Organized & Compiled by: Shenchao TechFlow

Speakers: Josh Kale, Marketing at AnthropicAI; Ejaaz Ahamadeen, Former Product Manager at Coinbase

Podcast Source: Limitless Podcast

Original Title: Leopold Aschenbrenner says "No More Stocks!"

Air Date: June 17, 2026

Key Takeaways

Leopold Aschenbrenner, considered one of the world's most aggressive AI investors, is shorting NVIDIA, ASML, and Oracle in the public markets with a notional position of approximately $90 billion, while rotating capital into deeper AI infrastructure and model assets such as power, memory, data center networking, and Anthropic. The two hosts believe this does not signal that the AI bubble has burst, but rather indicates a rotation signal within infrastructure trading from a "chips first" approach to "energy, networking, and data center construction first." This judgment's market implications are rapidly amplifying, especially after NVIDIA just completed a $25 billion bond financing and Anthropic's valuation was pushed higher.

Key Opinion Highlights

Leopold's Core Trading Logic

"The classic 'selling shovels' trade in AI has become too crowded, and Leopold's recent position changes convey exactly this signal."
"His judgment is not that AI infrastructure has peaked, but that certain layers within the infrastructure stack, especially semiconductors and traditional hot targets, have become overly crowded."
"If the question becomes where capital will rotate next, there are two answers. The first and most direct is flowing to the next real infrastructure bottlenecks, namely power, memory, and data center networking. The second answer is that mysterious investment exposed just a few weeks ago."
"What he bets on has always been very infrastructure-oriented things, investing in these optical companies as well as power-related companies."
"If he is cautious on NVIDIA, capital will go to places like power and memory; at the same time, he also wants to invest directly in the 'mine' itself, rather than continuing to just buy 'shovels,' and Anthropic is the mine he favors most."

Signals from NVIDIA's Financing

"The question isn't whether NVIDIA will continue to make money, but why a company with extremely high profit margins and substantial cash on its books would go out and borrow another $25 billion."
"If a company is simultaneously conducting large-scale stock buybacks, significantly increasing dividends, and borrowing money in the same month, it clearly isn't borrowing because it lacks cash. A more reasonable explanation is that this is cheap capital, and the financing methods in this AI cycle are undergoing a slight shift."

The Next Wave of AI Infrastructure Dividends

"The real bottleneck is no longer just GPUs, but power, memory, data center networking, and the ability to actually build these things out."
"Even if you raise as much money as you want, you can't build data centers fast enough, expand memory chip production capacity enough, or immediately scale up the power grid, transmission lines, and related infrastructure. There aren't enough people on the ground, and approvals, regulations, and various procedures are also holding you back."
"Whoever has the ability to build out the data centers will be the one to take the money."

Optical Modules, Copper, and Fiber Optics

"As GPU scale gets larger, copper wires will get hotter and energy loss will increase, making efficiency very poor. In this scenario, fiber optics become the next upgrade direction."
"For high-bandwidth, short-distance transmission scenarios, copper is almost the only material people really want to use. Only when it starts to become unsuitable, such as when the distance is too long or heat is too great, will they switch to fiber optics. So currently, market demand for the combination of copper and fiber optics is very strong."
"Copper futures have been very strong recently, essentially because everyone needs it. It is the most critical base material for short-distance, high-bandwidth transmission, and fiber optics is the next step."
Copper remains the most critical material for short-distance, high-bandwidth transmission, but once the distance lengthens or heat becomes too high, a shift to fiber optics is necessary."
"Capital will next fall on those infrastructure companies that don't sound sexy."

Why Energy is the Safest Bet

"I have always been bullish on energy, because even if AI demand slows down, energy itself remains a global necessity, and this demand will only get higher."
"The single trend that will continue to rise regardless of the scenario is our demand for energy, electricity, and power. These companies are the targets I am most willing to go long on for the long term."
"What I most want to follow are the companies Jensen is investing in, which also intersect with Leopold's logic. So the target closest to copy-trading for me right now is Marvell."
"The best long-term positions are not necessarily the hottest chip companies, but those power infrastructure assets that are unavoidable in any macro scenario."

Leopold's AI Portfolio

Josh Kale:

Leopold Aschenbrenner, this 24-year-old who specifically invests in AI, is now almost regarded by the market as the world's strongest AI investor. Rumors suggest his fund's notional position size has exceeded $200 billion. When we looked at Ejaaz's post a month ago, the fund size was only $137 billion, meaning it's essentially doubling every quarter.

