China’s catching up to US AI… Here’s why it won’t matter

22:40And so you already had an ability to do more CapEx, buy more chips. and in the context of the moment, we had that moment just before the Ant Financial IPO, which was pulled, where the Chinese government really sat on its entrepreneurs and a chill came into the Chinese market. Their valuations declined, the access to capital declined, the venture capital dried up, the balance sheets weakened. And in a way, I'm sure someone, a historian will look at this, but my argument would be that more constraints were applied at that point. Because even today, with the U.S. hyperscalers and the big tech firms spending$320 billion on CapEx, a little bit more than half of which will go on GPUs, that's above the capabilities of the Chinese big tech firms anyway.

23:27I'm sure that it's been very useful, particularly the ASML thing, which actually, in a sense, prevents China or slows China's ability to build its own semiconductor industry, which has been a strategic priority for many years. but in terms of procuring chips there was just much more money in the u.s absolutely yeah i think i mean to some degree this also might largely explain where we have like more frontier companies here right so i do i would like i would not argue against it that it's not one of the reasons but even if you have the money and you can just buy it like less good chips that's definitely going to hit you in the long run right and again depending on where this goes i think the point about extra controls this has been that confusing thing with deep seek it takes some time to buy it and we had many failures right the first times they got the numbers wrong Those are the chips they train DeepSeek on, right?

24:12So we basically have only working X controls since, what is it, October 2023? It's a little bit more than a year, right? You don't build a cluster every new year. So I expect we'll see like a major impact going forward if certain enforcement and things are being patched. And again, all of this also depends on will NVIDIA's chips continue to get exponentially better, right? To some degree, the good thing about computing is it gets exponentially better. Who cares about the computer built five years ago? Nobody uses that anymore, right? It's like my chips now are like 10 times better and I produce 10 times more of them.

24:42But if this would level off, then this might look actually different. You do hear, I mean, you hear Jensen say, I'm the chief revenue destroyer in the sense that he keeps bringing out new chips that are better than the old version. But you see the hyperscalers saying, we're sweating chips at six years old and seven years old, and they're still sort of fully used. In a sense, they have to say that to the market. What we have here, though, is we have a challenge. We have a real bifurcation here, which is national security needs, which in a way predicated on how good the frontier models can be. And there are the economic needs, which are predicated on the ability to deploy and diffuse high-quality, reliable AI into your firms.

25:20That feels like it's a really difficult pair of objectives to optimize for. You know, there's a line, you know, he who chases after two rabbits will get neither. How do policymakers make sense of these things? Because they're quite conflicting. They are. I mean, this is the concentrate of, right? Security comes at a cost. But this is also known for every company. And every company has ideally a CISO who takes care that your IP is secure, right? But this isn't so much about security coming at a cost, right? Security can be having the equivalent of biological secure levels for your labs, for AI labs, and that is a cost that comes out.

25:59This is something that's slightly unknowable, which is will scaling over five years create this super intelligence that builds self-reinforcing loops that take you up and away. And if it does, it's an existential national security issue. And therefore, you need to think in those terms. The other part of the argument, which is the economic argument, is that it's broad diffusion and use that matters. And it feels like those are really slightly at odds with each other. Yeah, I mean, look, there is like a view you just laid out, like it's a national security, which is like, well, the build AI, it takes off and then it's going to kill us all.

26:31I I think that's a national security risk. I think there are many other national security risks which are between AI is all good and between AI is going to kill us all. There are many different risks, right? We talk about cyber risk. We talk about biobiological misuse. We talk about loss of control scenarios, right? And they all need to be managed. And I think they're actually not too different from like normal security risk, which we're trying to manage all day long. If the US is trying to decide who to sell fighter jets to, pretty similar on this one, right? And then you're totally right. Eventually, if you decide to not sell chips to China, you might lose revenue right that's just the case i think this is just like again what the government began it has been bipartisan consensus why has decided it was like yep we take this revenue in but eventually in the long run that's the thing we want to do here because ai might go to kill us or china might take over the world whatever your favorite threat scenario is i think there's like a polarity of like national security arguments why people are doing this and again yeah my job as a policymaker is like to calculate the costs of both right all i would say is our industry has been doing pretty fine so far, right?

