Dylan Patel: NVIDIA's New Moat & Why China is "Semiconductor Pilled”

This is the biggest change in human history, maybe ever. What's about to happen with AI? This is the biggest revolution, bigger than industrial revolution. Jensen is very paranoid about losing. If he just kept making his mainline chip, people crush him on cost and performance, acquiring Grocca's how you get those resources to make more solutions for different parts of the market to stay king. At the end of the day, this is an economic war. If the US and the West win in AI, China will not rise to be the global hegemoni. But without AI, China definitely will rise. They're just gonna outrun America. Hi, I'm Matt Turk. Welcome back to the Mad Podcast. Today, I'm joined by the one person Wall Street and Silicon Valley turn to when they need to cut through the hardware hype, Dylan Patel of Semi Analysis. We go into many of the most important topics of today. In videos, massive move to acquire Groc, the truth about the Capix bubble, whether the US power grid can actually handle the AI boom and the geopolitical chess match playing out between the US and China. But I have to warn you, this conversation went off the rails in the best possible way and we ended up going into all sorts of fun tangents like the strange phenomenon of Chinese romance dramas set inside semiconductor factories. And what's really like when three AI famous roommates live together in SF? Please enjoy this fantastic conversation with Dylan. Hey Dylan, welcome. - Hello, how are you? - I'm great. I love to start with Groc and Nvidia since it's still fresh. So not so long ago, Nvidia was saying that one GPU could do it all and now they're doing this acquisition slash non-exclusive deal with Groc, what does that mean from your perspective? - It's very clear we're not sure where AI models are headed in terms of over the next few years. What happens to the architecture? But the thing that I think everyone is sort of like agreed on is models are pretty autoregressive, right? Next token generation is like the thing, but beyond that, right? Attention mechanisms change how it works. Everything changes, right? Could change. And so what's interesting is the reason in video one is because they just took like the widest surface area bet and then people kept developing models on that and that kind of shape worked. But now the workload is so large that there is room for specialization that will give you 10x increases in certain domains, right? In a general purpose workload, "Crock, rock, rock, rock, rock." Didn't work, right? You know, I can't train. I can't inference really, really large models. Cost efficiently, right? You can't serve many, many, many users, but what it can do is it can go block screamingly fast, right? Same with the Seruber's OpenAID deal. But that's like one workload, right? Very decode focused, right? Doing autorettr progressive tokens in a single stream super fast. Another direction AI models could head, right? We don't know. Are models gonna think in one token stream? Or is it actually there constantly context switching, right? And they're going from, they have this humongous, humongous context and they're generating in multiple parallel streams, right? And so Google and OpenAID have both released mechanisms of this with their pro models where the model actually doesn't just have one single chain of thought for reasoning, it has multiple, right? And then I don't exactly, like, you know, and how they choose which one and what the final answer to you delivers is an area of research. But there is room for that kind of chip, right? Something that works on very parallel. A lot of streams of chain of thought and maybe the latency requirements are not as crazy, right? Maybe you don't want to go blindingly fast, right? Maybe you're okay with it being, you know, because I can spin up 100 parallel, you know, streams of thought or agents or whatever you want to call them. Maybe I care a lot about cost there and because it's 100 in parallel instead of one going super, super fast, it's not as deep, right? The tree searcher, the depth of the inference is not as deep, but it is much wider. You know, there's other parts of inference, hey, creating the KV cache. So Nvidia has a chip for that, right? That's the CPX. So they've made the CPX, they bought GROC for decode and then they still have their general purpose GPU. So they've kind of trying to cover their bases because unlike the first wave of AI chip companies where they sort of just made chips and then tried to figure out where it would work, right? They had a thesis GROC and Srebris both, as well as Sanma Nova, right, which was put a lot of memory on the chip and not necessarily in the case of Srebris and GROC, no memory off chip and in the case of Sanma Nova less memory off chip or slower memory off chip with higher capacity, you know, they sort of all made similar bets in that direction and it didn't work for a while until it kind of did, right? Because there's a workload that now necessitates it. Nvidia recognizes they're the leader, they're at the tent pole, hey, in one respect, they can just run faster than everyone, but it's kind of hard to be two X better than Google or OpenAI or whoever else's internal chip, right? To justify their, you know, 75% plus margins, right? And then they have to be two X to four X better to justify four X better to justify their margins because that's what they're charging above COGS. You know, the question is, what architecture will deliver that? Well, yes, keep the programmability of their GPUs is great for training and for a lot of workloads, but you know, guess what? I think a lot of people will just be downloading an open source model, downloading an inference framework and pressing go, right? And a little bit more complicated than that, but that's gonna be the consumption method for a lot of enterprises, a lot of startups, a lot of tech companies, is there just gonna do that? Or they're gonna rent the GPUs or rent the chips and then download an open source framework and model and go, right? And Nvidia recognizes this and, hey, there is room for products that aren't general purpose, right? The general purpose GPU will still probably be the main line for training and for a lot of inference and for cost efficient inference, but maybe blindingly fast or workloads that have a ton of pre-fill, I creating the KV cache, maybe that those workloads could be different chips, right? And the CPX chip they announced, right? They say it's for the context processing, creating a KV cache. It's also really useful for video models because video models don't care about memory bandwidth and so, you know, why pay for the expensive memory that the general purpose chip has or why do what GROC is doing, which is tying hundreds or thousands of chips together and not having memory, but keeping the entire model on chip. The trade-off for that, of course, is you need thousands of chips and you have less compute per chip. And so, like, Nvidia's trying to capture the whole surface area 'cause, again, you don't know where models are headed. And it's hard to say where the research has had it. - And do you think it's a good thing for the market to yet another one of those deals that's structured as a license, but where you send a position? - I certainly think it's not good from an anti-competitive sense, right? I don't think people should just be able to buy companies without any antitrust process at all. Now, in the case of a large company buying a startup, I'm completely fine with it. The flip side is like, hey, we know the deal is happening, right? This happened for a company I was an advisor for in video acquired in Fabrica just maybe a few months before they did GROC and similar style of deal, right? If someone wanted to strike it down, that's the biggest limbo, right? We've seen this happen in venture and you probably know more stories of this, but a company trying to get acquired, they get stuck in limbo for a year. And then it falls apart. - It falls apart. - Yeah, it falls apart. The deal did because some regulatory BS and now the company was and the founders were focused on getting the deal done instead of like making the product better for a year. Now they're like behind or they weren't focused on growth as much, right? You only have so much time as a founder. In that sense, I like the license deals, right? - So now is Nvidia also dominating the the inference market? Is there any world where Nvidia is no longer the king or it seems to be getting stronger? - I think the thing about Nvidia is they take the Andy Grove mentality more serious than anyone else, right? Okay, find Google implemented OKRs because Intel did it, but that's management stuff, right? Only the paranoid survive, right? This is like core to the Bay Area, core to Nvidia. Jensen is very paranoid about losing, right? These specializations, if you just kept making his mainline chip, would mean people could point solutions for specific parts of the market would crush him on cost and performance than he can't justify his margin. That's a threat to Nvidia's business model as a whole, especially if the best model only changes every three months or the model you want to roll out. Okay, well then you can have three months to figure out how to make a model work on one chip architecture for that point solution. And it's fine software, software advantage of Nvidia is not that important then. Jensen's super paranoid about losing. And frankly, it's really hard to hire enough talented chip people. When you look across the market, there is only a few companies who have successfully created a chip architecture software to run the models accurately. Run the models accurately, right? Because you can look at random APIs of say, an Alibaba, a Kwen model, and different people are doing all sorts of tricks like quantizing it, but also many other tricks, which then end up like making the model quality lower, building a Rack-Scale solution, networking thousands of chips together, and then deploying an API, and grok did the whole thing. With frankly, not that many people, so now it's like, okay, well, I'm in video. I want to make four different chip architectures and actually four different point solutions, maybe the general purpose and then one here, one here, one here. And in addition, my general purpose thing is actually not just like a GPU chip. It's like GPU chips, CPU chips, networking chips, NV switch, nix. There's many, many chips, and each of those chips has many chiplets. You don't have enough engineering resources, right? And so acquiring grok is like how you get those resources to make more solutions for different parts of the market. And as far as like, are they threatened? Like I think, I think like obviously, there's some cool startups out there that are raising a lot, currently or have raised, such as Edge, Demat X, positron, these new age of AI companies. There's also the prior age of like, cerebral is out there still, right? Tens torn, et cetera. And there's a lot of AI chip companies on the startup side. But then there's also, you know, Google's TPU, AMD, GPUs, Amazon Trainium, who are all really credible competitors. And then, you know, Metas MTIA is somewhat credible. And then, you know, Microsoft Semiah is not credible, but like, you know, maybe it will be one day, right? So you sort of have like a lot of competition. They've got to hold the gates back. And so I think, is there a risk to them being, I mean, there's risk from all of those companies that I mentioned and effectively California slash Seattle, right? Only