Limitless Podcast - The AI Energy Stack: The Actual Industry Winners

Episode Date: July 1, 2026

This is the long-awaited Energy Stack episode. The AI trade is shifting from GPUs to the power, wiring, and data center infrastructure needed to support them. Today, we cover Bloom Energy, optical networking companies like Lumentum and Corning, Marvell’s role in AI hardware, and neocloud providers such as Nebius, CoreWeave, and IREN.------🌌 LIMITLESS HQ ⬇️NEWSLETTER:    https://limitlessft.substack.com/FOLLOW ON X:   https://x.com/LimitlessFTSPOTIFY:             https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQAPPLE:                 https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890RSS FEED:           https://limitlessft.substack.com/------TIMESTAMPS0:00 GPU Bottleneck Shifts0:37 Power Becomes the Constraint3:15 Five-Year Grid Delays5:47 Bloom’s Fast Energy Fix9:54 Data Transfer Hits Limits11:51 Light Replaces Copper17:25 Power Delivery Gets Harder18:22 Marvell Wins Power Design20:35 Renting Compute at Scale21:08 Nebius and NeoClouds25:17 Compute Becomes an Asset29:47 The Substrate Layer Ahead------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠Josh works with Anthropic as a contractor. All views expressed are his own and do not represent Anthropic, its leadership, or its affiliates. Nothing in this episode is investment advice.

For three years, the AI trade has been one simple idea, buy the chips. Invidia became the most valuable company on Earth by making these chips, and they've made countless of investors rich because of this. But this investment thesis now has become overcrowded. Not because AI is slowing down, but because there's another problem that has surfaced. No one can turn these GPUs on. Trillions of dollars being spent in AI, and all these GPUs are collecting dust in data centers. The problem, energy, electrical grids, wiring, networking, the transfer.

form is the web of infrastructure that sits around a GPU that allows it to keep alive, to talk to each other to transmit petabytes of data between each other. This is the next constraint that hasn't been solved yet and where the majority of the AI capital will eventually flow. Now, there's four physical things that keep a chip alive, essentially, the power to run it, the light to transmit all the data between them, the silicon wiring between them all, and then somewhere to rent it all from. And almost nobody is really talking about this.

And on this episode, we're going to unpack that specific layer. of the infrastructure stack, the power layer, and why it's so important going forwards in terms of capital. Yeah, and this may feel a little bit different because we're always talking about whose model is the smartest. That's the common conversation. But once you strip away all that abstraction, a token is really just what is output. And that is just a series of electrons flowing through chips, light flowing through fiber, and then heat being pulled out of the system. And all of that requires a lot of energy and electricity. The money cycle that we're going to talk about has kind of flowed towards this direction as well. I mean, everyone started with the crowd of GPU trade,

then it flowed to the semis that exist around them, then it flowed to the memory trade, and now it's kind of moving over to the seemingly the end game bottleneck, which is energy. I mean, US data power center demand, I think is projected to roughly double 80 gigawatts in 2026 to 150 by 2028. So the grid is not really built for that. And there's no shortage of demand for that. And when we think about energy as an idea, even if you believe that the AI bubble is towards the latter end of it, energy is still something that's not going away. We had energy problems prior to the LLM becoming a big deal from the chat GPT moment. This is just an extension and an exaggeration on top of that. And it creates a lot of interesting opportunities in the marketplace

to actually participate in this bottleneck that is now known as I guess we call it the energy battle neck. Yeah. And I think a lot of the reason why people don't talk about this is because it's sort of unsexy, you know, not everyone is an electrical grid expert. Not everyone knows how to network wires between these complex GPUs. I'm just still trying to wrap my head around the GPU itself, right? So like the fact that like we're bringing in all these other things, it's quite complex. Now, if you were to look at a budget of spending, right, for an AI lab or someone that's setting up a data center, only 10% of that budget gets allocated to the power and infrastructure side of things. The irony of it now is it's going close to 100% of the bottleneck of the entire thing.

So even though it's not the majority of the cost, still 90% of the cost goes into the actual GPUs itself. That's the most expensive stuff of the memory and the silicon required to build those GPUs. Only 10% is allocated to power, but that is like the main bottleneck that we're facing today. Now, the lead tire, if you are an A.m. You could be the richest, most important person on Earth right now. If you want to build a data center in the US, the lead time to get the necessary power to your GPUs is five years. And there are different sectors within the power thing.