27:31And again, when we talk about export roads to China, that's one market, there's still the rest of the world. The majority of the market is still available. And if you just compare these semiconductor exports, they're minor compared to what's happening right now with the tariffs, right? This is the way bigger economic deal. And it's still being done for many reasons, right? I want to sort of dig into this a little bit because, you know, in a way, there is a cost to export controls. There is a cost to controlling this. The cost comes as follows. it's about where the government chooses to intervene in a market, which has a doctrinal effect on where a nation operates and thinks about itself.

28:05It has a fundamental impact on the certainty and the risk profile of a particular industry and therefore its ability to secure capital. It has a cost of pallor over any country that becomes dependent on, say, for NVIDIA chips where they see that there is a line that is easily movable at whim in pencil about whether they'll be able to get updates or support and whether that investment in CUDA and that infrastructure goes. So it's more than a blanket, sort of simple, clean clinical cut. It's also one that spills over in many different directions. Then on the other side, we have this issue of to what extent does it stimulate innovation in other places, in China in this case?

28:53We haven't really talked about Huawei much, but Huawei came out with this cloud matrix, server unit. It has a sort of performance that's similar to NVIDIA's top-tier system, except it's four times as energy expensive. There are these other calculuses, right, that say, what are the costs that we have to bear? How do you think about that? Let's go back to the Huawei system, I think, because it's a nice example of exactly the point which I'm trying to make. They will be able to build systems which are, on paper, similar to the American system. The difference, just like, as you're saying, it eats way much power.

29:25It requires three times more chips, right? And, like, you have all of these inefficiencies. Does this mean they cannot not train, let's say, a GPT 4.5 model? Absolutely, they can't probably. This is kind of what I'm expecting, right? The difference is just, like, while TSMC is churning out five million chips per year, they're churning out less than a million chips and they need three times more chips to achieve the same performance, right? This is exactly the argument which I'm making here. And then you're right. There are costs to this, right? These export controls, they used to say small yard, high fence.

29:54I don't think that's the case, right? To some degree, when it's originally started in 2022, this was before Chattivity. And I give the government like a lot of credit for foresight here. Like knowing that computer's cool, if like people would have invested in 2022 into NVIDIA, they were still early. So they had this foresight. And then it was just like, oh, we only hit on these advanced AI chips which are being used for servers. Sure, AI was a niche application. But again, if you look at people like us, we believe AI is the future. AI is going to be everywhere. Then this is not a small yard anymore, right?

30:23If I use AI to literally plan my day, cook my food, and everything else gets organized around me, then you have collateral damage, which hits on the broad economy. And this is just a thing which is simply true, right? And this is generally the challenge of governing dual-use technologies on these kinds of things. What I would say, though, is China still has access to AI services. that can still use Microsoft Azure, that can use ChatGPT, right? So it really depends where on the value chain you want to intervene. What the S-government wants to do is like, hey, you shouldn't be able to build your own competitive big systems and like you make human rights abuse and a bunch of other nasty stuff with it.

30:59You want to use AI, which is like monitored by us. And like, again, you cannot build biological weapons. But sure, go for it. That's fine, right? And I think that's the best way to think about it. And here, again, the key is, chips can't go into China. they can access cloud computing right now without any restrictions. And any engineer, it's not like I need a computer in my basement to train a model, I go to a cloud provider. So right now they can't technically do it. You're totally right. They're the whims of the US government if they decide tomorrow to cut them off the cloud. That's a problem.

31:29But again, that's historically been the case for the monetary system the same. This is just how national security works and how you try to govern, right? And those things are trying to be balanced. But I think eventually we can find a good middle ground here. And ideally, it's not only the US doing all of this. It's the U.S. and its allies and partners with the semiconductor industry, the key allies there, building a broader multilateral governance framework, which is building with democratic values. That's at least what I'm aspiring to. Let's talk about technology change in all of this, right?

