Dwarkesh Patel: commitment

Transcript context

…The constraints, as I was mentioning earlier, are not necessarily EUV tools today or next year. They become that as we get to the latter part of the decade. Currently, the constraints are more that they physically just haven’t built fabs. Over the last three to four years, these vendors have not built new fabs because memory prices were really low. Their margins were low, and in fact, they were losing money in 2023 on memory. So they decided they weren’t building new fabs. The market slowly recovered over time but never really got amazing until last year. In 2024, we were banging on the drums that reasoning means long context, which means a large KV cache, which means you need a lot of memory demand. We’ve been talking about that for a year and a half, two years. People who understand AI went really long on memory then. So you’ve seen that dynamic, but now it has finally played out in pricing. It took so long for what was obvious: long context means the KV cache gets bigger, you need more memory. Half the cost of accelerators is memory. Of course they’re going to start going crazy on it. It took a year for that to actually reflect in memory prices. Once memory prices reflected that, it took another three to six months for the memory vendors to start building fabs. Those fabs take two years to build. So we won’t have really meaningful fabs to even put these tools in until late 2027 or 2028. Instead, you’ve seen some really crazy stuff to get capacity. Micron bought a fab from a company in Taiwan that makes lagging-edge chips. Hynix and Samsung are doing some pretty crazy things to try and expand capacity at their existing fabs, which also have large knock-on effects in the economy. So why can’t we build more capacity? There’s nowhere to put the tools. It’s not just EUV; there are other tools involved in DRAM and logic. In logic, for N3, about 28% of the cost of the final wafer is EUV. When you look at DRAM, it’s in the teens. It’s going up, but it’s a much smaller percentage of the cost. These other tools are also bottlenecks, although their supply chains are not as complex as ASML’s. You see Applied Materials, Lam Research, and all these other companies expanding capacity a lot as well. But you don’t have anywhere to put the tool, because the most complex buildings people make are fabs, and fabs take two years to build. I interviewed Elon recently, and his whole plan is that they’re going to build this TeraFab and they’re going to build the clean rooms. I won’t even ask you about the dirty rooms thing, but let’s say they build the clean rooms. I have a couple of questions. One, do you think this is the kind of thing that Elon Co. could build much faster than people conventionally build it? This is not about building the end tools. This is just about building the facility itself. How complicated is it to just build the clean room extremely fast? Is this something that Elon, with his “move fast” approach, could do much faster if that’s what we’re bottlenecked on this year or next year? Two, does that even matter if, in two years, your view is that we’re not bottlenecked on clean room space, but on the tooling? As with any complex supply chain, it takes time, and constraints shift over time. Even if something is no longer a constraint, that doesn’t mean that market no longer has margin. For example, energy will not be a big bottleneck a couple of years from now, but that doesn’t mean energy isn’t growing super fast and there’s no margin there. It’s just not the key bottleneck. In the space of fabs, clean rooms are the biggest bottleneck this year and next year. As we get to 2028, 2029, 2030, there will still be constraints there. The thing about Elon is he has a tremendous capability to garner physical resources and really smart people to build things. The way he recruits amazing people is by trying to build the craziest stuff. In the case of AI, that hasn’t really worked because everyone’s trying to build AGI. Everyone is very ambitious. But in the case of going to Mars, making rockets that land themselves, fully autonomous electric cars, or humanoid robots, these are methods of recruiting the people who think that’s the most important problem in the world to work on that problem, because he’s the only one trying really hard. In the case of semiconductors, he stated he wants to make a fab that’s a million wafers per month. No one has a fab that big. It’s possible that he’s able to recruit a lot of really awesome people and get them on this crazy task of building a million wafers a month. Step one is to build the clean room, and that I think he probably can do. His mindset around deleting things, that it can be dirty, it’s fine, is probably not right. Actually I think it’s 100% not right. You need the fab to be very clean. All of the air in the fab gets replaced every three seconds, it’s that fast. There have to be so few particles. But I think he can build the clean room. It’ll take a year or two. Initially, it won’t be super fast, but over time, he’ll get faster at it. The really complex part is actually developing a process technology and building wafers. I don’t think he can develop that quickly. That has a lot of built-up knowledge. The most complicated integration of very expensive tools and supply chains is done by TSMC, Intel, or Samsung. These two other companies aren’t even that great at it, and they’re tremendously complex.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.