It's not just the energy infrastructure grid you need to get access to. You need to get lead times for wiring and to get transformers designed, custom made, and delivered to you. All of that takes a lot of time. And we're looking at basically like half a decade. So what we've seen a lot of these AI labs start to do now is figure out what alternatives they could potentially procure to help them get GPUs online. Now, you've all heard the news of meta, Amazon, Microsoft, and Google spending trillions of dollars this year alone.

I think it was something along the lines of, actually, it might have hit a trillion dollars for this year, right? Because I think after the last quarterly earnings, they like up their budget, they haven't set those data centers up. In fact, there have been delays. Stargate from Open AI, which said that was their biggest project to kind of get. compute online has been delayed multiple times. And the point around this is because of compute. And not many people are talking about this yet. So I think on today's episode, as we walk down this infrastructure stack, and we did a previous episode before where we covered all the layers of the AI infrastructure stack, we talked about going from layer zero, which is the model labs down to layer one, which was the hypers, GPU semiconductors. We didn't dig as much into layer five,

which is the power and infrastructure side of things. And that's what we're going to do today. and we're going to do it through the lens of four specific companies, starting with Bloom Energy. Yeah, and just before we get into Bloom, there's one thing. Well, one of the things is, I think there's these bottlenecks within bottlenecks. You mentioned the lead time, but also the actual infrastructure is becoming much more demanding. Vera Rubin, the new chips, I believe, those server racks take over 100 kilowatts, which is a 10x increase from what these companies and these data centers we're preparing for. And then also before that, I want to kind of loosely define the power for people who are not familiar because there's watts, there's watts hours, there gigawatts,

just like a brief one-on-one quickly on what everything is. A watt is a rate. It's how fast these data centers draw electricity. A watt hour is an amount, which is the total you drew. So a gigawatt is a billion of those watts drawn. One gigawatt is equal to roughly one large nuclear reactor or enough to power 750 to a million U.S. homes. So this is a tremendous amount of data that we're talking about, or a tremendous amount of energy that we're talking about. And these data centers are demanding tens to hopefully soon hundreds of gigawatts. Now, back to Blum Energy, Bloom Energy has a solution for this. It is solving the most painful solution in AI, which is this energy data source. They use these kind of modular energy reactors that allow them to, allow these

data centers to remove themselves off of the grid and generate a little bit more energy closer to the data center. Yeah, actually, remember the first time I heard about Bloom Energy at all was in Leopold Ashrenner's 13-hour filings at the end of last year. It was one of his Biggest positions, I think in his recent one, I think it makes up about 12.7% of his entire portfolio. So he's very bullish this particular company. And the reason why is, as I mentioned earlier, the lead times to getting power generation on site for your data center. If you go through the old traditional way, is around half a decade. Bloom energy will reduce that down to 90 days.

They have a product. It's a gas turbine, which they can custom build and deliver on site. They kind of like fly it in and they re-accent. it and it can generate enough power for you to run your GPUs. Now, obviously, that's a very attractive thing, a ton of companies have signed contracts with them. Now, the one caveat that I will say is they haven't really delivered this at scale. Bloom Energy has yet to prove themselves from a manufacturing capacity and delivery capacity, but it is a company that a lot of people like, like to talk about and kind of like speak highly of. Now, the revenue kind of speaks for itself as well,

because obviously there's a lot of lead time. There's a lot of demand here. And as we mentioned, that money is going to flow down. As of Q1, 2026, their revenue hits $750 million. That is up 130% of revenue year over year, which is just like a crazy kind of swing. And of that, they're making around 17 million profit, which isn't large in the context of like some of these big hyperscaling infrastructure companies, but for a relatively small company, this is pretty impressive. And Oracle is powering its 2.4 gigawatt data center primarily using Bloom Energy. So they've signed some pretty major contracts as well. So Bloom Energy is one of these companies. And there's various different versions of this where it's not just gas turbines. Maybe it's solar energy or nuclear energy. But Blue Energy is basically the first company that is having an impact almost immediately after this constraint has been identified.