two places. There's also chips from other parts of the world, right? Obviously China has a number of different AI chip companies that are doing cool things. Anyone would have told you GROC was, you know, their business revenue, their revenue was not like stellar, right? In fact, they missed a revenue last year, significantly, and yet they got bought, right? Because the value that IP was there and the value of the team. Anyone else would have been like, well, why the heck would I buy this, right? Makes no sense. There's definitely a credible threat. Yeah. And do you think CUDA is going to remain that mode, I guess, a combination of CUDA and whatever came out of the Melanox acquisition, like do those persist as long-lasting advantages? I think they do. I think networking is super important. I think CUDA software mode is very important, but it's also like changing rapidly, right? It's an incredible amount of the software that Nvidia GPUs run on is not from Nvidia. It's the developer ecosystem that's open sourcing it. When you look at, for example, the LLM and SGLANK, right? These support AMD GPUs, almost as first class citizens now, and VLLM is getting significant support for TPUs for training. And there will be other chips coming out from startups that also support VLLM/SGLANK. Now, how difficult is it? The reason why CUDA is so important is like, OK, I can do whatever I need to do, right? Programming a GPU. I think most AI chips will not be consumed by people programming anything for it. They will download an open source inference engine. They will download an open source model. And then they will put it on the-- and it's really simple to download VLLM and make it work. It's not that hard to set up a server. And Nvidia is putting out a lot of open source software, like Triton inference server and Dynamo and all these things to make it easy, because that is the consumption model ultimately for the majority of AI, right? Is-- and it might be like, oh, it's my own inference engine. But most servers will not run code besides the inference engine in the model. It's not like people are actually like researchers are writing code for GPUs to see ideas, if they'll work and train models, and all these things, or just mess around with them to figure out in for performance or whatever it is. But most of it won't be there. And so CUDA has a mode. CUDA language is like, it's fine, right? No one actually writes CUDA, right? Most people write PyTorch, and then Torch Compile, and then they just run it on the GPU. They don't write CUDA. But a lot of this CUDA mode is like how does PyTorch translate into high performance GPUs. And that surface area from when people were just writing like hard, when people are hardcore writing CUDA kernels to like, hey, they're writing PyTorch, and then it's compiling down to GPUs versus, oh, I'm just downloading VLM. It is a curve of like not a ton of people that can do CUDA kernels. A whole lot more people can do PyTorch, right? Random, you know, PhDs and random people. It's very simple, right? Hey, crap load of people can do VLM, download it, run it on a server. Well, if it now supports other chips, what is the CUDA mode? And Nvidia's recognized this and they've been building software that is not necessarily the CUDA mode. And I can give some examples, right? So the name of the game is Fast Tokens and Lois Cost Tokens, right? And Lois Cost Tokens happens by your chip being fast, but there's also tricks, right? One example, right? Like I mentioned with, you know, the CPX versus GROC, right? As processing your pre-fill contacts, right? Super cheap CPX, right? If I care a lot about speed, then GROC, these are optimizations on the hardware side. There's optimizations on the software side as well, right? And so one example is when I'm doing, for example, if I look at a cloud code or a cursor type application, right? The workload is like, it takes your repo, it takes the relevant parts of your repo, puts it in the context of the LLM, it prompts, it generates, right? And if it's an agent mode, it circulates the context a couple of times, it'll collapse, put things off to the side, access to different contexts. But what's interesting, you know, especially when you think about an agent for software, and you can see this in codex, you know, codex, codex actually not as good as cloud code, but can do it work on time horizons of like, 9, 10 hours, and do like a big refactor better than cloud code can, even though most of times, cloud code is better. And what's interesting about codex does is it'll like, take your repo, it'll identify parts if you're asking it to refactor it, identify parts, write stuff, you know, make like these notes for itself everywhere, collapse the context, switch from this part of the repo to that part of the repo to this part of the repo. But when you think about it, it's like, oh, if this thing is just generating tokens all the time, plus it's switching what my context is constantly, that's really expensive, right? If you think about like, what's the cost of inference, I wanna say it's like, it's $10 per million tokens of output and $3 for decode or 10 for decode and 3 for pre-fill. And so if you think about, oh, it just worked for 9 hours on one task, one refactor, huge value, but if it changed context a ton of times, and your context is like 30k usually, or 50k, or you know, heading to hundreds of thousands, you know, how long your, how big your repository is, and how much context, which now you're spending all this money on pre-fill, right? Not the decode tokens, but actually, why am I like regenerating the KV cache? I can actually just like store the KV cache elsewhere, and then when I needed again, I could pull it and plop it into CPU memory, or into the GPU memory. And so Nvidia's got this like KV cache manager, and they've been working really hard on like, making it so they can interface SSDs and stick the KV cache on there and pull it out whenever they want, so for this kind of workload. And then if you do this and you look at like, coding as an application, and you like, look at these coding companies and how much they're paying for pre-fill versus decode, actually majority of their cost is pre-fill tokens, not decode tokens, because the context is just so large, and it's switching all the time even in agent modes. You know, if you can now not have to do the pre-fill, your costs go down dramatically, but that's a very complicated thing to do from a software perspective. You know, companies like Anthropic, Google, OpenA, have already done it, but what about the wide world, right? And so Nvidia's trying to make the open source software for this, and that's like, kuda mode, but it's like actually, no, none of this is kuda, right? Like it's like memory management and like, you know, storage management and when do you call what and how do you transfer it and how do you like spread the KV cache across a bunch of different storage nodes and what happens when you read it and the network congestion, just like all these things, yeah, it's like, Nvidia's wheelhouse, but it's not kuda. And I think like the easy way to say it is, it is the kuda mode, right? And so things like this, KV cache manager, and many other things they're trying to do to reduce the cost of inference, like is how they build the new code of kuda mode, because again, today it's, you know, it is quite, I mean, AMD's like not fully there yet, and TPUs being added right now and trainings being added soon as well to VLM, but all of them will have a very good UX for download model, run model on VLM by the middle of the year, I think, right? Certainly AMD is already there by the end of this quarter. We have something that like tests this, right? It's called inferencemax.a, it's open source, all the code is and the results are, but we run across, I think, $60 million of GPUs, which are donated to us by companies like Nvidia, AMD, OpenAI, Microsoft, Amazon, Crusoe, Corvieve, together AI, all these companies are sponsoring GPUs for us to run this, but running VLM and SGLINK every night on, you know, nine different kinds of GPUs, on a variety of different models and different contexts, lengths and all these things, right? To see the performance and you can see the performance moving every day, or pretty often because the software changes all the time. And so like, the fact that this exists is the kuda mode, right? It's not that like AMD, you can do this on their chips, and you can do this on their chips, it's, oh, when the new model comes out, how fast does it get to peak performance? 'Cause it's a moving target. Or, hey, can I implement this KV cache management thing? How hard is it? How many engineers do I need? Oh, just one? Great. Like, or 10 great. If I need 100 people to develop it, like Google and, you know, so on and so forth, did, then that's much harder. - Do you think AMD can catch up? - I think AMD will be caught up at times and very behind it other times. Like, currently, they're super far behind, right? 'Cause Blackwell is just way better than MI355. And then, you know, Rubin comes out, and they'll be way, way behind, but then AMD's new chip comes out, and AMD will be caught up, or even slightly ahead, on a hardware perspective, software's behind, right? And you have this like leapfrogging, and AMD is a very credible second competitor. I don't think they'll go beyond like, I think they'll stay in single digits market share, single digit percentage market share. Single digit percentage market share is still pretty good. Yeah, I mean, Nvidia's revenue this year is gonna be like, it's a lot. - It's three-gajillion. - Yeah. - $3. - I think it's actually four-gajillion. (laughs) - What about all the startups you mentioned of FUSE, so there's zero price on the one end of the spectrum, and then newer ones etched and others. If AMD has a, you know, uphill battle in front of them, like do you think those guys can take significant market share? - You're sort of the whole specialization game, right? You have to specialize because you're never gonna be Nvidia at their own game, right? They're gonna have the supply chain unlock, they're gonna get to the newest memory technology, or process technology, or what a packaging technology, whatever it is, sooner than you, and they're just gonna crush you, right? If you play their game, you have to, AMD is trying to play Nvidia's game, but AMD is like, extremely good at engineering silicon, right? Everyone else has to, has to, has to, try something weird or different, right? And so when you look at etched or Maddox or Positron or Suribris or Tenztorner, and you go to look at all these companies, right? There are unique things about what they're doing, and it's not clear if AI models still be within that realm when that comes out, right? Does, oh, now people use like Ngrams and other sparse attention techniques, is that, like, is this that change? Like some of the specializations people are doing, or hey, people are now doing like, you know, models are now sparse MLEs instead of being dense models. Does that change things? There's so many optimizations and changes on the model side, and you can't predict what's gonna happen with the ML research easily, at least, you can't. The thing you're optimizing for today has to be a vision of where AI will be in two years. And Nvidia's fully accepted, they don't know where that's gonna be. That's why they have a portfolio of chips now not just one GPU line, right? It's not just Hopper, Blackwell, Rubin. Now it's gonna be, you know, it's not Ampere Hopper, you know, it's not that line, it's like there's a variety