[00:50:20]What does Europe really have? We we've kind of failed on the model front. Mral is the leader, but sadly no nowhere near others.

[00:50:29]Should I be worried? You should be worried at the pattern the pattern of sort of lack of success across a range of technologies. It it's not just that that the leading AI companies are most of them are in the US but the leading chip companies but the leading software companies right of course there's some examples SAP and and some others but there has emerged in Europe a uh a sort of be afraid of it then regulate it tax it or or sort of mentality that that works against entrepreneurship.

[00:51:11]and and I think Europe uh and this isn't true across the board and clearly there are pockets outside of Cambridge and and in London and and in in in Stockholm where the guys at lovable are doing really interesting stuff and there all sorts of counter examples but on the whole given its population the opportunity to do vastly better on the innovation front across industries is sitting there unexed and uh that I think is is They worry.

[00:51:41]>> How much of your business do you think will be in Europe in 5 years time?

[00:51:44]>> I think along with this they have been slow to adopt new technologies. Not only have they been sort of slower to invent new technologies, but they've been slower to adopt new technologies.

[00:51:56]And so I I think the fastest adoption will not be in Europe. I I think but in the the 2 and 1/2 to 3 to 5 year range, it will be a meaningful portion. Is is that in line with your experience? I mean my experience is from a long way away and from visiting regularly and talking to customers. Is that your experience?

[00:52:16]>> Application layer. No, I think we have some of the world's best companies whether you're 11 labs or your synthesia or your deep mind. U I I think 100% on the infrastructure on the chip side on the model side unwaveringly so. So yes, in large part with a little bit of nuance which you to be fair added there with lovable and so hotspots. So I think we're totally aligned there in respect.

[00:52:42]I think you've done real work to to to argue against that and hats off to you and the others in the venture community.

[00:52:47]I I think capital plays an important role. I think a culture in which it's okay to fail plays a role that that that is uh not traditionally in Europe.

[00:52:59]careers are at one company and are long and that breeds a conservativism.

[00:53:05]I think one of the most powerful parts about Silicon Valley is the absence of a stigma. If you try to do something extraordinary, crash and burn, um VCs don't hold it against you. They they ask you what you learned and often it's great experience and a credit to you. I I think that is something that that I've not I don't understand it its history but is clearly present. Can I ask you before we do a quick fight? You mentioned the IPO at the start.

[00:53:32]>> You timed the IPO with the greatest of respects in my mind to absolute perfection before a SpaceX IPO, before Anthropic or Open AI. Was that strategic and deliberate with the greatest of respect or was it relative luck?

[00:53:47]>> No. Let let me share it was 100% deliberate. We tried to go public a year and a half earlier and we couldn't get it done because we bumped into CPHAS. We were 10 years old. We tried to get public for years of it. It was 100% luck and grit and sort of a relentlessness and an unwillingness to fail. And I I think that uh >> but did you have in your mind the other IPOs and when liquidity would be best, excitement would be highest? Did did we know uh when we set the date that chips would be on a run and that it was impossible for for for XAI and OpenAI and etc to get public before? We didn't know any of that when we set the date.

[00:54:34]Um but what we did know uh was that we had a chance to be the first and only AI pure play in the entire market. There's only one and that's us. and we had a chance to bring an extraordinary growth story to public market investors who had been shut out. We we tried again and again and that's how you get lucky Harry is smart, hardworking people, relentless work. They get lucky and occasionally they they they find the perfect timing.

[00:55:07]>> Should you be investing in companies building on top of you? You said about trying and trying again. You know, Jensen said before that he wishes there were companies he had invested in and about Bunny investing in the ecosystem around Nvidia. Do you think Cerebra should be investing more aggressively in the application layer built on top of you? I think that's an opportunity that is newly available to us.

[00:55:30]>> I I I think uh probably not with venture dollars or traditional venture dollars, right? I I I think you have to think very carefully about your investors.

[00:55:41]Um I I think when you're using venture dollars, the question is should we be investing in them or should our venture partners be invent investing in them?

[00:55:48]With public dollars, the the the mandate is different and your investors have different access. And so uh the opportunity for for us to do really interesting things with our our our customers and our partners grows. That includes acquiring companies, that includes investing in companies, that that includes different structures of partnerships.

[00:56:10]And we we have to explore them all.

[00:56:12]>> You mentioned the multiple times trying to go public and the persistence.