Yeah, well, in terms of gas turbines, I mean, that's one way of doing it. And the line to get a gas turbine, extends back to 2029. So you actually want a gas turbine at your data center and you place an order today. And for those who aren't familiar, a gas turbine is basically a jet engine, which is spinning a generator. And it's so funny how this works. It's like the way we generate energy today is the same way we did with like a steam engine way back in the day. You're just basically creating energy to propel a spinning thing. In this case, it's a jet engine.

If you want one of those, you're waiting at least three years, probably the next decade. So if you want something in the 2020s, that's not for you. The next option is nuclear or small nuclear reactors. Those are amazing. They're incredible. But if you want them in this decade, again, not possible. Those aren't probably coming online at scale until around 2030. And the licensing and legislation around that is pretty slow.

So that leaves us with these fuel cells. And that's what Bloom has. And these modular boxes, you stand up on site in 90 days instead of three to five years. And so far people have really been liking them. We've seen that reflected in the stock price with people desperate for energy. anyone who's able to actually deliver it at a accelerated timeline is going to win. I think about Elon all the time and how they were able to build the Colossus Data Center and how much premium it had to getting GPUs online quickly.

I mean, think about how much Anthropic and Google are paying for these GPUs. The same thing is true with electricity. If a data center is bottlenecked by this power, anyone who's able to provide it for them is going to get paid a tremendous amount of money and a large margin for it. And so far, Bloom has been one of those companies through their fuel cell program, that can be deployed in 90 days on site, they're collecting a lot of that value. So I think that's kind of the bullcase for Bloom.

That's why people are mostly excited about it. It's just a really difficult place for companies because it's hard to make energy. And a company that has figured out a way. I'm not sure this is the end state, but this is certainly a intermediary step. And right now, that's all people need. It's just power today. Now, moving on, one of the other bottlenecks that is appearing, particularly with GPUs talking to each other is data transfer specifically. Now, if we think of an AI model,

typically it has grown exponentially in size, in quantity of data that is required to train a frontier AI model. We've gone from, if you remember, Josh, billions of parameters to trillions of parameters and now tens of trillions of parameters. So the model weights of these actual AI models are huge. People are carrying them around in briefcases. They're like, incredible. incredibly dense. And when you're training a model, and when you're even inferencing a model, when you're trying to speak to it, when you're trying to send it a prompt and get a response, there is petabytes of data, which is an order of magnitude larger than anything we've ever really spoken about at scale in industry, at least, before, that are transferred between GPUs. So it's not just good enough to have the GPUs and wiring them up and having the power flowing to them. You need to transfer data between them. They need to talk to each other in rapid pace. If they talk to each other slowly, then they won't be able to. to work as quickly and as powerfully as you'd expect. Think of it like this. Every single way that a GPU requires to function,

the lifeload of it, the energy, the ability to communicate is a potential constraint. And this communication thing is another one. Now, typically the way you achieve this is through copper wiring. Copper has a really interesting chemical property that allows you to transmit data at lightning speed without heating up,

except we've run into a problem. We now have too much data. And so the copper wires are heating up, which causes energy to dissipate, and energy efficiency and cost efficiency, therefore, to not really work in the favor of the AI CAPEX that is being spent on this. So people have started scrambling around. They started looking around for another type of material, and you wouldn't believe it, they've settled on light, optical fibers to transmit data using light.

And one of the companies that is moving data as light is known as Lumentum. They're using optical fibers and laser transmitters to move data between, chips as light because copper melts at a certain level where data becomes too burdensome, essentially. Yeah, the bottleneck was very much copper, and it might still be to some extent, because copper is the fastest way of transferring data, just because it has so much bandwidth. It is so dense and you could transfer. I mean, now we have clusters of hopefully soon half a million GPUs. I know that's what people are working on. All those GPUs need to communicate. They need to do so as fast as possible. Copper has enough bandwidth. Unfortunately, there's this direct correlation

between the amount of data and the length it has to travel. It's physically impossible to put that many GPUs that close to each other to make it feasible. There's just not enough space inside of a room to make it happen. So as these data centers get larger, the distance in which data needs to travel gets further, copper becomes a worse and worse option. Because, I mean, every single bit that has to flow through that, it causes heat, it causes energy, and then like you mentioned, the copper, it's going to start melting. And it's really expensive to cool, and it's really difficult.

