Leopold Aschenbrenner Says No More Stocks
read summary →Leopold Aschenbrenner says “No More Stocks!”
Channel: Limitless Podcast URL: https://youtu.be/1uk8eZwOdhI Published: 2026-06-17
Leopold Ashenbrer, that 24 year old guy who invests in AI, he’s very clearly the best AI investor in the world. There is word on the street that his notional position in this fund is now over $20 billion. And EJ, we were looking at one of your posts from what a month ago and it was 13. So the fund seems to be doubling basically every quarter. And we have some pretty serious and interesting updates on what Liupold has been investing in. You’ll note that in the last episode we covered his portfolio. He was actually short a company that a lot of people are familiar with, the largest one in the world, Nvidia. And nobody could figure out why. He had over $9 billion worth of shorts on the largest and hottest AI company in the world. Today, we have some news that might actually uncover the reason why. It comes in the form of a debt offering. Nvidia is actually raising money, which doesn’t seem to make sense on the surface. Why would a company as big as Nvidia be raising what just closed as $25 billion in cash when their margins in particular are so so high? So on the episode today, we’re going to talk about Liupole’s portfolio, how he’s doing so well, what he’s looking forward to, and what he’s positioning himself to to go next, but also what’s going on with Nvidia. So Leopold Ashen Brener for context, exopai researcher, raised a fund about 1 and a half to two years ago. It was a small casual raise of $200 million, I believe. And since then, since his last uh 13F filings, the fund is worth $13.7 billion. And so, everyone obviously wanted to know what positions he was taking, what his thesis was, what was the next major trade. Now, it’s important to understand up until a month ago, Leopold was extremely bullish about everything AI, particularly picks and shovels. So, we’re talking about the Nvidas, the GPU makers, all that kind of stuff. And then a month ago, it was revealed um that he wasn’t especially bullish on the semiconductor landscape. He was still bullish on things like memory and power, those kinds of bottleneck constraints. Uh and maybe he was uh bullish NeoClouds as well, but he wasn’t bullish the most valuable company in the world, Nvidia. And we have a total of $9 billion worth of puts in Nvidia, ASML, and Oracle. So these are companies that are key to the infrastructure boom, which is what has been heralded as the best and most certain trade in AI. So people started getting concerned. They were like is this the AI bubble popping like we don’t see any signs of this. Nvidia is still selling so many GPUs. What could be the problem? Um and since then we’ve uncovered uh a few things. The major one being that uh Nvidia just raised $25 billion in a bond offering from external capital. So what this means effectively is they’re raising money outside of their balance sheet. So it causes the question which is why on earth is the most valuable company in the world that has the most money that’s making the most money that has the highest margins raising $25 billion. In fact, they weren’t only going to plan to raise they were only going to plan to raise 20 billion and they actually wound up raising 25 billion which was more than three times overs subscribed. And it’s funny in the last episode that we were talking about this portfolio. Are we in a bubble? How do we know when the top is getting closer to being in? We’re like don’t worry about it. All these companies are spending a tremendous amount of capex, but they have a huge amount of revenue to fund all that. It’s coming right off the balance sheet. This is the first time since 2021, in the case of Nvidia, where they’re actually raising money outside of the balance sheet. They’re not taking their money on the balance sheet, which I believe is about $12 billion they currently have. And that leads me to a few questions. It’s like Leopold is short. Nvidia is raising some debt when it seems like they have infinite cash and infinite margins. What’s actually going on here? So, maybe you could help us unpack what this deal actually was. This is a bond offering which is not just a general fund raise. It’s a little bit different and at the end of the day now Nvidia now has 25 billion additional dollars on their balance sheet for what I assume is a pretty low rate.
Mhm. So let me give you both sides of the story here. Um Nvidia has around $13.7 billion on their cash balance sheet. So this is money that they could just spend to do whatever they want. So the question then is why are they raising external capital? Well, the analogy is think of you purchasing a house. Typically, most people would, if they have the money, they would still take a mortgage out. Why? Because you can, you know, use your capital for other purposes and you can just borrow money at a really cheap rate. Now, interest rates have been pretty uh cataclysmic for a while now. But if you’re Nvidia, the most valuable company in the world that has the most valuable stock in the world, the most desirable stock in the world, hey, you could raise at pretty decent rates. And I’ve got the