of chips to serve the different markets in different possible scenarios. They think each of them has this vision today, but oh, it might turn out the general purpose one sucks, and actually AI models have developed in a way where CPX or Grockstyle chips are the best, right? Well, okay, now we have a solution for that market. And so I think that's the challenge with the startups. With that said, I think they're all taking very interesting bets. I think it's much more exciting than the first wave of AI hardware bets, GraphCourse, Rebres, Samanova, Grock, where they all made the same bet on memory and putting the memory on the chip. They sort of just made a bet, and they optimized for a certain kind of model, all similar kinds of model, and it didn't end up working out for a long time, right? They had to pivot and they had to work on a lot of things that took a long time. I think these companies have a really clear vision of what they think models will look like, right? Edge does, Maddox does, Positron does, and that's what's really cool about it between the three of them, these new Edge. So I mean, I'm excited for them. I'm very, very skeptical. I don't know what adventure capital is views as likely chances of succeeding, but I think all of them are less than 1%. Right? But that's a-- But the world where they win is a multi-silicon kind of world where any given customer uses a range of different GPUs. It could, it could, or it could be any given customer has like one workload they care a lot about. Anthropic clearly does not give a crap about video gen image gen, right? They just don't care. On the flip side company like Mid-Journey cares a lot about image and video gen, right? Image and video gen is very, very-- like I mentioned, it's a very memory bandwidth heavy. It loves, loves, loves compute, right? Whereas inference of large language models in the style of like, you know, these, you know, say for example, coding agents, cares a lot about decoding for long streams of time. And that's very memory bandwidth heavy, right? And so there's like, that's like a simple example, but there's a lot more nuance there in terms of like, even like the size of like the matrix multiply, you know, the tensor cores that you, you know, the systolic arrays that you use, or the ratios of networking and memory and like what's the memory hierarchy look like? And you know, what are you doing for different kinds of attention? And like, oh, like all these sorts of things, like, there's a lot of specialization here. And so some people are betting big on, on different types of specialization. And I think like, you could clearly see a world where companies do care about different stuff, right? Like, like, if for example, a chip optimized for video and image generation existed today and it was better than in video, or in video made it, I think mid-journey would absolutely only use that for inference. I think for training, they'd still use the general purpose thing. And as with like meta and Google, would like, they should do that, right? And hey, meta actually has two lines of AI chips. Their MTIA, there's a line that's focused on recommendation systems. And then there's a line that's focused on Genai. The Genai one is a new line, but that recommendation system's chip line is still continuing, right? It's not sexy, no one cares, because there's, and bite dance also has a recommendation system line of chips. And it's not really focused on Genai, which is fine because, you know, this is a $200 billion business or something, which is just deciding what ad deserve me, right? And what order to put my friends stories and things like this. So I think like, it's perfectly fine for there to be specialized AI chips, given the target market is big enough. And you have to have vision to know what that target market is. Unless you're hyperscaler, then you can like just like, you can just use general purpose until you've like, it's clearly there, and then you can make your ASIC, right? Fascinating. Turning to the geopolitical aspect of all of this, which is always fun, Huawei and Nvidia in China. Last year, there was like 10 or 12% of their role revenue. And this year, they were saying that their market share has basically dropped to not very much. Is that Huawei chips, is that restrictions, that tariffs, what's happening over there? - I think there's a variety of things. Actually, in some quarters last year, it was even north of 20, I think. But I don't remember exactly. But anyways, you know, if you look at 2022, China was almost the size of the US in terms of buying server hardware, right? Almost, not quite, but getting there. And it looked like they were gonna be the same size as America in like a year or two after that, right? And if you look at like global data center capacity, global cloud capacity, et cetera, et cetera, et cetera, it's American companies and Chinese companies, right? That dominate the world. American companies are obviously doing a lot better here, but both of those dominate the world. And if you look at like every industry, right? You know, it's very clear that like, China wants to insource stuff, right? So in 2015, they made these five-year plans for 2020 and 2025, where they set the percentage of semiconductors they wanted domestically produced. And they've missed the goal both times, which is fine, right? They set really aggressive goals and you know, shoot for the moon even if you miss, hit the stars, right? And that's sort of what's happened, right? Like, look, China is not caught up on, you know, leading-edge semiconductors, but microcontrollers from China are almost as good as the microcontrollers are as good and cheaper than the ones from Texas Instruments or ST Micro or, you know, et cetera, right? Or like this power, random power chip is better than, or the same as the one from like another company, right? And so they've really built up a semiconductor industry and started insourcing a lot more. I don't see why China wouldn't be buying, you know, 30, 40% of the world's AI chips and the US like 50, 60% and then the rest of the world like, you know, and when I say US, I mean US origin companies, that seems like a more natural state for the world. But there are restrictions and, hey, this is the biggest change in human history, maybe ever, knowledge work and everything that's going to happen there and then eventually like robotics and all these things. Like, you know, obviously there's a lot of geopolitical stuff and so there are restrictions. Nvidia's been handicapped, handicapped from selling their best chips to China. And so that's obviously impacted the sales a lot because like why would you do that? And so when you look at who rents the most GPUs in the world, it's three companies, right? So one of them is obviously OpenAI. Second one, actually they were bigger than OpenAI. They are bigger than OpenAI today. Or no, they were bigger than OpenAI than OpenAI clips them recently is ByteDance. ByteDance runs rents tons of chips from Oracle and Google and, you know, many other cloud companies because they couldn't get the chips they needed in China. They're mostly just serving TikTok, right? Okay, well, they're not allowed to buy them and that sucks but, you know, they're allowed to rent them. And so, okay, if I'm not allowed to get the best ones, I'm going to rent externally. And if ByteDance is the second biggest venture of GPUs in the world that's substituting demand that would have been built in China in many cases, it's instead being built in Malaysia. And Oracle has over a giga lot of capacity in Malaysia that ByteDance is going to take, right? So things like this are, you know, hundreds of thousands of millions of chips, tens of billions of dollars of capacity that would go to China but it's not. It's going to Malaysian stead, right? As an example, another sort of point around this is China's, like, you know, they've had these five-year plans. So, and, you know, the way these initiatives work from China's, there is like some top-down ordering but then they just kind of whip the whole, like everyone just kind of gets into it and it's really cool. Like, I don't think it's as top-down as many people think. Like, I think the entire country is like semi-conductor-pilled, right? There are dramas where people fall in love in the fab or dramas where people fall in love and they're photovoltaic, like solar cell researchers and engineers and it's like, this is just the backdrop and it's like actually, it's like super cool for your like significant other to be that semi-conductor engineer or to be that photovoltaic, you know, solar panel researcher. - As opposed to an influencer. - As opposed to an influencer, right? Like, I'm sorry, love Island is, I watched like for 10 minutes because I was forced to, I was like, this is freaking terrible. But, you know, like, we are so cooked. - No, you know, seriously, we're cooked. And so I think, I think like when you think about like, this happens, it's like, it's diffused into drama even. People, like, like, there's multiple dramas, like taking place about semi-conductor industry and they're like romance, comedy, like, the entire spectrum, right? Drama, really, it's like, it's like, what the heck is going on? Anyways, you have all these provinces, you have all these local cities, setting out ordinances and giving out subsidies and all sorts of stuff, right? It's truly like crazy. Like, there's some national level stuff, like, oh, no taxes on this. Oh, we're gonna ban a few things. But as far as I understand, the national government has not banned in videos H20 or H200. But the local ones have, right? A lot of local ones have said, no, you know, you must use China manufacturer chips and it's like, who told you that, you know, you're here to uphold this, it's like, doesn't matter, right? I mean, like, it's cool because then you have this, like, survival of the fittest, all these provinces and cities are trying to attract different companies with different types of subsidies and grants and industrial parks and like, all these different things. And then like, the ones who succeed actually develop an industry and they take over. - It's really high one thinks of China, right? They almost sense like more like the US or like it was a federal government in states with the provinces of what's already over there purchasing. - I mean, it's actually like, great, there's this one TikTok or not TikTok, and Instagram like person and they're like, they like singing, they're like, if you wanna buy things in China, make sure you go to the right place. And then they just say the most random shit and name the city. And then you look into it and you're like, wow, the city has the entire supply chain for this. And it's like lampshades and then it names the city. It's like, what the fuck? There's a city that specializes in lampshades. Like it's like, it's like, microphone arms. Like microphones, like it's like, literally, there's a city in China that specializes in things. - And guitar is as well, right? This one city that became the guitar capital. - It's literally everything. Literally everything there's a city. And it's not like, hey, specifically for camera arms, for example, there's ball bearings in this. And the ball bearings are, like, there's ball bearings. There's multiple manufacturers of ball bearings for camera arms. - Yeah, good. - And then most of the camera arms in the world come from that one city. It's like, what the hell is going on? And so like, the semiconductor industry, I think people don't realize is absurdly specialized. - I'm not answering your question. I just gotta let it or rat. 