So what's the next best thing is photonics. You just use light and you use fiber optics. And nothing travels faster than the speed of light. So it's just a matter of how much bandwidth we could squeeze into those rays of light. And by stacking this fiber and by using this glass fiber, we're able to actually transfer information from one ship to another at the speed of light. And Lumentum is a company that's working on that. So I think that's why people are really excited about this company in particular. As these data centers grow larger in footprint, they are going to require.

a longer distance in which data has to travel. And there's going to be a lot more data, which means copper probably won't be ideal. And unless we solve this material problem, the next best thing is photonics. It's just sending it right over a fiber optic line at the speed of light, or I should say, many fiber optic lines to store as much data as possible in this transfer. And that's why Lumentum is, I think, one of the more exciting companies. This is part of Jensen's portfolio, if I am not mistaken, right? Yeah, yeah. He's put in about two billion into Lamentum Holdings. But to your point, it's important to mention that this is a relatively new type of technology.

It hasn't quite hit that exponential wave that Nvidia GPUs did two years ago. And so we're waiting on that. The bet is that that is going to happen for these particular types of transmitters and receivers. Now, their units of sales is going up pretty aggressively. Last year, they sold, I think, around like 20 million units this year. It's now 60 million units. So it's kind of like a two and a half to three X from one. what they were last year.

But they have competitors in the form of Corning, GLW, who, and VDIV also invested $3.2 billion. Now, they're not direct competitors, but Corning creates the actual optical glass fibers that will be used to transmit light and data for this. And there are many other companies that are highly focused on this. It's worth mentioning though that scaling any of these material goods,

whether it's the glass fibers, whether it's transmitters, whether it's the power switches, or whatever that might be, takes an incredible amount of lead time. Like, one of the major critics of the memory bottleneck that everyone is quite familiar with right now is Apple just recently announced that they're raising the prices of all their devices because they can't get their hands on enough memory.

Nvidia is facing the same thing. And when they go to SK Hynux and the memory makers, and they say, hey, can you just build more, please? They say, it takes time. The machinery is incredibly detailed. Like, it takes a while. And we're going to be in this for a few years. and the optical fiber side of things,

the material side of power infrastructure in itself is where this is also playing out. Nothing's really changed in that sense. Also, Eja, they just sent you a custom link of a post that I'd like to share because this reminded me of an idea that I heard that I think is probably relevant because the metric underneath all of this

is joules per bit. It's the energy costs to move one bit. And at scale, light wins decisively. Like, there is nothing that flows smoother than light. And it reminded me of this post from John Carmack He is the guy who created Dune. He's like this internet legend. And basically, John was saying that you can store data within this light if it is transferred

over a certain length of space. You remember this? Right? Yes. Yes. It is amazing. So when we think about memory and we think about data transfer in general, I think fiber optics is a really interesting industry to consider that can go much further than I

think what we imagine it can do today. Because during the time in which this light, or these bits are being used. transferred at the speed of light, you can transfer a lot of data. And if that's spun over a long enough distance, it can actually store data in a way that is very fast and accessible, way faster than we get in memory and RAM, because it is actually moving at the speed of light. And there's a lot of interesting science experiments. And I think general experimentation as it relates to fiber in general, not only as a way of transferring data, but as a way of storing data, should you stack it and extend it over a long period of time. And I just wanted to highlight this as like, Lumenum is, like,

Lumenum is hot right now, but I think the industry of fiber and fiber optics and moving things around at the speed of light is ultimately where we're going to land on. We talk about data centers in space. They're going to be traveling at the speed of light. It has the lowest amount of friction and has the highest velocity of information. If we can experiment and kind of iterate on the way that we do that, I think that's something worth noting. And Lumentum being the fiber optic company may stand to benefit over a much longer period of time than I think people think based on this current bottleneck. I just want to highlight that because I was like, huh, that's pretty cool. That reminds me of something I saw back in the day. Oh, I love it. Now, you would think that the entire problem with the power infrastructure

stack is just generating the power, right, Josh? But there's also delivering the power, making sure the right amounts of power go between the chips at this time, and then when they are really into mind delivering even more power, right? That takes incredibly complex design patterns. Now, most people probably aren't aware of this, but Nvidia isn't a chip manufacturing company. They're a chip design company. They design. what the silicon wafer is going to look like, and then they send that over to their friends in Taiwan and TSM, and they actually manufacture the thing for them.