breakdown over here for the $25 billion bond offering. Um they’ve got uh bonds that range between 2 to 30 years. So it’s effectively free money that they’re raising at. The interest rates are almost nothing. It’s uh almost as good as government yield bond rates uh itself, which is like the best kind of offering that you could potentially get. Um and they were 4x overs subscribed for people who wanted to help invest in this money. So, $85 billion worth of capital wanted to kind of pummel into this $25 billion race. So, Nvidia could just have their pick uh of the litter. Now, if you want to look at the reasons behind why they’re doing this, um Nvidia’s stated claim officially is this is just financial bookkeeping. We want to pay off and refinance some debt that we have on our balance sheet sheet, which sounds familiar because Google did something very similar about uh 3 weeks ago and then they did it earlier on this year in February. So you could take the stated claim that this is financial bookkeeping, but the other side of it is does it seem pretty coincidental that Nvidia, Amazon, Google, and I think three of the other hyperscalers have all raised external debt financing in the last month and a half. And it’s a mixture of equity selling, which is what kind of like Google’s done just last week or 3 weeks ago, and it’s a mixture of bond raises as well. So Liopold could be right in the sense that this might be marking the start of a bubble popping or coming down. The house of cards are coming down if this is the sign of a levered bet or a levered race. But right now, if you look at the financial structuring of this entire thing, it doesn’t entirely say that. That’s what I’m thinking too is I’m looking at Leopold’s portfolio. I’m like, man, $9 billion is a lot of money to be short Nvidia. And another thing that I found out as we were researching is on May 18th, the board of Nvidia actually authorized an additional $80 billion in buybacks and raised their dividend from a penny to 25 cents per share. So they multiply the dividend by 25 times. They pledged $80 billion to buy back shares from shareholders. And a company doesn’t really hand $80 billion back to shareholders and raise their dividend 25x in the same month that it borrows out of need. So there are these two conflicting things. It’s like, well, they’re clearly not borrowing out of need because they’re actually giving back to the shareholders. So, the question, why? Why are they doing this? Well, because it’s cheap money. And it seems like the way that this AI pump, this AI bubble is being funded, is just slightly shifting to a new model because everyone wants to be participating. Everyone wants to be involved in these capital raises. Nvidia has recognized this. They could do so even more cheaply by issuing their own bonds than they can from going anywhere else. So, they’re like, “Hey, if we could take some money off the table, we’re going to do that.” And it seems like Nvidia is actually just doing just fine. Which leads us to the question of like what is Leopold actually thinking? How is his mind changed over time? I mean, you looked at the Nvidia stock. It hasn’t been doing great. The one you were just showing on screen. It’s not like Nvidia’s been crushing recently, but it’s still like $5 trillion, most valuable company in the world. I mean, only down 7% in a month is is no worthy considering everything else has been pumping, but it’s still like I don’t know. It it doesn’t seem like it’s doing that bad. So, so listen, I think that Nvidia is not going anywhere. I think their products, their GPUs, and now the CPU line that they just opened up a few weeks ago is going to do overwhelmingly well. There is a surplus, an exponential surplus of demand for AI products, and there’s only one main machine provider that can uh support that type of demand, and that is Nvidia. So, I don’t think they’re going anywhere. What I do think is the picket and shovel AI trade is extremely overcrowded and that’s what Leopold’s most recent portfolio positioning tells us like I mean let’s let’s let’s take a look at this right so as of their last 13F filing they were pretty bearish on the semiconductor landscape uh through the form of a short position or put position in Nvidia ASML Oracle and a few other infrastructure scale providers but at the same time Leopold was very long companies in the memory sector or in the power sector or in the Neocloud sector. Iron is one of his biggest positions. Um, and that since his 13F filing is up 20%. So like you know again he’s hitting some nails on the head but he’s bearish some layers of the semiconductor infrastructure stack specifically and that tells me one very clear thing which is Leopold isn’t bearish um AI infrastructure. I don’t think he’s calling it a top on the bubble. he just thinks that the trade itself on picks and shovel is overcrowded and that money is going to rotate somewhere else. So if the question is where is that money rotating? Um there’s two answers. The most obvious answer is it’s going to the next infrastructure bottleneck which we’ve spoken about on previous episodes is power, it’s memory, it’s data center networking, it’s stuff like that. We’ve spoken about this on previous episodes. Um and then there’s this other secret investment that was uncovered a few weeks ago, right? Yeah, this is the weird one for me. I was like, wait, this came out of left field. AJ, you actually told me this yesterday. And I was like, no way that’s true. There’s surely no chance that Leopold’s fund situational awareness, 20% of the fund owns anthropic equity. Yes, the AI company that everyone knows and loves, Anthropic. 