This time, we've obtained several quite important new changes in his recent investment moves. In the last episode, we discussed his portfolio, and the most surprising point then was that he was actually shorting a company almost everyone is familiar with, namely NVIDIA, the world's highest market cap and hottest AI company. Many people couldn't figure out why he would set up a short position of over $90 billion against such a company.

Now we have a new clue that might explain this. NVIDIA is actually financing, and doing so by issuing bonds. On the surface, this seems very unreasonable. Why would a company as massive and highly profitable as NVIDIA need another $25 billion in cash it just completed raising? Today, we want to combine Leopold's portfolio to discuss why he can make so much money, what he's looking at next, and what NVIDIA's financing really means.

Ejaaz Ahamadeen:

First, some background. Leopold Aschenbrenner was formerly a researcher at OpenAI. He raised a fund about one and a half to two years ago, initially quite small, I recall around $200 million. But judging from his latest 13F filing, the fund's public holdings are already worth $137 billion.

So naturally, the market wants to know what positions he has bet on, what his core investment logic is, and where his next big trade will land. To understand this, you first need to know that before a month ago, Leopold was actually very bullish on the entire AI sector, especially favoring the "selling shovels" logic, meaning GPU and upstream hardware suppliers like NVIDIA.

But about a month ago, the market discovered he wasn't that bullish on the semiconductor line. He remains bullish on real bottleneck areas like memory and power, and possibly on new types of cloud providers, but he specifically is not bullish on the world's most valuable company, NVIDIA. More specifically, he has placed a total of about $90 billion in bearish positions on several companies considered core beneficiaries of AI infrastructure, including NVIDIA, ASML, and Oracle.

The Logic Behind Shorting NVIDIA

Ejaaz Ahamadeen:

Once this news came out, many people started worrying, thinking the AI bubble might be about to burst. After all, on the surface, NVIDIA's GPUs are still selling like hotcakes, and demand hasn't visibly weakened, so where exactly is the problem?

Later, we dug up a few more new clues, the most important of which is that NVIDIA just took $25 billion from external sources through bond financing. This means it's not simply using the money on its own books, but is adding extra leverage. So the question arises: Why would the world's most profitable company, with the highest margins and strongest cash flow, go out and borrow $25 billion?

Josh Kale:

And initially, they only planned to raise $20 billion, but it ended up being expanded to $25 billion, with subscriptions exceeding 3 times. When we discussed this portfolio last time, we said not to worry about a bubble yet, because although these companies have huge capital expenditures, their revenues are also high enough that, theoretically, they could fully support expansion using their own balance sheets.

But this is the first time since 2021 that NVIDIA has clearly financed off-balance-sheet instead of directly using its own cash on hand. I remember it currently has over $120 billion in cash on its books. Putting all this together creates a strange tension: on one hand, Leopold is shorting, and on the other, NVIDIA, seemingly with unlimited cash and profits, is still issuing bonds. So what exactly is happening?

Deconstructing NVIDIA's Bond Financing

Josh Kale: Ejaaz, can you help us break down this transaction itself? Because this isn't financing in the ordinary sense, but a bond issuance. Ultimately, NVIDIA now has another $25 billion on its balance sheet, and the interest rate seems like it should be very low.

Ejaaz Ahamadeen:

I'll lay out both explanations. NVIDIA already had about $13.7 billion in cash on its books, meaning it could have easily spent its own money outright. So why raise external financing? The simplest analogy is buying a house. Many people, even if they have the full amount in cash, still choose to take out a mortgage because their own capital can be deployed elsewhere, and if borrowing costs are low enough, it's actually more cost-effective.

The interest rate environment hasn't been friendly in recent years, but if you're NVIDIA—one of the most valuable and sought-after companies in the world—you can borrow money on very favorable terms. This $25 billion bond offering, with maturities ranging from 2 to 30 years, can almost be considered extremely cheap money, with rates approaching U.S. Treasury yields.

And this financing was reportedly oversubscribed by about 4 times. In other words, there was $85 billion in the market trying to pour into this $25 billion offering. NVIDIA could practically pick and choose its investors. If you go by the official explanation, NVIDIA says this is primarily a financial arrangement to repay and refinance a portion of existing debt. Google did something very similar a few weeks ago, and also back in February. So you can certainly accept that explanation and view it as financial optimization.