Now, one key component of manufacturing this design is figuring out how the power is transmitted between these chips and within the chip itself. And this is a huge role that is played within the power infrastructure because it determines where the power gets dispersed. It's an incredibly hard thing to figure out. Now, about three weeks ago, There was this very popular compute conference called Compute Text, which was basically in video GTC2.0, Jensen was like the bell of the ball there. So he goes on stage and he talks about this company Marvel and he says, this is the next trillion dollar company. And this is an investment advice, by the way. This is literally verbatim what he said in Taiwan. And the stock subsequently went up 76% after he mentioned this. And I was looking up on this just to see, you know, have they like, you know,

rescinded since then. Josh, they broke into the SEP 500 based off of this single news. And the reason the TLDR of why Marvel is attracting so much attention and capital right now is because they are experts in designing that part of the chip, which determines where power gets dispersed, how it gets dispersed, and where it gets used within the chip and between chips. And if you're wondering, oh, okay, well, they're just solely relying on Nvidia. The answer is no. Broadcom and every other single AI company that is trying to design their own chip and diversify away from Nvidia is using Marvel. Wow, it's up 80% on the month I'm seeing in one month since Jensen mentioned this company. I actually don't know much about Marvel and I feel bad about it because

clearly I missed out. But it's one of those things where it's been very lucrative to copy trade the people in authority. It's like if someone who is running the largest company in the world says this company is going to a trillion dollars, people are going to react, and you have to just trust the fact that he has a good reason to do this. I mean, Jensen seems like he holds, he holds pretty strong generally on his morals. He's not trying to pump and dump. He genuinely believes this. He has made a sizable investment. What is the total amount? It's a couple billion dollars, right? Two billion dollars. Yeah. Two billion dollars, by the way. So he's up, he's up pretty big. Yeah. Man, my God, exactly. That's the point. And it's like, yeah, he actually has the ability to influence

the success of this company, not just by talking about it, but by actually interring the technology into his company. Same thing happens when you talk about the person in charge of the United States of America. Like when Trump says something, it generally, it goes up. Generally, there's a good reason. People are looking for projects to invest in, for companies to invest in, and Marvel has been one of the most recent winners. Now, there is another section of this stack in which you have to rent out the compute. Say that you are not, you don't have hundreds of billions of dollars of KAPX at your disposal this year. Sorry to hear that. Chances are you're not winning the AI race, but you can at least make an impact. You can do something novel, something interesting in AI,

and if that's the case, you're going to need to rent that compute from someone. There's a bunch of companies that we've talked about on the show, in which you can rent compute from. We talk about Iran, a lot of the neoclouds, but we have a new one to talk about today, which is called Nebius. Can you explain to everyone what Nebius is, what makes them so different and unique? It's actually not what makes them different or unique. It's just, it's a scar service that, that they supply. So if you've heard of a core weave, or if you've heard of iron that Josh just mentioned,

they're known as a Neo Cloud. A Neo Cloud is basically a company that will come in, they will construct the data center for you, they will handle all the wiring, they'll make sure that the GPs talk to each other, they'll make sure that the networking and transformers and receivers, they'll make sure that the copper doesn't melt, and they'll make sure that these beautiful, precious GPUs

that you just spend billions of dollars, on function optimally. The reason why someone will go to a Nebius or a Corleave is simple. Google, Amazon, and all of those companies, they don't want to focus and spend time operating and maintaining and designing the flow of a data center. They want someone else to handle them. That's what Nebius does, and they do it pretty damn well.