31:56We're having to hold a couple of distinct ideas in our head. One is this idea of the very powerful national security model. I'm not thinking about one that kills us, but I'm thinking about one that runs the million drones rather than the 20 drones. The other is the economic side of this, which is the ability to scale this out across an entire economy and all the workers and give them all superpowers. One of the things that strikes me over a five-year period of time is that isn't long enough for certain things to change. I mean, the CTO of AMD, Mark Papermaster, has said that by 2030, he expects most inference loads to be run on device, right?

32:30So on the smartphone, because in reality, to undertake 28, 20, 29, 20, 30 tasks, the quality of a smartphone chip and the RAM that's on that device will be enough to do that, to act as a software coding agent and so on and so forth. So to some extent, at that point, the battle is not about these super sexy$40 ,000 GPUs or their competitors. It's about what can exist on device for which China can build lots of devices with great ships, that's the mobile phone industry. But it might also be about new architectures that emerge predominantly for inference, right? Because sometimes we think the training and inference are the same, but they're not.

33:10They require different things of the chip architectures, the memory bandwidth, the latency, and so on. So that also changes the way in which the leakiness or the effectiveness of export controls, given what's needed for the inference and the deployment. Absolutely. Well, I mean, we just talked about history. We talked about going from two chips to 100 ,000 chips. The two chips was not a transformer architecture. At some point, we developed transformers. At some point, we had a new loss function. Me as a technologist, I love when people develop new papers, but I'm interested in the macro trends.

33:41That's why I just love like line going up, compute, Moore's law, more transistors per area. Generally too. How do we do it? Completely different architectures over time, right? If you find the right friction layer, again, you see these exponentials being the case there. Then some people are arguing like, oh, we have a new architecture that does this. And they're like, true. That's what happens all the time. This is what we call increasing algorithmic efficiency and compute efficiency. I think it's quite unlikely that somebody tomorrow pulls out a new architecture and says, oh, look, AGI on a smartphone.

34:10I don't think that's the case. I think we would first build AGI on a big cluster, and then a couple of decades later, we would build it on a smartphone. This is just what we always see for these kinds of things. So the thing to acknowledge here is diffusion, right? A given capability gets cheaper over time. I used to run it on a big server. No, I can run it on my smartphone. That's definitely the case, right? So yes, we will see more workloads being on edge. For privacy reasons, We see Apple being bad on these kinds of things. But even Apple is kind of acknowledging. It's like, damn, there are certain questions you can't do them locally.

34:39You're going to run out of battery power. We sent them up to the cloud and they came up with a really fancy architecture to preserve privacy. That was super fancy. Right? Yeah. And these are the things I'm expecting, right? So I'll usually say, if I want to set a reminder, sure, my LLM on my smartphone can. If I ask for the meaning of life, well, hopefully it goes up to the server and it's processing and generates five books and tells me what the meaning of life is. Right? So I expect this trend to generally continue, that we just see these kinds of things. While, again, one curve is going up, the best capabilities will always be in the bigger servers, roughly, whereas a given capability goes down over time and will diffuse across.

35:14And that's definitely a national security challenge. You're totally right. If GBP5 poses a national security risk, we can control it for, what, a year or two? And then every guy in a garage with five Mac minis potentially could reproduce it. And that's just the thing we need to deal with eventually. Well, and eventually may, in that case, be very soon if we're talking about GPT-5. Okay, we've got a few minutes left. I want to jump into just some thinking about the future and what the future ought to look like. When DeepSeek released its R1 model in V3, we're not much fanfare from DeepSeek, but lots of people were very excited at the time.