'Cause I think people don't understand China semiconductors. It's really sick, or semiconductor is a general, but like, you know, like in Japan, they like focus on a few different types of chemicals and they're the best at it. And it's like almost a cultural thing, right? Like, Japanese people are so precise, like with sushi and like, it's all about the trade and the craft. And like, you know, the French food in Japan is better than the French food in France because the Japanese chefs went there and then come back and they perfected it in Japan and like, 'cause they're so precise. And there's so many different things that like, Japan is so good at because they're so precise and like, dedicated to the craft. And it comes out of like, I don't know, like, samurai culture or something. I don't know, right? Like, I don't exactly know how that culture came up. And so when you look at like, and it's like across the world, there's different places where things like this happen, right? Like, oh, like the Netherlands makes UV tools. Cool, I guess so. And you look across the semiconductor industry, there's a famous economic essay called Eye Pencil or something like that or talking about how the pencil, like a simple pencil comes from like, oh, the rubber comes from like Indonesia for the eraser and the graphite comes from this mine here and the wood comes from these Aspen trees in Canada and like, you actually can't make a pencil without aggregating this entire supply chain. semiconductor industry is like way crazier because like, I would say there's like 15 or 20 countries that can shut down the entire semiconductor industry, right? Even like Austria could, right? And it's like, what? And it's like, well, yeah, there's two different companies there who have like 90% share in like some random niche stuff. And it's like, okay, cool. I guess Austria can, and oh, yeah, those two companies only like have less than a billion of revenue, but they just happen to have Lynch pin critical things. And there's Lynch pin critical things everywhere because the process is so complicated. And so China's been trying to replicate this. Is everyone thinking they're missing that they don't have yet? I think there's a lot of things. I think if you were to close your eyes and say, or if you were to cut off every country and say there's no more globalism, China has the most vertical stack in semiconductors today and they're the best semiconductors in the world because their fabs could still run somewhat on a lot of things because they have built some of these chemical supply chains, right? Like TSMC for certain kinds of chemicals 100% share from Japan, right? Or Intel same thing, right? Or you know, for certain kinds of tools 100% share from Netherlands or 100% share from, you know, this American company or that, you know, Austrian company or this or that, right? Like there's just all these like, you know, this Swiss company, like it's just all these different places have 100% share. It might be one company, it might be three companies but geographically or in the same area and China's built that up, right? Because they've created this made in China initiatives which just plowed money into it and they've got this culture of like the diffused, like, you know, these provinces are like, yeah, I just decided I'm gonna fucking focus on or it might not even be the city, right? It may be the like, you know, someone brought it there and decided and then people are like, oh wow, you're doing that me too. Like I'm a Patel and I grew up in a Motel and guess what, we like almost all the Patel's I know grew up in a Motel and it's because some random Patel immigrated to America and like worked at a Motel and then bought a Motel and then like, it just started happening, right? Like these things are so ridiculous into those sorts and like, I don't know, like it's like, I view it as the same kind of specialization, right? Chinese cities are like starting to do these things. China's missing a lot of things, right? I would say like if you say minus 10 years tech, China's complete and no one else is complete, right? Taiwan is not complete. The fabs are shut down without forward supply. And you go down or you grow across the stack. But if you go to 10 year tech, maybe more like 20 year tech, you could get a fully vertical supply chain in China, which I do not think any country could do. Like America could not build a fully vertical fab without stuff from elsewhere, even if it's 20 year old tech. - Yup, yup. - Probably not even 40 year old tech. And so that's interesting. But then when you look at the flip side, it's like, well, like you kind of do need specialization. That's how that chemical gets the purest, best, most engineered, you know, or that slurry of chemicals or that gas or like that tool, because every smart person or a lot of them in that country grew up around that culture and like the supply chain is there and like everyone kind of knows and like it's like a drive away and like sort of like this is what makes supply chains work, is that there is this specialization. And the best of the best only comes when you have that hyper specialization. So China doesn't have lithography. Their lithography is like 10 years behind. And I think it'll be five years behind in a couple of years, right? They're catching up fast. I don't think they'll be as good as ASMR for a long time. You know, maybe I don't know. Maybe they will be, you know, you know, China, you shouldn't ever underestimate China, but like in Chinese engineers or, you know, but like for a while, right? Or like, you know, I don't think they'll be able to make leading edge chemicals like many Chinese, Japanese companies or many American companies in their tools and like you just go across the supply chain. They're not, hey, forefront on really anything in the manufacturing supply chain. On the design supply chain, there's some things that they're starting to be similar par but like cheaper or like a year or two behind, but cheaper and that's like fine for a lot of stuff. An example of that is Huawei, right? Huawei in mobile phones was on par with Apple, like entirely. And they had become Apple TSMC's biggest customer when they were designing the best thing. And they are number one in telecom and their tech is just literally better. And so when you think what happens, you know, is China missing anything? They don't have the best of much today in the AI supply chain. They have a complete package and a couple of years behind and they'll figure out how to make it cheaper, slash, do more, slash catch up and create a robust industry. But there's a reason like, I don't think that like, Jensen is scared of AMD really. He's paranoid, I mentioned he's paranoid. I'm sure he's a little bit scared of them, right? And I think some of the things that they've done are reactions and competitive dynamics with AMD or Google's TPUs or whatever, right? There's a core we've deal today. And I think that's directly the result of what Google's been doing. - Yeah, the 2 billion pipe that Nvidia announced in two billion. - Yeah, and it is. - 2 billion in Core-Rive. But what's more important is that's like sort of just like the sticker, what's really relevant is Nvidia is gonna work with Core-Rive to acquire and backstop and all these things, the land, the power, the energy, the transmission, the help build the data center, all this capital side stuff that, because Nvidia has so much money, they can backstop Core-Rive doing it because Core-Rive then can be the one who generates the money. Anyways, there's like, 'cause Google was doing that. And they did that with like a couple companies, such as Floyd Stack and Terrible Off and Cypher, these are some public deals that have been announced. And so Google was doing that with TPUs and Nvidia reacted, right? And so in the same way, I think Nvidia is reacted to AMD. And in the same way, I think the thing is, Nvidia is like deathly terrified of Huawei. 'Cause Huawei has caught up to Apple and actually surpassed them as TSMC's biggest customer before they got banned, right? They did just crush Nokia, Sony Ericsson, et cetera, right? Like the entire telecom supply chain. They just like completely destroyed them. And there's so many other areas like, they straight up made a folding phone, right? You know, I have a Samsung folding phone. They have a folding phone that's better than Samsung's folding phone. And it's like, bro, what? Like, you know, Huawei is really, really cracked. And so of course they're terrified of, and Huawei is the most vertical company in the world. No company is more verticalized than Huawei, which then leads to huge innovations. It's something that we don't fully appreciate in the US, but when you travel in Europe, you see everybody who's like honors on our phones and it's like the footprint of Huawei is huge in phones in a way that people-- - But on just phones, you know, security cameras. - Yeah. - Actually, they think they have like, you know, - There's like a lot of training on the, (laughing) that a captive group of testers. - Exactly, exactly. I think Huawei is terrifying, right? And so like, yes, their chips are not as good today. - And is that already happening? I mean, obviously the US and China are the two biggest markets, but like for other markets, I don't know, UAE, Middle East, Europe, our Nvidia and Huawei already had two head in the deal. - Well, I shipped a little bit, but like mostly just like sticker capacity, like there's nothing like, no, no, like, I would say like a little bit as in like a few servers, not like a billion dollars worth of stuff, right? The thing is China's supply chain has to ramp up, right? China's express goal is to have all internalized, but then like a company like, while you're back, I was like, I don't wanna use Huawei, right? Like, I wanna make, I wanna use Nvidia and just make the best freaking models, right? 