I believe Nebius's background is they operated a bunch of data centers that focused primarily on maybe it was mining in the Web3 side of things, maybe it was something else, but they pivoted pretty aggressively to focus on AI specifically and their expertise of managing hardware and most importantly, the software that talks to each of these GPUs and the different hardware components is their expertise. So they aren't so truthfully is they do exactly what CoreWeave does. They do exactly what Iron does, but they do it at a massive scale. And their Q1 revenue is up almost 700% on the year. It's just insane. Yeah, that's crazy. I saw it at $400 million. And this is the thing with the demand issues is that like if you build it, they will come. It doesn't matter who

you are. One fun fact about Nebius is they were previously known as Yandex, or this is what was built from the remains of Yandex. If you are from Russia, you probably know this because that was very much the Google of Russia up until 2024 when they severed and they started to build a neocloud. So that's a, yeah, they are familiar with managing data and storing data and uptime with servers. Wow. Yeah. So that was interesting. And then I think one of the things that makes this company special, again, NVIDIA is a backer. And video has placed $2 billion, a direct equity into this company. And Vitya should just like, man, if they weren't making GPU, Jensen could just make it as a VC, because this dude is making pretty unbelievable.

Yeah, two billion. Yeah, maybe we just screenshot this portfolio and I just send it to my broker and say, can you buy all of these for me, just split it across my whole portfolio? Because clearly Jensen knows what he's doing. But the company is doing remarkably well, because they have contracts from companies that you actually have heard of. They have $20 billion contract from Microsoft. They have a $27 billion contract from META, and then they have a backlog that's approaching $50 billion for the years, 27, all the way out to 2031. And this is the thing that we talk about a lot, too, is that the useful life of a GPU has extended so far that I think these companies are getting revalued and re-underwritten based on that, because a lot of the time

people would say that, oh, if you have a GB100 today, it's going to be worth 10% of what it is today five years from now. And there's a five-year depreciation schedule. The reality is, is that GPU from three to four years ago is worth more today per hour than it was back then. And so long as this demand curve continues, these GPUs, the useful life of them is going to continue to extend out over longer and longer durations, making the inventory of these companies more and more valuable. And if that trend continues, everyone's going to have to continue to re-underwrite these companies based on this new depreciation scale. And that is going to continue to bring the value of the company up. They have the scarce resource, which is the ability to actually build these data centers,

which means they have the land, they have the power, they have the shell. And this is another company that is doing it pretty well. So Nebius is a good example. Neoclads in general are a good example of just like this insane demand that is insatiable. It's like no one can actually provide enough energy, enough GPUs that are powered and online. And any company that's able to contribute to that is just going to continue to grow so long as this demand stays there. And there's no signs of it slowing down. You know, as we talk about this, it's so interesting to see this sort of, it's like this entire ecosystem that has built around the GPU. GPUs are basically the gold bar, right? There's this whole economy around this. There's news that I was seeing earlier

this week that the Chicago Mercantile Exchange, CMMC, I believe, is setting up a compute futures trading index so that you can like hedge trades and do trades based off of compute and GPU infrastructure. It is becoming a financial instrument that you can kind of like eventually borrow money off of. That's what Nvidia and a ton of other companies are doing. And And so it's this kind of like pseudo-proxy investment vehicle where it's just based around GPUs and this whole ecosystem is now blossoming around it, whether it's neoclouds,

whether it's optical light fibers, whether it's the silicon itself, whether it's the chip design that helps transmit all this kind of power. It is formed around, the gravitational pull of these GPUs is pretty insane. One final note that I wanted to mention for Nebios specifically, and they have the same advantage that CoreWeave and Iron has, is permitting and licenses, like, yawn, right?

Like, who the hell cares? As you mentioned earlier, a lot of the five-year lead time to get, like, power and energy to your GPUs is regulatory. It's red tape. It's trying to jump through the government hoops to get all of this figured out.

Nebius, iron, cor weave, all have this permitting in advanced stages. And that's why they're signing these multi-billion-dollar deals with Meta, Google and all the other hypers because they have this advantage. It's all a regulatory game. It's all a hardware game.

Whoever can figure out a way to arbitrage the red tape that they're currently facing will end up winning pretty hugely. And if you look at what Leopold Ashenbren is investing and if you look at what Jen Sir Puan is investing in, the people who are on the ground that are seeing these bottlenecks in real time, that can kind of give you an idea of where that AI capital is eventually going to flow. Yeah.