35:47One unique feature was that it was an open source and open weights model. And Marc Andreessen, who is a Silicon Valley venture capitalist, really behind the Build America movement, viewed this as a gift. I think I forget the exact text of the tweet, but it was something like, you know, this is a gift to humanity or something similar. And what that does is that helps to diffuse AI capabilities because anyone can get those weights. They can do things with them, customize them, and build their own applications. And at this point, you know, Google and OpenAI have been getting progressively more closed.

36:19So there is this question about, as we look at the future, between a world where compute resources are highly concentrated, concentration lends, of course, to control, but in many places, typically in political systems, we don't like concentration of power, right? We like decentralized power. We like checks and balances. We like some competition. We also like that in economic systems, right, in markets. There is this simple two by two, right? Maybe it's just two boxes, two parts. a world of compute concentration and a world of compute abundance and a world of closed source models living with that concentrated compute and a world of capable open source models living on more available compute.

37:02What are the considerations for each of those scenarios? You know, how should people making their own mind up on what will ultimately be a political choice for us and hopefully think about those two? Well, that's indeed the question to which it will be a political choice for us, right? Or if one needs to intervene. I think the best way to start with is just let's look at the status quo. Most compute is largely concentrated. Where is it concentrated? The large majority sits in industry. This didn't always used to be the case, right? Academia was really strong on traditional high-performance computing.

37:29This lead was eradicated over the last decade, basically. Then secondly, where is this compute concentrated? Within industry, within our hyperscalers. Amazon Web Service, Microsoft Azure, Google Cloud, Oracle, they own a large majority of it. If you look at NVIDIA's shareholder output, they need to tell if it's mostly for big customers. That's what you see. That's simply the case. These are all American companies. Which countries are the most concentrated? Most in the U.S., bad news. We just published a data set where we see, I think, 60 % to 70 % of all supercomputers are in the U.S., right? Like Europe, 10%.

38:04China's another 10%. Something along these lines. I think roughly this is also what I expect. All of AI computer results being there. So that's the status quo, right? So if people just say, like, power to the people, they're worried about more concentration. I was like, look, the status quo, it's already concentrated across all these companies. And there's a good reason for investigation. I think there is ongoing investigations into the cloud stack, right? All of these cloud companies have just been getting bigger and bigger and bigger. Again, I'm not an antitrust lawyer, but I think this is at least something one should look into what we could do about it.

38:32So those are the computer resources being concentrated. Argument against it is like, this doesn't mean who's using it. We both right now use computing power by Substack. Well, I don't think Substack has a cluster. It's probably Amazon Web Services, right? So it's something that I can just access it. The world is just like, what if they cut us off, right? What in the future, I need my AI lawyer, and they turn my AI lawyer off, and I'm just like on my own. That's a big problem, right? Right. And I think that's where we just need to make sure, like, which role will AI play in the future? And then I think it's even more important.

38:58I don't want compute in the future. I want AI. If I have 10 GPUs in my garage, but I don't have the leading AI model, that's the way bigger problem, right? So I think that's the thing which we need to think about, like what kind of services do people need access to, right? And then is it a compute with the AI model or not? If my mom needs a lawyer, I'm sorry, she's not going to set up a cluster and spin up the new DeepSig R2. Ain't going to happen, right? There, I think it's like way more likely you potentially want access to AI systems. Even here right now, whoever can afford the more tokens already has a lead on this, right?

39:30There are more countries, of course, than the US and China. So this also explains in a sense why the EU and why India and also the UAE are pushing for their own compute infrastructure. And that can either be physically owned by local companies that fall under their jurisdiction, or it can be the right kind of contracts with a hyperscaler. And then you've got this emerging tension that we've seen, and certainly Microsoft has come and spoken about, talked about it in the last couple of weeks, regarding Europe and the sense that their assets in other countries, sovereign nations now want to know under whose legal purview does this reside?

40:09How certain can we be of this? So this sense of this concentration risk is really being noticed increasingly over the last five years. I guess one question is, how do you imagine it plays out? You're totally right. You see more and more countries kind of waking up to this computer, right? Like I just said in the beginning, computers survived in San Francisco. I was just in Brussels. It's all survived there. They just announced an AI factory, right? I was sort of hanging around London around the time they announced a future of compute review in the UK. So like, I think people woke up to this compute idea earlier.