'Cause that's my business. My business is not, you know, using a Huawei thing, but it's like, okay, it's being pushed upon me. There's other companies through the camera con and so on and so forth. And so there's sort of like supply chain, you know, companies in China don't wanna use, probably they're kind of encouraged, obviously, or pushed, you know, you must. Some local provincial government be like, well, you're doing this much business here. You gotta do this, right? Like there's all sorts of like crazy stuff that, you know, pushing of companies to use Huawei. The challenge is, Huawei can't manufacture enough, right? We've like done a lot of work on this, and we've just put it for free instead of like to our customers because it's like something that's like national security, which is how was Huawei actually building chips? Well, actually, they were using shell companies to get chips from TSMC and using different methods of like sneaking HBM, which is memory from, you know, Korea through Taiwan to China, right? Like all sorts of crazy stuff we've reported on. And people, it's like in black emul, right? They shut it down, or like tools that get shipped to China and they shouldn't be for, you know, making leading edge chips, but they actually are. And all these sorts of things are happening because they can't make everything. And if they wanna make the leading edge stuff, they do need to rely on the foreign supply chain quite a bit in terms of the upstream supply chain, right? Memory, logic chips, tools for fabs, chemicals for fabs, et cetera. Huawei cannot satisfy the market because there's not enough advanced leading edge capacity in memory, logic, you know, and all these other things, domestically in China. And they're trying to build it as fast as they can. But that means there's just not enough to satisfy the market. So Nvidia has a market. I think they'll figure out how to sell chips to China. And Jensen's in China, I think, like right now, or was yesterday. And so like he's clearly like a wheeling and dealing to trying his chips in China because, you know, I think Nvidia's argument is if we sell them chips, then they won't, you know, there won't be enough of as much of a domestic market. The feedback loop for software and everything else won't be there. That was sort of like really challenging, right? Like most of the open source software for AI has a lot of Chinese contributors, right? The OEM and PyTorch, SGLANG, and like all of these other like libraries and things are just like, you know, and it goes to low level software especially, right? Like a lot of the best open source stuff is actually just from like a Chinese company who decided to open source it and same with models, right? And so like it's like, okay, well, if they can't use Nvidia chips anymore, then this open source stuff won't be designed for Nvidia chips. So be designed for Huawei chips. And now does that like weaken the Kudamo and now like not only is China domestic, now they have like a feedback loop internally and then they can externalize across the rest of the world, right? So this is like argument in video makes. I'm not sure if I am like, I'm like, you know, I think my AI timelines are so fast. I'm not that fat, like not in terms of like AGI, but like, hey, AI is $100 billion of revenue across the industry. I think the industry could hit $100 billion error by the end of this year. Like 45, 50 for open AI, like 35, 40 for Anthropic and then, you know, Vertex, DeepMinds models at Google Gemini, right? And then Vertex API for Anthropic models and bedrock APIs and Azure Foundry APIs, like I think $100 billion like ended this year. That's a lot. And then what's the economic value of that $100 billion? Now, how much of that is in China, right? Like China's number is probably 10x lower, right? Because they just haven't been able to pervasively push AI, right? Chad GPD has a billion users roughly. And you know, then you add on Gemini and Metaclames have 500 million users. I don't know. I think people are accidentally click like a generic sticker or something. But like anyways, there's like, there's like a lot of usage of AI in the West already. And it's going to climb, it's going to keep climbing. And like you kind of have to get used to it. And so like the question is like, do you, you know, what's the economic benefit to the world, right? And at the end of the day, this is an economic war, right? If the US and the West win in AI and control, you know, more powerful AI systems that have this feedback loop that improve economic growth and weapons systems and whatever else, right? Engineering of grids and cyber attacks and all these sorts of things. They have this like advantage over China. Then China will not rise to be the global hegemonie. But without AI, China definitely will rise to be the global hegemonie. They're just going to outrun America. And so the question is like, you know, that's I think like the other view, right? And how fast are super powerful AI systems versus you know, China building a domestic ecosystem for chips and models and everything that is a few years behind. Like what's what's what's actually the value, right? Like that's sort of like around restrictions and regulations. - What do the US on-shoring efforts fall in that category? What do you make of them from the chips act to like all the things that is being built? Everything looks like it's massively delayed by the way, which perhaps is not surprising. - I think TSMC's manufacturing waifers and they're like building real waifers and there's real fabs and like, you know, there's some other fabs that have been announced and like they're doing well. And there's like a bunch of like different kinds of plants, like a Korean company making a random gas plant in Texas for you know, they're chips, right? Like for chips and all these like sort of things are happening, I think the chips act did really well with its $50 billion. It's just, I don't think people understand the scale of the semiconductor industry. It is the most complicated supply chain in the world, right? It's much bigger than you know, say manufacturing airplanes. It's much bigger than like, you know, really anything else, right? If you look at the top 10 companies like of the world, I think eight of them design semiconductors, right? Now obviously like Google design semiconductors, but it's like, oh wait, no, but their cost of search would be like 10x higher if they didn't have TPUs. And TPUs are super optimized for search, right? Or like, you know, you go down the list, right? Like MetaServes recommendation systems with their chips, right? Like you go down the list, it's everyone is making their own chips. Apple devices would be materially worse if they didn't have their own chips, right? And you just go down the list, it's like, it's the most complicated supply chain and they're spending something on the order of like $150 billion roughly in subsidies a year to the chip industry. We are doing 50 over like a decade. Yeah. There's a difference in scale here, right? The collective total amount of like capex that has been spent in Taiwan is like 500 billion plus, right? Across the industry, across all the companies that are making semiconductors in Taiwan. And Taiwan doesn't have a domestic industry. How is 50 billion dollars of subsidies gonna change America's needle, right? It does move it a little bit, right? I wouldn't be clear. Like the chips act is awesome. I don't understand why like EVs or like solar was given this massive, massive like trillion dollar package. 700 is only given 50. Like semiconductors need a lot bigger package to actually incentivize the on-shoring. I think what's happened so far is proven that it's working well. TSMC is literally making chips for Nvidia and Apple and AMD and others in Arizona today, right? And I think that's really great. It is your sense that the broad American government is just aware of all of this. I wouldn't say it's roughly in the back of my mind. It's only passed because the automotive prices went up because car manufacturers are like the worst 'cause they do just in time inventory, right? Or worse, but it is just like a thing, right? Just in time inventory systems. COVID happens, sales plummet, fabs that were making, you know, random power ICs or random microcontrollers for engines got repurposed to the boom from COVID, which is, which was data centers and PCs and smartphones. So that stuff was booming. And then when people were like, oh wait, actually like, you know, I have some money, I stayed at home, I didn't go out, I didn't drink, I have a lot of, I have some cash, right? Let me buy a car, they went on buy cars and car started skyrocketing in prices. Oh, let's restart and let's, oh yeah, can you sell me that microcontroller for the engine again? It's like, no, I'm making a slightly different microcontroller that works for, you know, let's say a keyboard or a mouse, right, or whatever. And it's like, and they actually didn't just leave me flat footed and they were like a partner through COVID, right? You know, versus you just left me, screw you for it or whoever, Toyota or automotive OEM, you know, that supply chain. And so chipsack did not get passed, only got passed because that happened. And people were like, oh my God, the semiconductors are why cars can't be made. If that didn't happen, we wouldn't even have the chipsack. It's like, it's like silly. So like, I don't know, like I think, you know, whereas like, and even though that's what was pitched to all the senators, like I know people who were running around Capitol Hill just pushing that narrative and story and that's why it finally got passed. In reality, it was all for advanced leading edge chips, right? Nothing that goes in a car, right? And so it's like this like funny thing. - So you know the words of your thing, my words, my words not yours, but is it hopeless that the US is going to-- - I'm very optimistic. - Okay, I mean, do you think that's world where the US just decides to invest in semiconductor at the scale that's in it? - You know, I thought we just needed a bigger chipsack, but look, Trump's kind of gotten TSMC to promise to invest a fuckload more and they're moving on it, right? They're like actually like just building it. It's like, I'm gonna tear off the shenanigans unless you build a fab, but it's like, you'll build a fab. And they're building it, right? Now the timeline for our fabs just takes forever. 'Cause again, it's the most complicated thing in the world. The cleanest space in the place in the world is not like a hospital or a biotech lab or whatever. It's a semiconductor fab. And the most expensive tools in the world are not, you know, any of these medical tools or whatever it's semiconductor tools or it's not rocket, it's semiconductor tool, right? Like everything, you know, I describe it as, I remember when I was a kid, I was like, I wanna be a rocket scientist. And then I was like, oh, I wanna be a surgeon and I'm like, wait, chips are like rocket surgery. And even cooler, right? Like I think, anyways, like sort of like, there are fabs being built in America. They won't take America to self-sufficiency. I don't think that's a relevant, I don't think that's a goal relevant, like that's relevant, right? Like, globalism is generally just good. Hot take. Like in terms of economics. - I'm not moving on this into a short, a YouTube short. Globalism is good. - Dude, you're gonna get me like Ketzles. (laughing) I tweeted about ice. It was a complete joke, but somebody people got mad at me. 'Cause I kept peace, you know, I'm too much of a joker. These are serious things. - Yeah, yeah. - No, I know the feeling, yes. - Anyways, I think, I think, you know, I think we are building fabs and I think it's like gonna move and now, even Elon's talking about building fabs now. 