[00:56:18]What do you know now about going public that you wish you had known when you were trying multiple times? No, look, I I think what happened was uh we bumped into a a CPHUS challenge that was sort of sort of obstructionist and uh the there were sort of unnamed concerns that that never got articulated that that sort of lived in the ether uh about uh some of our our large customers. Um, and then we got a new government and those concerns disappeared and we were able to move through it really quickly and thoughtfully with a a really fair resolution. And by the way, a resolution that we had proposed a year earlier.

[00:57:05]>> Is the Trump administration unwaveringly better for business? Again, I said here in the UK >> are unwaveringly better for business.

[00:57:13]Um, you know, there are things I agree with, there are things I disagree with in this administration, but unwaveringly better for business. You got to be at bat taking swings and you've got to be building every day.

[00:57:25]When we got public, we were a much stronger company. We had larger sales.

[00:57:31]We had we're further down our road map.

[00:57:33]We had better customers. And so you you sort of have to separate you. We we we we didn't get public because of we kept building the business. The business got better and better and better. And that gave us the opportunity to try again.

[00:57:47]And that's I think the message to to your builders, to your audience who who who builds companies is a lot of stuff will happen that is not in your control.

[00:57:57]Right. There will be bad times. There'll be Right. And and there will be, you know, I was raising money in the summer 2008. Yeah, that's right. That look on your face is exactly right. Right.

[00:58:08]Summer 2008. Bear Sterns falls apart in March. Layman Brothers is exploding in September.

[00:58:14]VCs didn't want to put money to work.

[00:58:16]And you know what? The only thing we could do, we could keep trying and keep building.

[00:58:20]>> I was 11. I was playing Pokémon, dude.

[00:58:25]>> That's right. When you were out of nappies, uh, we we were out raising money. Uh, and what you can do is run with the things you can control. And you are always stronger if you keep building. And if you keep adding customers and you keep moving your technology forward, you keep adding space between you and your competitors, that's what you can control. Good times, bad, that's what you're in charge of.

[00:58:53]>> I have to move into a quick fire cuz I I >> uh number one, dude, what have you changed your mind on most in the last 12 months?

[00:59:02]>> As you prepare to go public, the the number of people who call you and try and sell you stuff is insane. just I mean suddenly developing a presentation wi which should cost $20,000 is a $200,000 project. Suddenly you get 20 emails a week about wealth management.

[00:59:22]Suddenly you get I just the garbage that that sort of it's like when you get married it's the same. You want a photographer to do a corporate event $3,000. You want a photographer to do the exact same thing only you call it a wedding?

[00:59:37]three times as much. Same for a caterer.

[00:59:41]Same for everything. Right.

[00:59:43]>> Why?

[00:59:44]>> Because you can't put a price on love. I know you're >> That's the same reason. No, be be because um they can. And that's something that I didn't expect and it's sort of uncomfortable.

[00:59:58]Uh it's sort of the number of people trying to take a little nibble of your IPO and get paid on it. That was a surprise to me. I I didn't really think carefully about that prior to getting out the door.

[01:00:11]>> I mean, listen, you touched on Europe from an American's perspective. If I touch on America from a European's perspective, there's always a take. It's like it's always about the money in America. The transa the money, the money, the e Oh, it's like in we we have a problem with that in our society. I think that's right. I I think it is both the the source of some of the drive and the entrepreneurship and some of the source of of the uncomfortableness.

[01:00:37]>> How how did money change you as an entrepreneur? Like it changed me as an investor. I go for way bigger upside.

[01:00:43]I'm not so fearful of losing money.

[01:00:45]>> I grew up on the Stanford campus and the only currency was intellectual horsepower.

[01:00:54]I mean, my dad's tennis match, he played doubles every Saturday and Sunday, and they were like six or eight guys in rotation. And I look back and four ended up with Nobel prizes, and one had a Fields Medal, right? You know, William Shockley lived next door to us. Dude invented the trans transistor. And what we knew about him growing up was on Halloween, he gave full-size candy bars. That was what we thought about as kids. after I sold my last company, nothing changed. Nothing.

[01:01:26]Um, uh, nothing's changing now. I I think what's made me proud, what made me proud in my last company is we we made 100 millionaires. What made me proud in this company so far is we made 800 millionaires. And that if you if you don't like doing that, you have no business being suit. If you don't like delivering for your team, you're not a real leader. And uh that feels good every day here.

[01:01:53]>> 800 millionaires.

[01:01:55]>> 800 millionaires.