But the other side is hard to ignore: Over the past month and a half, NVIDIA, Amazon, Google, and several other hyperscale cloud providers have all been adding external financing. Some are issuing bonds, some are selling stock. Leopold's view might not be entirely without merit—could this be a signal that the bubble is starting to crack, that the house of cards is beginning to wobble? But if you look purely at the financial structure, this doesn't yet clearly point to danger.

Josh Kale:

I see it the same way. A $9 billion short position on NVIDIA—that's a really massive bet. But when we were doing our research, we noticed something else: on May 18, NVIDIA's board just authorized an additional $80 billion buyback and raised the dividend from 1 cent per share to 25 cents, a 25-fold increase.

If a company is simultaneously doing massive stock buybacks, dramatically increasing dividends, and borrowing money in the same month, it's clearly not borrowing because it's short on cash. The more reasonable explanation is that this is cheap capital, and the financing methods in this AI cycle are undergoing a slight shift. Everyone wants to participate in these capital operations, and NVIDIA realizes that raising money by issuing bonds is even cheaper than other financing methods, so it simply went ahead and did it. At least for now, NVIDIA itself is still doing very well.

Why He Adjusted His Positions

Josh Kale: This brings us back to another question. What exactly is Leopold thinking? Why did his judgment change? The stock chart you just showed also illustrates that NVIDIA's recent performance hasn't been particularly strong, but it hasn't been terrible either. It's still the world's largest company with a market cap near $5 trillion, down only 7% in a month. Against the backdrop of other AI stocks surging wildly, that's nothing.

Ejaaz Ahamadeen:

I don't think NVIDIA is going anywhere. Its GPUs, including the CPU product line it just launched a few weeks ago, I think will perform extremely well. Demand for AI products is exponentially oversupplied right now, and the core machine supplier that can truly meet this demand is still primarily NVIDIA.

But I do think that the classic "pick-and-shovel" trade in AI has become too crowded, and Leopold's recent position changes are signaling exactly that. Looking at his latest 13F, his bearish positions are clearly skewed toward the semiconductor line—names like NVIDIA, ASML, Oracle, and several other infrastructure-level companies.

Yet at the same time, he has heavy positions in memory, power, and new types of cloud. This tells us that his judgment isn't that AI infrastructure has peaked, but that certain layers within the infrastructure stack, especially semiconductors and traditional hot names, have become too crowded.

If the question becomes where capital will rotate next, there are two answers. The first is the most direct: it will flow to the next real infrastructure bottleneck—areas like power, memory, and data center networking. The second answer is that mysterious investment that was only exposed a few weeks ago.

The Unexpectedly Revealed Anthropic Position

Josh Kale:

This was the most surprising to me. I only found out about it yesterday after you mentioned it, and my first reaction was that it couldn't be true. Does Leopold's fund "Situational Awareness" really have 20% allocated to Anthropic equity? Rumors now suggest that this company accounts for about one-fifth of Leopold's fund. The Wall Street Journal and several other media outlets have reported this, and sources very close to the deal have confirmed it.

This becomes a card in his portfolio that the market completely didn't anticipate. Because 13F filings only disclose public market holdings, not private equity, and Anthropic is precisely a large block of unlisted equity. That's also why people are starting to understand why his portfolio valuation is being estimated as high as $20 billion.

If 20% of the fund is in Anthropic, and he likely invested around early 2025, then one year's return on Anthropic has been like seven years' worth. This change would significantly revise our understanding of his entire investment portfolio.

Ejaaz Ahamadeen:

Yes. He first invested in Anthropic through private channels or funds around March 2025, when Anthropic was valued at about $60 billion. Now, based on the latest funding round, it's been priced at $965 billion.

That's nearly a 15x increase. According to the calculations we showed in today's episode, the liquid portfolio value disclosed in his latest 13F was $13.7 billion. If you add the Anthropic position as reported by The Wall Street Journal, that could add roughly another $7 billion, bringing the total assets under management to $20 billion.

How staggering is this? Bill Ackman, a top-tier investor who's been in the market for thirty or forty years, has a Pershing Capital of roughly the same size—about $20 billion. Leopold has been in this game for only a year and a half, and he's only 24 years old, with virtually no real investment experience.

But he has made some incredibly stunning calls, and the crazy thing is he actually wrote it all out in advance. When he launched his fund a year and a half ago, he published a 65-page AI treatise called "Situational Awareness," laying out the entire logic almost completely, including how capital would rotate from semiconductors and certain infrastructure segments to other bottleneck constraints. Now the market is developing along those lines—it's truly astonishing.