20% of Leopold Dash and Berners fund is involved in anthropic. How do we know this? Well, it’s not totally confirmed, but the Wall Street Journal and a few other publications have apparently sources that are very close to the matter that confirmed this. And it seems as if this is a huge wild card that we were not anticipating in his fund because when you publish these 13F filings, the way it works is you just publish your public positions. It doesn’t have anything to do with the privately held equity. Enthropic turns out is a huge amount of privately held equity. And that’s where we get to this 20 billion dollar expected valuation of his portfolios because I mean 20% in the fund being in anthropic and he’s been in it what EJ for over a year now like he invested early 2025 that’s like being invested in anthropic is like dog years each year is like seven. So a year in anthropic feels like a tremendous return and that that’s a huge readjustment to his portfolio when we see this because now it feels like we have a much more clear picture. Yeah. So the first time he invested either privately or through the fund in a VC investment in Anthropic was in March 2025. That was when Anthropic was worth $60 billion double digits. Now of Anthropic’s recent valuation, they are valued at $965 billion. That is a 15x markup. And if you do the math, uh, as we’re showing on the screen today, his liquid portfolio, so that’s the thing that’s reported in the 13F as of his recent 13F filings, was worth $13.7 billion with the anthropic stake that he supposedly has, as reported by the Wall Street Journal, adds another $7 billion, bringing him to a total AUM of $20 billion. Now, if you’re trying to understand how crazy this is, Bill Aman, one of the most famed investors in the entire world, who’s been in this game for 30 to 40 years, his fund is worth just around $20 billion. Persing Capital, Leopold’s been in the game for about a year and a half. And everyone’s and this guy is 24 years old. He has zero and I must emphasize zero investing experience, right? Um, but he’s made some of the most amazing calls ever. And the craziest part about all of this, and I’ve said this on previous episodes, is that he gave us the entire thesis in a 65page uh AI essay uh about a year and a half ago when he started his fund, situational awareness called situational awareness that broke down his entire trade. He even described how the money will rotate from semiconductors and infrastructure specifically into other bottleneck constraints and that’s the trade that’s playing out today. So it’s just extremely impressive and that tells me where the uh the rest of the money is flowing. If he’s bearish Nvidia money’s going to power memory, but also he wants to invest in the mines themselves. Screw the picks and shovels. Invest in the mines and Anthropic is his favorite bet. This seems like the trend and I think this is probably a little bit ahead of his time again as always, but it seems like the new trend is pivoting away from the bottleneck thesis. Over the last what 12 months or so, it’s been everyone’s trying to find the bottlenecks. Where are they? It’s in the precious metals. It’s in the memory. It’s in RAM. It’s in all these these things that people perceive to be bottlenecks. And it’s true. And that run has happened, but it seems as if those are getting close to more fairly priced. Everyone kind of understands the business. They understand the market. They understand the revenues that are projected forward. and they kind of have a reasonable valuation. So a lot of that value has been captured. The next rotation is what we’re interested in. It’s where does the money flow after this as you mentioned land power shell kind of the physical infrastructure. This seems to be directionally correct. When we think about the companies that are most important moving forward when you think about the things most critical to AI it’s the actual physical buildout. It’s it’s what XAI has been doing. When you look at XAI the company or SpaceX I should say which is now publicly traded. Where does all of the revenue come from? It has absolutely nothing to do with rocket ships. It’s the AI buildout. Look at the deal they signed with Anthropic. Look at the deal they signed with Google. Those account for more than Starlink and Starship and their entire satellite business combined. There’s clearly a tremendous amount of value, a huge amount of demand. And then the question now shifts to who are the people that are building this? Okay, SpaceX is the obvious answer. SpaceX, I mean, after hours last night traded at $230 a share. That’s $3.1 trillion in value. And we’re going to have an entire episode about SpaceX this week because my god, what a run. They just closed the Cursor acquisition. They’re now valued at $3 