40:40And to some extent, it's easier. I think as a government, it's easier for me to set up a cluster than to actually build a leading AI company. This is my worry. I'm just like, guys, I don't want you to overly focus on compute, where there's a kind of force leading people to compute. It's like, wait, are you actually solving any real problem here? You need a whole tech stack strategy. Do you want to build a different AI company? Sure, go for it. Do you need more energyless compute? There's a problem here that I think is really hard for these governments to wind their way through, which is that the AI layer is going to be the fundamental infrastructure layer of the global economy.

41:12AI systems will be our interfaces as citizens, consumers, whatever role we take, to the bulk of the digital services that we need, and increasingly so. And so if the iPhone ended up being a political issue, which in a sense it was, it's something I wrote about in my first book a few years ago, the AI layer will be even more so. And I think that governments will have to figure out how they can articulate what they need and expect and what kind of guarantees they need from that layer, especially as you say, it's concentrated in American hyperscalers delivered admittedly through their European subsidiaries, but ultimately all routes lead back to Washington in some sense.

41:55And that seems like a really difficult, thorny problem for them to address over the next few years. Absolutely. I mean, you were saying it's actually not a new one, right? When Snowden came out, I was like, oh, god damn, all of our data is going over there. We want more sovereign cloud. So to some degree, I think it's not a new problem. It's just repackaged with a new tech layer on top of it. The thing where I think it just differs, I do think AI is a fairly big deal, probably the biggest deal in our century. So we really need to get it right here. So I'm excited about all of these governments having these plans, right?

42:25And I want more techies in there and doing it. But I'm just saying I have not seen a sovereign AI strategy, if you want to call it this way, which actually makes me feel like you're actually tackling the root cause here right and eventually what i believe here is i don't think the plan is like every country is going to like redevelop the whole tech stack and come up with down solution i think eventually what i want again maybe i'm a little bit too much kumbaya it would be nice if we can just get together and just say like here the standards here's how the democratic world is doing ai and be like have these scales right to work on this and like you're not at the whims of somebody just shutting off your ai tomorrow because they just feel like it i don't think that's great i don't think that's how we should go about it That is super kumbaya, so I'm going to have to pick that up.

43:05Number one, what did you have for lunch? I want some of that to feel that great. But number two, I mean, the issue that you have there is this sort of original sin of the internet, which is that we never really figured out who was going to ultimately make the rules. And it was fine in the late 80s and early 90s when it could be, you know, IETF and IANA and so on. But now it's a much, much bigger issue. And the decision ultimately needs to rest from the view of a national government or the EU, which means that for that to be cast iron, those rights have to be subsidiarized by the U.S. government in a way that is non-repudiable to whoever else is using that.

43:41And I actually don't know what that mechanism is anyway. And even if that mechanism existed, I'm not sure I can see the political circumstances in the next few years where it's politically tractable. So let's start with that first question, which is what is that mechanism? I mean, we made progress on these kinds of things, right? There are certain data localization laws. And this at least stops other governments to literally raid the data center, grab the hard drive, and take your data. There's also encryption. So we've got methods to something to defend. And there's nothing new. We've been doing this with different types of cloud acts, initiatives, and more.

44:16So that's already going on. I think for AI, we want the same. And ideally, you just want to be part of the broad AI supply chain. I think to some degree, we're having these discussions because Europe is just like, where is Europe on the tech stack? I see it all the way below with ASML. right where they're producing the machines which produce the chips but across else it takes it looks like not that well i don't think the u.s government can say tomorrow like oh actually taiwan netherlands south korea meh we don't want anywhere then they got a problem right because these are key allies in the semiconductor supply chain so i'd be more way more keen for countries to pick like their part in the supply chain right this is just how it works be it a local community be it on a global scale to pick it there and then localize certain things where they feel like They've got a say or they've got potentially some bargaining there if they eventually needed to.