The Next Wave of Infrastructure Plays

Ejaaz Ahamadeen:

So this also tells me where the money will flow next. If he's cautious on NVIDIA, then capital will go to areas like power and memory. At the same time, he also wants to invest directly in the "mine" itself, rather than continuing to just buy "shovels." Anthropic is his favorite mine.

Josh Kale:

This does look like a new trend, and once again, he's ahead of most people. Over the past 12 months, everyone has been searching for AI's bottlenecks—rare metals, memory, RAM, and so on—and the market has chased all of them. Those calls weren't wrong, because that wave of moves did happen.

But looking at it now, the valuations of those directions that were seen as bottlenecks are gradually becoming more reasonable. People have a better understanding of these companies' business models, market potential, and future revenue, so a lot of the value has already been priced in. What we care more about in the next round is where the subsequent money will continue to flow.

You just mentioned land, power, physical shells, physical infrastructure—that direction looks right. Because if we think about what truly matters most for AI, the answer increasingly looks like physical construction capability. Look at xAI, or more precisely, look at SpaceX, which is now publicly traded. Its core revenue isn't the rockets themselves, but AI infrastructure construction.

And look at its recent deals with Anthropic and Google—the value generated from those has already surpassed the combined value of Starlink, Starship, and the entire satellite business. There's clearly enormous demand and enormous value here. So the question becomes: who can actually build these things?

SpaceX is clearly one answer. After hours last night, its stock price reached $230, corresponding to a valuation of about $3.1 trillion. We'll do a dedicated episode on SpaceX this week, because this run has been so extreme—it just completed the acquisition of Cursor, and now its valuation has reached $3 trillion. Elon made more money in a single day than Warren Buffett has earned in his entire career.

Who Will Capture the Next Wave of Gains

Josh Kale: What we care about is which companies are best at doing this kind of hardware infrastructure, best at developing the "machines that make the machines." Combining Leopold's direction and the broader trend, we think capital will flow here next. So Ejaaz, in reality, which companies will this rotation actually land on?

Ejaaz Ahamadeen:

Many will be those infrastructure companies that don't sound particularly sexy. One name that's been frequently mentioned in the past month is Marvell. A few weeks ago at the Computex conference in Taiwan, Jensen Huang said directly on stage that this would be the next trillion-dollar company.

And just three months before he made that statement, NVIDIA had invested $1.5 billion in Marvell. I'm already a bit unsure whether this counts as insider trading or market manipulation, because after he said that, the stock rose another 70%.

I think it's actually easy to jump to the conclusion that AI infrastructure has peaked right now. But if you compare it to historical financial crises, like 2008, that flavor of high leverage, financial engineering, and systemic manipulation hasn't fully emerged in this cycle yet.

There are two key differences. First, the products these companies are making today—people are actually paying real money for them. Neither the dot-com bubble nor the financial crisis had such solid real demand. Second, constrained by the laws of physics, we actually can't infinitely leverage up right now, because the entire system is bottlenecked by manpower and construction capacity.

Even if you raise as much money as you want, you can't build data centers fast enough, can't expand memory chip production capacity fast enough, and can't immediately scale up power grids, transmission lines, and related infrastructure. There simply aren't enough people on the ground, and approvals, regulations, and various procedures are also holding things back.

So I actually think this gives investors an advantage. Since you already know the most popular chip and pick-and-shovel trades are overcrowded, the money will next flow into power, data networks, companies like Astera Labs, and then into other related links. What you really need to think about is when these contracts start to materialize, when these fabs actually get built, when SpaceX rockets can send AI satellites into space, and even when we can start using solar energy to train AI models.

The timeline determines the rhythm of placing bets. At least that’s the framework I use for investing, though of course this is not investment advice. The reason I see it this way is that over the past year and a half, we’ve already witnessed how money has flowed from broad AI stocks into semiconductor and infrastructure trades.

Josh Kale:

If you continue looking down this portfolio chart, you’ll actually find that this story is clearly written into his position structure. By category, what is his largest allocation? It’s power and energy. Next is memory, followed by cloud and GPU miners — the most tangible infrastructure.

He wants to hold next-generation cloud providers like CoreWeave, and also miners that have pivoted to cloud computing.What he wants to own is this physical infrastructure, because he believes this is where the real bottleneck lies.As you just mentioned, there are certainly many more granular links in the chain — actual construction, hardware manufacturing, and data center construction itself are all extremely difficult.