trillion. Elon made more money in a day than Warren Buffett did in his whole career. So that is a whole separate thing. But what are they best at? They’re best at that hardware infrastructure at developing the machines that build the machines. And that is where we think based on Leopold’s direction, based on just general trends, that’s where the money heads to. So what does it actually look like EJ physically manifest? Who are the companies involved? Who’s there at the rotation ready to just receive this hot ball of money that’s flowing around? Yeah. So it’s uh a lot of the unsexy infrastructure companies as you mentioned. Um one uh popular name that’s been passed around uh over the last month is uh Marvel. Marvel is the company that Jensen Huang went on stage at Computex, which is like a major uh AI conference 2 weeks ago in Taiwan and he said this is the next trillion dollar company. Um 3 months prior to him making that statement, Nvidia had made a $1.5 billion investment in Marvel. I don’t know where the lines are in terms of like insider trading or market manipulation at this point, but the stock did pump 70% subsequently after he made that statement. I think it’s easy to call a top on uh AI infrastructure right now. Um and if you think about uh previous financial uh crisis like in 2008, there was a lot of levered positions. There was a lot of financial and market manipulation. We don’t quite see the same here yet. And I’ll tell you what the distinct differences are. Number one, there are people paying for the products that these companies are creating. So there are real customers buying the products. Back in the dotcom boom, back in the financial crisis, you don’t really have this. Um, and then the second thing is we physically by the laws of physics can’t lever up right now because we’re constrained by human capital. What I mean by that is no matter how much money you raise or throw at the thing, you can’t build data centers fast enough. You can’t scale memory chips fast enough. You can’t build the power lines and scale the energy grid and the infrastructure fast enough. We don’t have enough men on the ground. We don’t have the capacity to do that. There are laws, regulations, and a ton of red tape that is currently blocking you. So, what you have here is an advantage in my opinion, which is you can view where the money is going to go, right? So you think that the picks and shovels trade is overtraded or it’s overcrowded. That money is going to flow into power. It’s going to flow into data networking like Aster Labs. It’s going to flow into a bunch of these different companies. And so you need to start thinking, hm, okay, when will these contracts start to be materialized? When will these chip fabs end up being created? When will SpaceX’s rockets start being launched into space, releasing his AI1 satellites, and harvesting the sun’s energy to train AI models? What is that timeline? And then my my the bet I’m taking at least is like I’m investing accordingly based on that. Again, not investment advice, but I just it’s how I generally see the money flowing because it’s how we’ve seen it flow from uh general AI stocks into the semiconductor infrastructure trades which we’ve seen over the last year and a half. And if we scroll down a little bit in this artifact, we we see that story playing out in his portfolio where it shows the ownership stake that he has on a relative basis. The number one category, what is it? It’s power. It’s energy. Second to that, it’s the memory. And then it’s the the clouds and the GPU miners. It’s the actual physical infrastructure. He wants to own Corewave, the NeoClouds. He wants to own the the miners who have switched over to clouds. He wants to just own that physical info cuz that’s where that the true bottleneck is. And so sure, there’s smaller parts along the way, but like you mentioned, the the actual buildout, the actual hardware, the physical manufacturing of these data centers is such a challenging thing and it’s it’s the biggest bottleneck if anything just to get the permits to go build it. Who’s solving this problem? Well, SpaceX is trying to put them into outer space. Who’s solving the human problem? Well, Tesla’s trying to build humanoid robots. But both of those things are a pretty long way out. So, in the intermediary, there’s a lot of opportunity in whites space, and that’s what he’s going for. So, as we start to wrap up this episode there, there’s like one last thing I think we we probably want to mention, which is the the subtle nuances that exist in his portfolio that I think are a little surprising, and we didn’t highlight before. For anyone who’s looking to go a little bit deeper, a little more alpha, a lot of it comes in optical and like kind of lower down the stack technology. EJ, I know you’ve been in the weeds about this. How does it work? What is his position? What is his thinking around I guess optics in general? Well, if you just generally look at his portfolio position that we’re showing on the screen here, Coreweave and Iron uh are basically some of the top neocloud providers. So what I mean by that is think of uh Amazon Web Services like they create cloud services for any internet company that wants to create a