'Cause he sees the shortages in the world, right? There's a lot of semiconductor related shortages for building out AI. And so, I don't think it's hopeless. I think I'm like very optimistic that we're gonna do more and more and more. And maybe this administration threatens tariffs and they get the deals and the next administration comes back with the carrot if it is the Democrats or whatever happens, I don't know. - Like, is that a comedy club on Sunday night? And like, he's like, I use Chad G.P.T. and then like, there are a couple of people who booed. He's like, yeah, I'm one of those guys, I know. And like, it's like, wow, just people hate AI. - And that has not even started, right? Like, the actual impact of AI. - Or like, New Jersey power prices are up, right? Is it because of a data center? Well, New Jersey, the governor's election, like, I think literally, like, there's like an election that changed recently in New Jersey because power prices were up. And people blamed a Microsoft nebius data center in New Jersey for that reason. But in reality, that data center has nothing to do with power prices going up. It's super storm-standy, like, five years ago, knocking, or whatever, how many years ago, knocking down the state's electrical infrastructure and then improving all these improvements. And then those improvements have to be paid by someone and it turns out the consumer has to pay for them with higher power prices, right? And so like, you know, like, there's like, there's a lot like going on in that regard, right? That kind of is a sad. And people hate AI and they're blaming AI on it. An artist hate AI and like, you see all this deep fake stuff. And like, I think it'll be the hottest button issue, especially as like, we're really getting into like, I think last year, Google spent $3 billion on Waymo and we're waiting for their guide for this year, $3 billion on Waymo taxis. But their waymo's went from like 300k to like 100k or 90k the new Waymo car. And they're gonna spend more than three because they've just launched in like four cities now, right, or five cities, and they're testing it a lot. And the same with robot taxi, like people are gonna hate AI for that reason, people are gonna hate AI because the slop on the internet, people are gonna hate AI because you know, the perceived job replacement, people are gonna hate AI for all these reasons. And so yeah, it's gonna be a hot button. Political issue, don't you think? - Yeah, talking about that. So, CapEx, is there a CapEx bubble? Are we investing too much? Or actually, are we investing not enough given what you were saying earlier about the rate of revenue increase and therefore we apply demand that you expect for this year? - I'm obviously a Maxi, I think we're gonna need a lot of in front. I think I'm literally paid to like analyze the supply chain and do consulting, like that's what my company does. So like obviously I'm very biased. I think we're pretty good at calling when things go down though, right, before like a part of the supply chain reverber. But anyways, you know, again, going back to the economics of it, it's north of $100 billion of revenue exiting this year for AI from a base of, you know, sub one billion, a Gen AI from a base because ads and stuff is like already a multi-hundred billion dollar AI industry, right? You know, go back to 2023, it was like less than a billion, right? And 2024, I don't know exactly what number, maybe, let's call it 10 and 25 was maybe like 3040. It'll be north of 100 easily. If you're like about a hundred billion dollars of revenue, let's say at a 50% gross margin, so that's 50 billion dollars of gross profit and 50 billion dollars of cogs, that 50 billion dollars of cogs needs to run on Infra, which cost roughly, if you're talking about five year depreciation, call it 250 billion dollars, right? Of Infra, four hundred billion dollars of revenue. Okay, what is the actual spend on AI? And for this year, it's going to be like, it's, I mean, it depends on what layer, if you're talking about energy, those are longer lived assets and all these other things, right? Data centers are longer lived assets, the chips are not as much. People are putting catbecks down. And the hyper-scalers catbecks is going to be like 500 billion dollars a year, something like this. And then besides them, there's also a lot more catbecks, elsewhere. And so, you know, is it a bubble? I mean, theoretically, like, you know, what's twice as much as it should be, but it's also like, well, no, there's an R&D component to this. And the excess spent that wasn't revenue generating last years, but led to models being so good this year, and led to like, everyone who can, using cloud code and like, that changing their life, this is like, it's not a bubble, right? I don't think it's a bubble yet. I think if AI model progress stops, and that's the main thing, right? The moment model progress stops, all the spending is for not. But so far, we've had consistent improvement as you put in more compute, you get more performance and better models. >> Yeah, model performance being lagging indicator of hardware progress or data center. >> Oh, yeah, of catbecks, right? Ultimately, the catbecks that Microsoft spent in 2024 for OpenAI is what results in 2025 for OpenAI or Cori over whoever is what results in their models being so good this year, same with Anthropic and Amazon Google. And their models now being so good now is that catbecks. And actually, they still haven't paid for those chips yet because those chips still have a useful life for another few years, right? I think model progress is very clear. The moment that stops happening, right? If we hit a wall, there's no new research directions, then it's cooked, right? >> And that assumes that better model leads to more demand, which is a reasonable assumption. >> Yeah, for sure. >> But yeah, I mean, this till the adoption curve, regardless of how good the model is, in the end of the year. >> Yeah, but like 2% of GitHub commits today are cloud code. >> Yeah. >> As committed by cloud code, you can disable that where it's not automatically committed. But 2% of GitHub commits today are cloud code. Two trillion dollars of software wages paid in the world. If it was 2%, then you know, like you're like, you're like, wait a second. This is an insane amount. >> Yeah. >> AI is under earning the value that it's producing in the world, right? >> Yeah, yeah. >> By a significant margin already today. >> Boris Cherny from a cloud code who had, who we had on the, but was saying that he's written all of a cloud. What is it called, co-work, like the new product, entirely, that was called code yet. So we've very much in that world, yes. >> Yeah, one of my roommates, I was asking him, 'cause he's like always been a really low level good programmer. And he started, you know, I was like, he's like, he had this holiday obsession, right? I mean, he was using cloud code for work already, right? Like whatever. But he had this holiday obsession. We got into playing Age of Empires too. Myself, you know, my roommate, a handful of people from like Open Eye, GDM, Anthropic. We just would do LAN parties of AOE2 over the holidays a bit. Not like Christmas, but like a little bit before, a little bit after, you know, 'cause most of us went home for Christmas. But like, we'd do these lands. My roommate got so obsessed with like the game that during a Christmas week, 'cause he didn't go home, he just stayed in San Francisco. He just worked on an RTS game. And he built an entire RTS game. And I think, I kid you not. I think he used like $10,000 of cloud in one week. And built an entire RTS from scratch about, but instead of like being a standard RTS, where it's like, oh, Age of Empires are advanced through ages or StarCraft. It is an RTS where it's China versus the US. And you're in the AI race. And you go from the start of the information age all the way through to, you know, AGI and like robots and humanoids and like space-faring civil. Like it's crazy. And built it in a week. And he did type a single line of code, right? He only dictated to the model. And he told me, yeah, like we have an indicator internally at Enthropic where you see how many people, like actually write code now. There's only a few holdouts left. But I guess the question to the bubble is where you question of timing as well, right? It's whether the build, which is a supply side and the demand side are going to land sort of at the same time, is that fair? Yeah, but also the economics of like say, you spend, let's say you spend, you build a gig a lot, you put down roughly $50 billion across the data center, the chips, the networking, blah, blah, blah, blah, right? I'd say it has a five year useful life. So it's $10 billion a year. Is it a bubble if the first year you didn't make any money at zero, the second year at zero? And then third, fourth, fifth year, you're at 50% gross margins and see make 2020. Now you've made $60 billion off of this $50 billion investment. It's not the best return on invested capital, but it did pay for itself. Yeah. Is that a bubble? Well, that's what's happening today is that people are spending all this in front of money on infra and there's no return for a lot of it, right? A lot of it is just doing research and like trying to get adoption and as free users and like, what does that mean? Yeah. Depends a bit on the. The timing, that's the timing though. Yeah. But oh, that $50 billion capEx was spent in year one. What about energy in the data center world? You had this fun post about the gas replacement for energy. So is the AI basically destroying the grid? What if the utilities were willing to let it? But I think the utilities are so slow and dumb that they don't want to not destroy, but like expanding the grid. I think the US can have a way better grid, but we just don't want to. Like no one's made the effort or initiative. You know, there's not enough power. America's not built power for 50 years, really, right? It's like converted from coal to gas and like things like this, but like really just have not built wholesale new power on a large scale. And there have been a lot of times where the industry blew up, right? And so the producers, IPPs have blown up multiple times in the 2010s when Korean and Japanese investors like flooded the market with because they saw such a good return there. Or before in the early 2000s, power was growing a little bit for a little bit. And so people overbuilt on power. So power industry has been burned a couple times. So no one really builds power. And then you've got data centers now all of a sudden coming online and going from 2% to 10% of the US grid in just a handful of years. And so you've got this humongous, humongous change in the industry. We don't have the labor, right? I think ultimately that's the biggest problem is the equipment in the labor. And equipment is basically, you know, again, labor and time takes time to build a factory. So you can build the things. I think the equipment side of things will be solved, like more reasonably. And one example is like gas, right? People initially thought, oh, you can only use like the two vendors, right? Siemens or G. E. Vernova for gas turbines. So they have the best ones, the most efficient ones. It's like, OK, well, like, OK, also Mitsubishi exists. And they're ramping up production fast. Oh, Dusan and Korea exist. Oh, actually, I can just take Cummins engines, right? Like, you know, if you've ever like ridden a pickup truck or like, you know, like diesel trucks, like, everyone loves Cummins, right? You see the ram on the street and has the Cummins like bad. It's like, that's like, or a symbol for a certain kind of redneck from South Georgia, which I have a little bit of. Anyways, I don't have a car on a truck. I have though. But anyways, like, there's like all these engines, like people are figuring out how to make the equipment. You know, solar socks, it's 200 mits, and wind socks, it's 200 mits, nuclear socks, it takes forever to build. Cold socks, it's way too dirty. How do you make power for data centers besides gas? And like, okay, the grid's not willing to put the gas on your site, right? So Elon did and now everyone's doing it, right? These are the cold pose to just last week and two weeks ago that was about the water consumption. Do you want us to talk to that? Yeah, yeah. So there's this annoying thing where everyone's like, oh, AI is using all the water. Oh, wow. AI and data centers are going to like use up all the water. And now we don't have any water. And it's like, that's so silly. Water is a distribution problem. Not a, like, we don't have enough problem, right? Like you look at California, it's like California has a shitloads of water. But people decide to make oat milk, which consumes like 1000X, the water of like anything else, like regular milk even. And and cows obviously consume a lot of water. But anyways, like, you know, data centers consume very little water actually, right? So the US grid will