Chips

“We should zoom out and recognize the biggest problem right now in robotics is actually data. One day it'll be chips, but for now it's data.”

“It's very obvious at this point, wherever there is chip, there's compute. Wherever there's compute, There will be AI if it's not there. So from that point of view, both from a business as well as use case point of view, AI is the future.”

“And of course the chips are 50% of the data center build out cost.”

“Everyone is compute constrained and everyone You know, at every manufacturer is trying to develop, manufacture chips, RAM, the entire supply chain is operating at, you know, complete capacity to fill the demand.”

“Because right now there isn't a path to value capture because right now you've got a lot of people doing basically the same thing with basically the same technology and the same data and the same chips.”

“Accelerated computing is the composition of thousands of technologies. If any of them fail to deliver acceleration, the value is destroyed. It doesn't matter whether the chip is great if the compiler sucks.”

“You know, they have this crazy belief that, oh, you know, our own internal chips are good enough. They're not.”

“In many cases, we believe that the reliability of the chips, the failure rates will actually be lower on our platform because we can provide much colder cooling temperatures than is typical on land, and we have no oxygen. We eliminate the oxygen from the paylo…”

“Today in that, in Jensen's five layer cake of AI. The profits. They're accruing to energy. They're accruing to data centers. They're accruing to chips. They're accruing to models. Not really accruing to the applications.”

“You have commitments to spend money, but the actual ability to benefit from the spend of money, and even the ability to deploy it, those are Severely limited by actual supplies of chips and workers to build the buildings themselves. And of course, even the tel…”

“Certainly the energy, if we solve fusion or, you know, superconductors or, you know, optimal batteries or some set of those things, which I think we will do with material science, energy costs will be essentially zero, but there'll still be the physical creati…”

“Maybe the chips are the refined uranium more than the actual weights, and the weights are merely one piece of the puzzle.”

“I don't think these companies are all that profitable, partly because, you know, the underlying chips are not all that cheap and there isn't a lot of competition.”

“What do you think that is in AI? Oh, it's data storage, data centers, it's chips. It's not actually the LLMs. It's figuring out where do we store all this data? How do we move it faster? Those are the choke points.”

“And so I think you're gonna see The chips be moved around inside of all the labs to like compute will find the most optimal output. Like the tokens of the most value will always be the ones that it flow that the compute flows to and endless random generations …”

“The sheer level of, I mean, one trillion dollars of demand for their chips over the next year.”

“The one other thing about data centers, which is a little bit more nerve-wracking, is it's not like putting, you know, people compare it to putting fiber in the ground or building the railroads. You put the tracks down. Because once you build a data center, yo…”

“There's more data centers being built. The data centers that are being built are bigger than the ones that used to get built. And then the third thing is where historically the investor would fund the rack and the shell. Increasingly now that investor is fundi…”

“Well, you know, fast forward to twenty-twenty-six, and what we see is, you know, there is obviously more availability of chips, but to build and operate these, you know, data centers requires people, power, infrastructure, a lot of these things that have a lot…”

“I think we're looking at a massive short, shortness in chips in 20, 29 or so.”

“If you think about it, if you're up here in the software stack, it's more important to own one level down, which is that model layer than to kind of start to optimize around chips.”

“There's plenty of space. There appear to be plenty of chips but there isn't enough power.”

“The inventiveness of new chips and the most advanced chips the United States is a bit ahead, is, is ahead. And then but the application to use the chips, whatever AI there is, to be able to integrate it into life and making it for the society Most applications…”

“Yeah, we have plans underway. We think we can do some great work there, but by the time we realize that it will take another eight years, I think from now before this will even be fruitful at all.”

“And it also, it takes at least three generations to make a good chip.”

“These details, like, it does actually, not even for the feeling, it has a direct impact on the usage. Like, people will probably click these things, and there's just a higher chance that they'll, they'll use it more properly if you have this.”

“Jensen is not shown. I mean, he he's lobbied hard to sell chips into China. So if this was really an issue and you really cared about it, you would not be doing that.”

“Yeah, and even if the bottleneck within the infrastructure moves, so if it becomes power, if it becomes cooling or anything else, then you'll have a chip glut for sure, yeah.”

“The reality is that the energy, total embedded energy cost in compute is small compared to the capex of the, of everything else, but mostly the chips.”

“NVIDIA still keeps selling chips.”

“If you're liquid cooled, the thermal doesn't just displace itself, so you have to continually run liquid through that so you don't damage the chips”

“So you multiply those things together. The algorithms, the chips, and then the raw compute that's available. You're talking about a million X increase, some of which will be captured in price reductions. Some of it will be in the performance ceiling, and then …”

“We need to measure and protect four critical areas. Everything else, I think we can deal with some inefficiency, some single points of failure, but the following four areas, Ezra, I would say are Sacrosanct. Number one is all of the technology, both the chips …”

“So my bet is, I think there will be, you know, chips on the market that do that at much lower price. And there's two companies I see going in that direction. One is called Etched. And the other one is called Visor.”

“Turns out, the number of chips was also doubling every 18 to 24 months. So rather than two X, it was four X.”

“Like, I think the Chinese are a bit behind in the chips and But they're ahead in the applications.”

“The data center is now the unit of computing. To me, when I think about a computer, I'm not thinking about that chip. I'm thinking about this thing. That's my mental model, and all the software, and all the orchestration, all the machinery that's inside.”

“Yeah, I mean, the, your requirement in any supply chain pipeline is to eat the value. If you're, you know, like, that's the whole idea of, you know, how those things work. So, like, if the, if you're spending a lot of money on chips, you go into chip making. J…”

“There's an enormous chip constraint right now. Every AI that anybody uses today is its capabilities are basically being gated by the availability of chips, but like that, that will resolve over time.”

“We'll find a way to make more chips. There's a lot of capital in that space. I, I'm more concerned about can the electrons keep up with the growth.”

“Yeah, I think, so I'm not really thinking about technology, because that's a bit too complicated for me. I mean, that's, because that changes so fast, and God knows who's going to be the winner in, in AI, or chips, or these kind of things.”

“Expectation at the moment is that the cost for training these models, you know, may actually sort of top out or even go down a little bit, you know, as the chips get faster, but we don't discover new training material as quickly.”

“You know, the interesting thing about AI is, in the end, it's going to be limited by things like power. You know, if you're, and, you know, chips, and if you're trying to build these really massive clusters in 10 years or whatever.”

“We're probably gonna win chips, and we're gonna probably win material science and batteries, and we're obviously gonna win biotech.”

“I mean, it's a pendulum, and we've swung very, very far to general purpose compute, and now we're swinging back towards more specific purpose type chips.”

“Because if you look at sort of the autonomous cars that are out there today, like I live in San Francisco and there's autonomous cruise cars driving around. They have 10, 20,000 dollars of chips in them, right?”

“You know, there's a longer conversation to have around chips, which I think have become strategically interesting to the rest of tech for the first time in a very long time like beyond, beyond just Apple.”

“You know, even, even if you're an EA type of person or, you know, you're a pacifist or whatever, or if you're into climate change or like, you know, pick, pick any industry that you think is important. Like good luck running compute on GPT-III if there are no …”

“And so they've always been very, they've always been actually been very asset light in almost all of this stuff. It's only, it's A, retail, and B, chips. Those are the places that they push deep into assets. Because that was where you could deliver something f…”

“It's all based on these chips and 70% of them are fabricated in Taiwan.”

“The end state is going to be the chips are going to be free, which means chips will be embedded in everything. You'll be able to use chips for literally, literally everything.”

“I think we're still in the exploratory stage because the algorithms are not matured enough yet.”