thing. That’s what these companies do for any AI company. They have the GPU infrastructure. They set it up for you. They do all the networking such that you don’t need to focus on it. You can just train your models, get access to compute, don’t worry about any of the other um kind of stuff. Core and iron have been Leopold’s biggest positions since he started the fund or he’s had very concentrated positions and they have provided him the biggest returns. Now, I think it’s worth pointing out that he still has the biggest positions focused in these two companies. In fact, he’s privately invested in a company called Core Scientific, which helps unlock a lot of Core Weave’s um infrastructure provision, which means that he’s effectively taken a lever bet on Core Weeb. And the fact that he still has major positions in these companies tells us that that trade in his opinion typically isn’t over yet. And then if you look at Coores and Light, these are all optical fiber providers. Now if you want to understand what that means at a very basic level typically if you need semiconductor GPU chips to speak to each other requires like a lot of copper wires and at some point when you have so many of these GPUs copper wires get overheated there’s a lot of heat resistance you lose a lot of energy it becomes extremely inefficient you know what isn’t as inefficient if you start using optical fibers and that is kind of like the next leap the next iteration to build data centers to network and connect these GPUs which will allow much quicker data transfer that’s much cheaper, much more cost efficient and you can end up making more money on the inference or the training compute that you provide. That’s another it’s very infrastructure heavy is the point that I’m making. So he’s um invested in those companies and invested in the at the power level as well. And I think you know it’s very unsexy to say and talk about but that is inevitably where the money is flowing right now in my opinion. Yeah, the copper thing was interesting for me because I learned how critical copper has been recently for a lot of the close data transfer. Copper is the only material that anybody wants. And the only time you stop using copper is when it becomes unviable. When you have to transfer data over too long of a distance or like you mentioned, it gets too hot. That’s when you switch over to fiber. But the combination of these two seems to be all that anybody wants. So, it’s been interesting watching the copper trade as well. Even though Leopold’s not really invested in it, oh yeah, you got the copper futures pulled up. Copper, if you go to that one-year chart, it’s probably looking pretty strong. Yeah, it’s a very strong commodity here because everyone needs it. that is the the single most important critical element when transferring high bandwidth data over a very short distance. But fiber is the next one. So that’s been an interesting thing to watch play out. I think the materials play is always interesting. That is the the base of all base layers is like what is the core material that goes in in order to get intelligence out. Copper’s one of them. Of course, lithium. There’s a whole bunch of them that I think we need to do for that material episode. Josh, I I think we need to do this. Um we should go down there cuz you know what? Leopold hasn’t made his way down there yet. And maybe we could frontr run him for the next rotational thesis. We could go all the way down to the bottom of the stack. We could go visit the copper mines, see how they’re making all this stuff happen. But I think that is generally speaking the next rotation is from these perceived smaller bottlenecks into the the real hard thing, which is the hardware and building out these large data centers. And whoever’s capable of building out the data centers will collect all the money. We just saw what happened with SpaceX, how much money they were able to make because of how in demand these data centers are to have them online. And anyone who’s capable of spinning them up, of producing enough of producing power GPUs, if you could create GPUs with power, you are going to make all the money. And that’s kind of where Leopold’s at. I guess in summary, we’re not in a bubble. Leopold has rotated. Should probably still copy trade. Is that right? Does that sound okay? Yeah. Um, I I’ll admit when I read his 13F, I was like, you’re bearish the most valuable company in the world that has like demand projected into 2029. Like that is the worst take ever. And now I see this raise and I’m like, wow, if uh Nvidia continues to raise external debt, sell equity potentially in the future if this trend continues. Lioo might be right again and his uh his portfolio AM will eclipse the best traders in the the best investment funds in the world. Um, I don’t know. the guy keeps winning and it’s it’s very impressive to see. Yeah. So, I mean, here’s the thing is he he’s never needed to sell anything. Dude’s just been long only his whole life. So, we’ll see how like you talk about we talked about Bill Aman earlier in this episode and dude’s been around for 30 years and turns out being around for 30 years is actually harder than making like a 30x. If you can actually sustain