get to like 10% of power by like 28, 27 is data centers. For water consumption, it's not even going to crack 1%. Yeah. By the end of the decade. And what was the metric? And so the comparison we made is because like, you know, it's a bit of a shit post, but it was like serious research. Yeah. Basically like we were doing serious research because we keep getting this like question and demon kicking it and we would do it seriously. But then I was like, no, no, no, this is like too like complicated. Like let's make it very simple. So I was like, guys, why don't we just compare it to like hamburgers, right? Because, because you know, I've heard that argument from some like vegetarian people before or some Hindus or like, I'm Hindu myself, although, you know, and I do eat beef sometimes. I was like, but you know, like I do, but like, you know, so we made this comparison to hamburgers, right? Hamburger requires shitload of water. Because cows, you know, when you, for them, they require a ton of water. And what a cow is taking a lot of water, it's not the cow itself, it's all the feed you're feeding them, right? Because no one grass feeds their cows, you know, and just lets the rain take care of their grass. They like either rain the grass or most likely, they do mass industrial farming of corn, soybean, alpha, et cetera, which uses shitloads of water, right? Like, you know, or like almond milk, like, uses tons and tons of water, like, produce is like the main user of water. I think the metric was the entirety of Elon Musk's Colossus data center, right? Uses as much water as two and a half in and outs. Because that's, you know, you do the calculation on how many, how many birds, what's the average revenue per in and out? And how many hamburgers does that translate to, right? If you, everyone's ordering like a combo, right? Okay, let's ignore the drink, let's ignore the fries, let's just talk about the hamburger, let's ignore the bread, which does use, have grain, let's just do the meat and the cheese. And all of a sudden, all this water is, there's so much water, right? Like a single query, like all of your AI usage from chat GPT of the average user is like a hamburger, right? Like it's like, okay, this is nothing, right? You know, because these things are, the data centers actually are like, they're mostly closed loops and like, sure they evaporate some water for like cooling reasons, but like by doing evaporative cooling, they're using less power, right? And that's actually better for the environment than not using evaporative cooling. There's only, all these reasons why this myth or hoax of AI using all the water is just nonsense, right? Like Meta's data center in Louisiana is getting protested because the water, it's gonna be the largest data center in the world, it's gonna make four or five gigawatts, at least announced so far. We're tracking some other ones that are, that may be as big or bigger. But Meta is getting protested because the local population around that area is like, oh, the water's dirty, it's because of this metadata center. And like there's these trucks on these big trucks on these back roads that used to be empty completely, they're just like mad and annoyed about that, right? But at the end of the day, what actually made the water dirty is that that's an area where you go fracking. Like, fracking is absurdly worse than almost all of that gas is being a ship to an LNG terminal and being shipped to Asia, like, you know, like Japan or Taiwan or China or Korea, and some Europe as well, right? Like, like actually all of this water is dirty because of regulations slash fracking. Like I support fracking by the way, but you know, that's an insane take too, maybe. But like water usage is like not a relevant argument. - Are you bullish on the sort of energy companies thinking constellation for nuclear or Vistra, I guess is an independent power producer? - I think IPPs will do well. I think IPPs can secure contracts at premiums to what they've previously been able to for new power plants that are either dedicated or grid connected, but come with a pairing of a grid load, right? For example, utilities won't let you just do data centers now, but if you come with a pair, right? You're like, "Hey, I'm gonna build this massive data center." But we're also gonna have this massive power generating asset, right? Say, you know, whatever it is, right? Some IPP they're gonna partner with, and they'll build the load and the consumption, even if it's connected through the grid for better stability and more reliability. Or it's not, it's behind the meter, by you not connected to the grid at all. Like some data centers, like partially like clauses from Elon, the original one, or part of Abelines, Texas, open AI, right? Like, Crusoe, there's a lot of room for power producers to get outsized returns. I'm not necessarily bullish nuclear. Existing nuclear fine yet, it'll, it can find a higher buyer, higher priced buyer, but majority of it will be gas. But like you can do like renewals back by gas and then just turn off the gas and like it's cost more, but whatever, right? Or you can do a win back by gas. - No, why not? - Nuclear. - It takes too long. - It takes too long. - No one can build nuclear fast. Even China takes like five years to build nuclear, right? Like it's complicated. It's unsafe, right? You know, I love nuclear. I wish it would work. It's just not relevant in the time scale that like AI's power is going crazy. But yeah, there's a lot of interesting stuff. Like have clients with like, had a client buy a coal plant, and we're advising them on the transaction based on they just like showed up and they're like, yeah, we want to buy power assets. We believe this power story is like, okay, great. So yeah, so here's all of the like power plants that we know of, like you can get some from EIA, blah, blah, blah. Um, which of these like, and then we like work through the economics and we looked at new data centers being built in the region and all this, and then they decided to buy a coal plant and they restarted it. And they're like making tons of money now because now someone, a certain hyper scaler, wants to buy the entire pipeline of power and put a load near it, right? Instead of just being a grid connected asset. So it's like a super awesome investment. So like, you know, power is, power is gonna do great. - Yeah, I was gonna talk about peace dividends of the OliI boom. - Generally yes, right? Like hyper scalers are paying for transmission grid upgrades, which people will benefit from, right? Or like, you know, investors are obviously gonna benefit. People who work in the industry, electricians wages are skyrocketing, you know, et cetera, right? Like plumbers wages are skyrocketing. So there's like a lot of trades that are doing really well too. I think that's definitely also part of it. Yeah. So I'm gonna come back quickly to that Nvidia and Co-We've deal that you mentioned as we sort of close the discussion on CapEx and a bubble. It seems like there is circular deals, but also a lot of debt kind of like flushing around. I don't know the specific, of that deal, but I did hear variations of this were, effectively you have a large player guaranteeing the debt being the last recourse for a lot of infrastructure build. It's sort of this plus the whole like Oracle commitment, there is a fragility into this whole thing that can be a little nerving, what do you make of it? I think it's completely fine. And I think like people are like freaking out and making narratives where there really shouldn't be one. It's like, well, okay, Google doesn't have enough data center capacity and they need people to build data centers, but no one can build a data center because they don't have the cap, but like don't have, you know, make cases capitals not the, you know, they don't have capital, right? No one will give them alone because they don't trust some random fucking company. And it's like, but then Google is like, well, no, we do diligence them. We think they can build it here. We'll like even guarantee we'll buy the thing or start using it once they build it. You know, just having a customer alone spoken for it was enough, right? In the case of core, we've they're actually able to, no backstop, right? They were able to just say, hey, look, here's our Microsoft contract for this many GPUs. I want to put in that data center, that data center, here's the contract for renting those GPUs. I want to hire these people and do this. No one will like, they don't have any money, but then they were able to like have it work out because they were able to get people to lend to them. I think like core, we've did that and there was no circular financing, but that was when there was like, the scale of investment was like single digit billions or less than a billion, right? Now the scale of investment is hundreds of billions. Yeah. And so the question is like, oh, well, if I want data center capacity, how do I get data center capacity? I just go to everyone who's gonna build it, looks smart, is smart enough to do it, but can't afford to do it until them, I'll take it and in fact, I won't just take it. I'll go to your debtor and be like, I'll guarantee you. Because, you know, obviously you're a new company, I've vetted you, but the debtor hasn't. And so, you know, like, you know, they don't want me to just be able to walk away. Because like in the Microsoft core, we've deals, Microsoft could have walked away if core, we fucked it up. Right? Yeah. I mean, yeah, there's always like, sort of like cancellation or whatever possibilities. And so there's just a further form of guarantee, as far as on like a lot of these back stops, as far as on like Oracle getting the money and then opening an eye, getting money and Nvidia, you know, paying and it's a whole circular. It's kind of nonsense because it's like, Nvidia is getting equity in opening eye, they're basically saying, hey, every gigawatt you buy will also buy some equity. Yeah. Right? Okay, well cool. Now Nvidia owns an asset, which they think is valuable. Opening eye. Opening eye is turning around and is like trying to rent those, uh, use that equity. They buy, what are they, what are they, what are they, what are they use of equity? People's cash pay isn't that great, right? It's mostly just 99 plus percent of their spend at the company's probably just compute. Yeah. So, sort of like, it's like, okay, well then I, I raised this money. I'm going to do the whole thing where I explained earlier, right? Year one and two. I lose money. I'm going to be on it, right? Um, and opening eye has been doing that, right? So I'm going to, okay, I'm going to go out there. I've raised $50 billion. I've raised $10 billion. I'm going to raise it. I'm going to rent a cluster for five years for $65 billion. And I've rented that contract and now I only have enough to pay for the first year to be clear. But I think, you know, you trust me or cool, you think I'm going to grow and you think I'll be able to pay for oracles again. Or if you're not, I think I'll be able to sell to someone else. So like, okay, cool. I'm going to spend $50 billion this year to build that data center. And this, this is like four giga lot. And so is it like circular that open a eyes every amount of GPUs they consume, maybe gives an investment, that investment has turned around to pay for the first year of the rent of the cluster or second year. And then first two years go, you know, it's sort of like, it's fine. Yeah, yeah. Like, it's like, it's like, it is a little bit funky, but like, I don't think it's