And then one final point on the neoclots, which I think is fun and underrated, and a lot of people don't recognize what the incentive is to go with them, as if you don't need the GPUs. But if you're a large company like Meta, for example, who has that $27 billion deal with Nebius, you can actually write off that spend as an expense spread over years instead of just like hammering your free cash flow with this increased Kappex today. And for a lot of earnings reports we see, the first numbers we go to look at are the Kappex. How much are they going to be spending on building out compute? And if that number is too high based on what the market is. it believes they can actually build, then they get crushed for it. So a way to kind of obfuscate that is by investing in these neoclouds, signing these deals that can then be extended out over the years and don't immediately crush your balance sheet like we're seeing a lot of these companies doing. Google being one and Vida being one where they just raised $25 billion. So it's an interesting financial instrument as well that these companies are using. But I think that is mostly it in the hardware energy stack. That's where we are. How do we do? How do we do? I'm curious. a bit asking for this episode for a while. I feel like we did a good job kind of shepherding the story

between these companies. Josh, do you feel the same? I feel pretty good about it. I think there's this like, the layers go so deep. And you can even go deeper than this with like the actual, um, manufacturing of this and like who is supplying all the raw materials in. But I think this is about as low as you need to get. Because this is the, this is the interesting part. Past this, you start getting into all the science and physics that I'm not sure it's super interesting, but, I feel good about this. I actually learned a lot in preparing and recording for this episode as it relates to just like where the bottlenecks are, who's responsible for them. I didn't even know what Marvel was until Jensen talked about them on stage. So we're learning a lot about

them. There's a lot of alpha and coming across these early. I mean, all of the companies so far, they've done pretty well. It's no lie. Like people know about this. This is not this like deep alpha. But there is a really great opportunity and these companies are building valuable goods and services for companies that need them. So yeah, I think, That's probably it. That covers the whole thing. We went through the stack all the way down to the bottom. We previously covered the higher parts of the stack. If you're interested in that episode, you can find it. We'll link it in the description. It's also in our channel. You may have heard it if you've been around. But yeah, I think that's pretty much it. That is covering the base layer, the energy infra layer of this AI stack. There is one more layer, Josh, that I, not today, not next week, but maybe in a few months time that we need to prep hard for, which is the, the substrate layer. That is the one at the bottom of the bottom. That is, we're talking, we're going back to old school days. We're talking about raw materials here that actually help

form the materials, the silicon itself, the wiring that we've been speaking about. Where did the atomic layer? Like, fun little nugget, an Easter egg before you even hear anything about this episode is the majority of those materials, they're in Japan. And Japan has a company that manufactures toilets. It's known as Toto. In fact, if you look at your toilet right now, you might notice that... We filmed an episode about this. We filmed an episode about Toto. They happen to have machines which get access to this really funky material that is only accessible in Japan or primarily based in Japan that is used to make GPUs. It is an essential component in Nvidia's manufacturing stack. Just a little fun fact for you. We're going to make an

episode on that eventually. There's a teaser. What do toilet bowls and AI have in common? Ceramic is one answer. So billions of dollars of market. cap. Yeah, maybe we can do this again. We'll do it for the furthest down the stack. It's amazing how granular it gets. Like when we talk about these two nanometer chips that are coming down the line, the difference is a matter of atoms. Like they are literally constructing on the atomic level, five atoms. And if one of them is in the wrong spot, then these things don't work. So the, the precision required to make this happen is like, it's unbelievable. There's nothing else in the history of humanity that has ever been so complex as this. There's just nothing that humans have made at least.

And going down the stack, it's just unbelievable to see. And the resulting output is you could type into your chat box and you get intelligence out. And it seems imperceivable from magic. So that is the AI energy stack episode. I hope you enjoyed. If you did, please be sure to leave a comment. Share it with a friend who you think might be interested. Be like, yo, bro, you should check out this company.

These guys on Limitless talked about it. And they made a pretty interesting case. Again, none of this is investment advice. I'm not exposed to half these companies for better or worse. But we should start following your own book. Yeah, we probably should. Yeah, we should copy trade our own episodes because like everything that we mentioned has been going nuclear. We'd be way rich.

That's where that's the current state of the union. So if you invest, listen, that's on you. We're too stupid to take our own advice sometimes. But that is the episode. Thank you all so much for watching. Don't forget to share like, subscribe. Leave a comment on something we missed. Is there another one?

Is there another layer of the stack that are missing? Is there another company that is going to go nuclear? that'd be good to know. Just be out of curiosity. But yeah, I think that's it. So we'll see in the next one. See you guys.