[01:01:56]>> Yeah, that must feel pretty great.

[01:01:58]>> Feels pretty great. And these are people who bet many of them bet long periods of their career with us, right? I mean, and maybe you get 35 years of their career as a top working engineer. Many of these guys have been with me for three or four companies. Some of them been here 8 n 9 and a half years.

[01:02:13]>> We've spoken before in off record about kind of personal lives. I'm intrigued.

[01:02:16]When you are a public company CEO and you're going public, the world wants a piece of you. You're public now. You're public. Any advice on how to sustain an amazing marriage and an amazing relationship while also being a public company CEO and going through that process? I I would say that pick a wife with patience. Pick a partner, a husband or a wife partner um who understands what it is to to be an entrepreneur. I I don't think and I look at my co-founders and and our leaders. It it is every day when you're a leader of a of a startup, a pressure test on your soul. Every single day. And if you're a real leader, when you are 30 people, a little a little company picnic, you look out and what you see are mortgage payments and braces that need to be done that you're responsible for. And that doesn't change.

[01:03:13]And uh I I think that if you really believe that and you you hold that in your heart every day, you carry real weight with you. And I I think uh you have to share that with your partner so they understand. It's really hard if they don't. I think almost everybody and and maybe your your your your partner has felt this and I think every CEO I know has told the story of their partner telling them that they're more lonely when the you're sitting next to them thinking about work and your mind is just ripping on work than they were when you weren't in the house. And I I I think that that that what we do is a family thing. There's a price to be paid in how often you see your wife. I mean, I'm on the road 3 weeks a month. I mean, put it this way. Uh Emirates Airline sends me a Christmas basket.

[01:04:13]This is an Arab airline sending a Jewish guy a Christmas basket. You know how frequently you have to fly for that to happen? It takes a toll. And I I think you have to think really hard about how to put some credits back because otherwise they're just a scream of debits against your relationship.

[01:04:29]>> Final one for you, dude. What's the kindest thing anyone's done for you? You know, we see a lot of uh whether it's your investors publishing on um your IPO day and oh, I met Andrew once at a coffee shop, you know, oh, I opened the door for him once. I was part of Cerebrus. Um >> yeah, >> what's the kind of thing?

[01:04:47]>> I think uh and and this is for you, Harry, and and the VCs is to have empathy for how hard our job is. And I think one of the the things that I was really lucky with was we had a board that that understood they didn't need to put more pressure on us, that if the pressure doesn't come from within, all right, they bet on the wrong people. You you know, hardware is is extraordinarily difficult. And we had and and we attacked a problem that had never been solved. And we had an 18-month period where we were spending 8 million a month and we couldn't build it. Yeah. 8 million a month. We were burning for 18 months and we couldn't solve the technical problems. You know what it's like to have a board meeting every six or eight weeks and come back and say, "No, I can't do it. Still can't do it."

[01:05:38]Um, and they were >> Did you doubt yourself at that point 18 months?

[01:05:44]>> Of course. I I think there's this myth that CEOs don't doubt theirelves. It's not driven by relentless fear of failure. Of course, of course you do.

[01:05:53]But I believed in the methodology we were using. I believed that that each time we failed, we learned a little bit and we didn't fail the same way again.

[01:06:03]And that that where it started, we failed in the first two seconds. And then then a year later, we were failing at at an hour.

[01:06:13]had each time we we did a full failure analysis each time I mean every single one we failed at and for 18 months and I think uh that's some of the proudest work of my career was that problem and nobody else to this day has solved it nobody else knows how and I I think you can imagine getting back to your previous question I wasn't a peach at home right I I I I I wasn't chipper. I wasn't light. I wasn't happy. I was failing every day at work, every single day, and for a long time. And I I think if you want to attack hard problems, you have to come to grips with that. You have to learn to manage it. You have to surround yourself by people who who who you believe in, who you want in the boat when the hardest problems are are present. And I had all of those things.

[01:07:10]Um, and my wife was an extraordinary partner.

[01:07:13]>> Dude, listen. I so appreciate you. I so appreciate you putting up with my incredibly weward questions from partnership.

[01:07:19]>> Very interesting. I I think Harry, you're you're an extraordinarily good interviewer. Um, you can cut that part if you want. Um, >> no, I loved it. That's fantastic.

[01:07:28]I trust me, we're going to start the teaser is Harry. Thank you.

[01:07:33]Extraordinarily good interview.