a big deal. Yeah, love it. Control and take. Maybe let's finish with the models and the software side of things. We talked extensively about hardware and supply chain and all the things. I get a sense that you are super, super bullish on what's happening next in in AI your roommate, Shulto, I assume was the roommate that you were talking about earlier on this part, particularly making the point that we're just starting to scratch the surface. And there was so much looking in fruit around, you know, RL and all the things. You're in Silicon Valley circles. Is that your sense as well? And what are you tracking on the model side? One thing is like, you know, simple stuff like GitHub commits. Other things are like, what's the amount of usage? How much are people using like all these sorts of things? I think there's so many different alternative data sources for tracking AI model progress, area, tokenomics, token economics, tokenomics. And so that's like an entire practice for us. Are you really branding the term from crypto? Yeah, I don't believe in crypto people. Like, I've always hated them. So now you're taking the term. Yeah, and Jensen's used it now. So I've like, I've convinced him to use the word, he's used it to sovereigns. And so I think we want. That's awesome. I've said it to him. We've written it in articles. It's an entire practice of consulting that I just started in like 23, 2023 was token economics. And we've been trying to build out these like, you know, but basically I think the main things are like, people who don't code can use cloud code now. Right. I think people don't understand that. Like, even if you don't code, you've never had any training in cloud software development. You've never taken, how to job as a software developer. You can code. Let's take an example of what's one of the, one of the analysts at my company did, right? Comes from a engineering background, but on like a semiconductor systems, right? Like worked on mechanical systems, worked on these sorts of things. And they coded this thing, which was they want to do an analysis of area of clean rooms, right? Clean rooms are the building that you, the fab has all the tools and the most complicated kind of building in the world, has every all sorts of chemical systems and all this. Area of that, a company who builds systems, builds these systems and revenue of that company, right? And so it was like, okay, we have this fab data set pointed to that is like, hey, here's this fab data set. What's the square footage of all of them? And we have this like thing that we built, which just pulls with cloud code separately, which for data centers and, and fabs and everything else just calculates the area of something from a calc from a satellite inventory. Very simple. So we have the square footage of all these things. What's that? Here's the company name. Okay, go find the filing. So it dig, dig through all these filings. It pulled the data, right? Okay, great. Now I'll tell it to compare these two. Make a chart. Great. Oh wait, there's this like weird inflection. Oh, that's because they bought a company five years ago. Can you do a pro form of this analysis without those financials of that company they acquired? Okay, great. And then we were able to like, like, figure out an investment case for clients, as well as like, you know, some other interesting details from someone who's never really coded just using cloud code and it like doing this all. And this is like not even their, and it wrote the note. And they just like, they didn't even like work on this full time for like three hours. Right, they just told the model and would go work on other things and told the model and worked on other things. It just did this. People don't understand that like the skill sets that like, I think like if you go talk to an analyst, right, a very junior analyst at any company, right, whether it's venture or, especially growth venture or public markets or private equity, their job is like finding data, cleaning it, making charts. It's like, this is cloud code. No, you don't need junior analysts. Just like a lot of companies have stopped hiring L4 engineers because it's useless. Why would I hire L4 engineer? I just tell cloud to do it. You sort of like have this has happened. And this is a really big like shift, I guess. Like, is that like low level knowledge work just doesn't matter, right? Why would I, why would I use Excel when I can just help cloud to manipulate CS fees? Why would I use word when cloud will just generate the mark down and I can copy and paste the mark down directly into our WordPress and then, you know, that WordPress is fully formatted now. And it's like, oh my god, like what's the point of word, right? And what's the point of doing all sorts of stuff? I think when we look at model progress, that's just four opus 4.5. Opening as new model, I think will be better than opus 4.5. And it's coming like somewhat soon in March-ish timeframe. I think maybe February March-ish, but yeah. Because OpenAI has a better RL stack than Anthropic today. It's just their pre-trained model suck compared to Anthropic's pre-training, right? And so like if they catch up a lot on pre-training and keep their better RL stack, they would actually have a model that's much better, right? Flipside, Google has a better pre-trained model than Anthropic or OpenAI, but their RL stack sucks. So if they catch up on RL, like these models are going to get ridiculously, and then Anthropic is obviously advancing as well, right? And then, and then you look across the ecosystem, everyone's advancing really fast progress. These moments are happening, right? You know, chat GPT was a moment, Ghibli was a moment. Those are more consumer, those are less like, I mean, there are chat GPT's everyone using it for work too, but like, I think Cloud Code is like a new moment, right? Opus 4.5 on Cloud Code is a new moment where the way you work has forever changed. And so now we're trying to force everyone in my company. There's 54 people here, I think like half of them have coded. The other half, we're trying to force them to use like, Cloud Code, and it could be like, oh, well, actually you come from a consultant, semiconductor consulting background. Oh, you come from a semiconductor engineering of like, package, oh, you worked in a fab, right? Like these kind of people, they're using Cloud Code now, right? And their productivity's being boosted. And it's like, you know, workspace, Cloud workspace is new, it sucks compared to Cloud Code, but it'll get there, right? He said he coded it entirely in Cloud Code, you know that, right? I mean, it was on your pod, right? Yeah, yeah, yeah. So like, you know, I've heard that, and I think maybe that might have been from your pod, original disclosure. My part was before that, but yes. Oh, okay, okay, well. It's the guy on my pod, but he subsequently said that. Oh, okay, okay. I think it's like a brand new agent, and like, there's so much the hanging fruit. As Shultza said on the episode when he was here, there's so much the hanging fruit. Yeah, I mean, for the models progressing, and then I think model progress will translate to revenue. Adoption is difficult, but like actually, the UX of Cloud Code sucks, but like give it six months, the models would be good enough that the UX can be like, talking to it. And you don't even have to have like, you know, CLI integration, right? It's something even easier. Or like, Cloud for Excel was released recently, and it's like not bad. You know, building models and like, all these sorts of things are just gonna be like, tell someone, right? Like, why tell a junior analyst, right? When you can just do it yourself. I think it's a whole new world, and it's a two trillion dollars of software work, but also of wages, but it's also, we have more north of 2% is Cloud, and then there's code acts and cursor and all these other guys. So probably like 5% of code committed today is, hey, I generated, if not higher. Marked as I generated, what's gonna happen when normal workers who do spreadsheets and office processing start automating their workflows, I think it's a whole new world. - And speaking of a short toe, we both agreed that he was a perfect specimen. - Dude, I've been, I'm straight, but I've been accused of being almost extra, which is perfectly fine for how much I like, praise this man because like think about it, right? He's like six foot four. He's like really good looking. He's like Australian accent sounds amazing. Like you've heard his episode, but I have like a annoying voice, probably. His voice sounds amazing. He's absurdly good at coding. He's an Olympian level fencer. Like he picks up any sport. He's really good at it, right? Because he's athletic. It's like holy crap, you're a specimen. - Yeah, yeah. - So this is going to be a good tip for sure. (laughing) - Yeah, you must be, you know, I guess maybe some people don't follow the play-by-play on Twitter and like don't even heard of like the fact that all of you guys are roommates, you roommate with short toe, and then with Dwarkish and Dwarkish is like the podcaster's podcaster. So it must be absolutely. - What's a podcaster's podcaster mean? - The podcaster that all the podcasters aspire to become all alone from. - Yeah, yeah. When he's preparing, you know, it's like he's so locked in and he prepares so hard for interviews. - Yeah, it's great. - No, he's just incredible. - And then he might only say like 100 words on the episode. But he's prepared so hard. And then like I think people just realized, oh wow, he's not just like, you know, it's like oh, he just has good guess. Like no, no, no, like he's preparing really hard. But you can't tell if you're not like realizing that. And then once he started writing more and he started writing more, people are like, oh wow, he's actually really, really smart. It's like yeah, 'cause he's studying like crazy. Like it's like, oh, I'm interviewing an AI researcher who worked on this. I'm gonna try and train a freaking model. - Yeah. - Right, it's like that's the level of like commitment he goes to when he records his stuff. - What do you guys talk about when you bump into each other? He's like, is that AI nonstop? Will you talk about everything but AI? - With Show Tots, like the age of Empires game, you know, 'cause we got super into it for a bit. We talked only about that in his RTS that he made with with with Tbarcash. It's all sorts. It's like normal roommate stuff. It's like how's your dating life? Oh, okay, you went on a date. It wasn't well. It didn't go well. Okay, well, okay, yeah, you know, like oh, you know, like that's me, that's me. You know, my dates don't go well. No, she's kidding. Or like it's like, oh, you wanna like have dinner. We can invite a few friends. Like, you know, great. Or like, you know, it's like all sorts of like normal stuff too. Also, obviously we also do talk about a lot about tech, right? Like we are like, this is our lives. And tech is the most fun thing. - Awesome. Well, great. Great San Francisco lore. Dylan, thank you so much. That was absolutely fabulous. We enjoyed it. Learned a lot. So we appreciate you're coming on the pub. - Thank you so much. - Hi, it's Matt Terk again. Thanks for listening to this episode of The Mad Podcast. If you enjoyed it, would be very grateful if you would consider subscribing. If you haven't already, or leaving a positive review or comment on whichever platform you're watching this or listening to this episode from, this really helps us build a podcast and get great guests. Thanks and see you at the next episode.