this growth, if you can learn how to sell the push the sell button, learn how to kind of hedge against risk. We’re starting to see that. I mean those $9 billion in shorts, he’s not actually exposed um with $9 billion of cash sitting there short. That is through options. That is through some leverage. So it’s not a direct onetoone short. But it’s fascinating to watch. We will see. I am going to continue copy trading the boy Liupole because dude has just been correct. He is not wrong. So of his entire portfolio, Josh, what is the one stock that you’d be buying from him? Uh I like the energy stocks. I’m a big energy guy. I think no matter what, even in the case that AI demand slows down, energy is still very much in high demand and the the just demands of it from the general world are going up only. So even in the absence of AI, we need more energy. We need more electricity. Bloom Energy, a lot of companies that are building electricity throughput, I think are the ones that I’m most excited about because they’re just they seem to be the most hedged. is like what is the singular trend that is not going to stop no matter what is our demand for energy, electricity and power and companies like Bloom Energy deliver on that. So that’s the companies that I’m most excited about being long in. When I invest it’s it’s on a very long time frame. I don’t really like to trade much so that’s why and that’s kind of how I think of things. What about you? Do you have any favorites? I uh uh this is a cheat answer but it’s wherever the intersection of what Jensen’s investing in and whatever Leopold is investing in. So, the company that I’m like kind of copy trading is uh Marvel, which isn’t something that Leopold currently owns, but it does sit very firmly in his optical fiber and power investment bet. And it’s the one that Jensen put $1.5 billion behind. I’ve just noted that whatever Jensen invests in via Nvidia, whether it’s Intel, Corewave, or whatever it is, has just gone up only. Um, and so that’s currently where I’m at right now. I own some Core Wee as well, which both Jensen and Leopold have been extremely bullish on. So, dude, Marvel up 270% in 6 months. I think this is a good pro tip. When people like Jensen, when people like Trump, they say go buy this stock, you should probably go buy it. It’s paid off in pretty big ways. Like, Intel is up how many multiples from the time he bought it. Marvel is up how many multiples from Jensen. It’s like these guys know what they’re talking about. Or at least they have the ability to influence the outcomes of these companies. And in that sense, it has been a very wild ride. I hope it continues. I hope this keeps going. It seems like it’s probable to keep going. I think we’re all still long. We’re all still pretty optimistic and excited and we’ll just continue to evaluate day by day. Coming up next is a SpaceX episode which I think we’re very excited about. SpaceX has been the biggest IPO ever. They just closed their deal with Cursor and there are a lot of new updates and interesting takes I believe on how to navigate this IPO, how the unlock schedule is going to work, where all the shares are and and how this is going to play out over the next 6 months. I think is something a lot of people may be interested. So, that one’s coming up next. Any final thoughts before we wrap up here on Leopold portfolio update? I’m curious about the skeptics that are listening to this that think that everything we’ve just said is incorrect and might potentially be wrong. Um, tell us where we’re wrong. Like I’m I’m genuinely curious. Like I I stared at the $25 billion Nvidia raise news yesterday and I was like I want to hate on this but look at it looking at it financially this seems like a sound thing. Like why wouldn’t you raise debt like risk-free money? Like that makes complete sense. Like borrow other people’s money and spend that versus like selling your own equity. Obviously that makes sense. Keep more equity, earn more money in the future. But if there’s something that we’re missing, if there’s a pattern that you’re seeing that we aren’t seeing that we haven’t spoken about, uh leave us a comment, DM us, do whatever you can to get in contact with us. Um and then if you’re listening to this as well, and if you enjoyed this episode, if you’re one of the majority that enjoyed this episode, um please make sure you’re subscribed to us. Turn on notifications if it’s on YouTube. It helps us out massively. Do the same on Spotify. Do the same on Apple Music or wherever you’re listening to us. Uh we also have a newsletter that goes out twice a week. We put out an essay that’s the first essay of this week is coming out uh tomorrow. And then we’re going to have the uh five weekly highlights that we always do at the end of the week. We are doing anything and everything to keep you up to date with uh everything that’s going on with AI. And with that, I believe the episode’s over. It’s done. That’s it. Well, thank you all so much for watching. And yeah, let us know. Share with your friends if you enjoyed this episode. Please leave a comment and um yeah, I think that’s everything. So, thank you so much for watching as always and we will see you in the next one. See you guys.