45:01And then again, I think just building data centers gives you at least the physical control over its computing power. And then we should just ask the question, which part of the tech stack do you want to localize? What is the most important? I don't have the answers here. I'll make two comments about that, right? So the first is I like what you've said about pick a point where you can be the winner and you can be the best. And therefore, everyone needs you in the team. it's a little bit like how the primes in the US, you know, build even a hammer in 39 different states. So that there's always a congressman fighting for your size.

45:29And, you know, the wood is done here and the plastic is done there and the screws are done there. You could, you know, argue that. But of course, as you say, three countries right now are part of that, which is the Netherlands, South Korea for memory and Taiwan for the chips. So that is quite an interesting approach. And you're right, I haven't seen that in any sovereign AI strategy. And you've read more than me. But on the second point about the physical data centers, I think one of the problems there is Moore's law or Huang's law, whatever you want to call it. Because essentially, given the price performance improvement of chips every year on a static amount of CapEx every year, you're sort of adding the same amount of compute as you already had.

46:08And so even if you have physical data centers and you can't swap out the chips because the chips are sort of physical, they have different cooling requirements and energy requirements, you have to build new data centers for the new families of chips. Even if you do have the data centers, in two years, if you get cut off for whatever reason, in two years, they're going to be anemic and they're not going to be able to run anything and they're going to be irrelevant in that new modern AI economy. So you have to be able to guarantee the consistent supply of uprating of the chips. And it's not just going to be hyperscale data centers, these big ones like Stargate and Colossus.

46:43It will be MetroScale and Edge data centers as well that will need kind of constant replenishment. So it feels like that may be an indicator of your commitment to be a customer and therefore create incentives to not be cut off. But it doesn't feel like it's as strong a defense as the proposal that you have, which is let's become the best photoresist in the world, which I think exists in Japan right now. I mean, two points. Let's flip it around. You just said, well, you build a data center and the next year is outdated. Nice thing is, you can enter the race pretty late. You can get a good share of compute.

47:14If I now build a data center with B200 and Vitea's newest chip over H100, here we go. It's way easier. And if you look at these compute shares, and again, me crunching all these numbers, You just see how, like, to some of the volatile they are, right? If you just build a couple of big data centers, you're suddenly in lead. You're right. It's not guaranteed. Next year, it changes again. But it goes both ways here, right? And then you're right. You need to always upgrade and get the newest chips. This is where the semiconductor supply chain, again, plays a key role, right? And again, the U.S. government just identified AI chips.

47:44Oh, nice. They're like these four countries which play, like, this important role here. We got, like, this nice authority. We can just decide where it goes. But again, why did we have all of this outcry, for example, like Poland and the diffusion framework not being one? same with Portugal, it's exactly for this reason. The vibes are off. The vibes are off if you just feel like you're subject to and they can cut you off at any single point in time. And it's a product you literally need to replace every five years, right? And this is where I wish for more bilateral and multilateral to have these engagements, right?

48:11But I think we need to talk a lot about AI here. I think we found this like across many different industries where we found a way and ideally we can work together. That is a positive note, Lenart, for us to bring this conversation to an end. And I'm so disappointed to do it because I'm really enjoying it. So I hope we will get a chance to do this in person or in the near future. Just have one fun quickfire question for you. Don't overthink it. Who makes a more fun dinner companion? An AI governance person or an AI developer? I mean, speaking for me personally, I love the developers. You know, I'm in DC.

48:43We got lots of high level talk with little substance. And you know what I love about engineering? There are truths. Things either work or they don't. And again, this is my bias as an engineer. So right now, maybe I'm just speaking what I need right now, while Gath is spending a lot of weeks here in DC and not being in the Bay Area for a long time. I would just love to hang with an AI developer and ask him a bunch of questions. Well, you know what? The other thing that has worked has been our conversation this afternoon. Thank you so much for making the time. You have a great weekend and thank you to everybody who tuned in.