If you ask where the biggest bottleneck is, it might even be permits and approvals. So who is solving these problems? SpaceX wants to move data centers into space, and Tesla wants to use humanoid robots to solve labor issues. But both of those are still far off. In the short to medium term, there are actually plenty of gap opportunities, and that’s exactly the direction Leopold is betting on.

The Advantage of Optical Modules and Fiber

Josh Kale: I want to add one detail we haven’t expanded on yet. For those who want to dig deeper and find more excess returns, many of his clues are actually hidden in optics and the deeper layers of the tech stack. Ejaaz, you’ve been researching this recently — can you explain his thinking?

Ejaaz Ahamadeen:

If you look at the positions on his screen, CoreWeave and Iron are basically the top-tier next-generation cloud service providers. Simply put, they’re somewhat like Amazon Web Services, except AWS provides cloud services to internet companies, while these companies provide ready-made GPU infrastructure to AI companies.

They set up everything — GPU, networking, deployment — so AI companies don’t have to worry about the underlying infrastructure and can directly train models and access compute power. CoreWeave and Iron have been among his largest concentrated positions since he initiated them, and they’ve also delivered the highest returns.

And notably, he still keeps these two companies among his largest positions today. This also indicates one thing: in his view, this trade is far from over. Going further, he has also privately invested in Core Scientific, a company that can help unlock CoreWeave’s infrastructure supply capacity. In a sense, he has added another layer of leverage on CoreWeave.

Beyond these, look at companies like Coherent and Lumentum — they are essentially suppliers related to fiber optics and optical connections. To explain it in the simplest terms, semiconductors and GPUs need to communicate with each other, and traditionally this relies heavily on copper wiring. The problem is,as GPU clusters get larger, copper gets hotter, energy loss increases, and efficiency becomes very poor, and that’s when fiber optics becomes the next upgrade direction.It can complete data transmission faster, with better cost efficiency, and allows companies providing inference and training compute to earn more money. So you’ll find thatwhat he’s betting on has always been very infrastructure-oriented — investing in these optical companies as well as power-related companies.It may not sound sexy, but in my view, this is where the money is actually flowing now.

Josh Kale:

The copper angle is also very interesting to me, because I only recently realized how critical it is in short-distance data transmission.In many high-bandwidth, short-distance transmission scenarios, copper is almost the only material people really want to use. Only when it starts to become unsuitable — for example, when the distance is too long or heat is too high — do they switch to fiber. So there is very strong market demand for the combination of copper and fiber right now,which is also why observing the copper trade is so interesting.Copper futures have been very strong recently, essentially because everyone needs it. It’s the most critical base material for short-distance, high-bandwidth transmission, and fiber is the next step.

Thinking from an even more fundamental level, the materials thread has always been very interesting. Beneath all the underlying layers, at the very bottom, it’s really about what core raw materials are needed to achieve intelligence. Copper is one, lithium is another, and there are many more. We really should do a dedicated episode on materials. Maybe Leopold hasn’t reached that layer yet, and we might actually be able to see the next rotation one step ahead.

Josh Kale:

If you keep going down the stack, you could even go directly to a copper mine to see how these things are made. But returning to the core judgment, I think the next rotation is indeed moving from those seemingly smaller bottlenecks toward the truly difficult things — namely hardware and large-scale data center construction.

Whoever has the ability to build out data centers will take the money.We’ve already seen how much money SpaceX has made because of how strong data center demand is. Whoever can bring more data centers online faster, and whoever can provide sufficient power and GPUs, will earn the most money. This is basically the direction Leopold is betting on now.

Is a Bubble Emerging?

Josh Kale: In summary, we don’t think we’ve entered a bubble-bursting phase yet. Leopold’s positions look more like a rotation rather than a full retreat. So, should we still follow his lead?

Ejaaz Ahamadeen:

I admit, when I first saw his 13F, my initial reaction was that this guy is actually bearish on the world’s most valuable company, one with demand backlogged all the way to 2029 — that’s insane. But now, seeing this financing, I’m starting to think that if NVIDIA continues to take on external debt in the future, or even potentially sells shares down the road, and if this trend continues, then Leopold might once again be proven right.

If that’s the case, this fund of his might end up surpassing the world’s top traders and the best investment funds. He really keeps winning, and it’s hard not to respect that.

Josh Kale:

But there’s another important point. For most of his life, he has almost always just been long and has never really been tested by a large-scale sell-off. We mentioned Bill Ackman earlier — generating a 30x return and surviving in the market for 30 years are two different things.

If he can truly sustain this kind of growth and also learn when to hit the sell button, how to manage risk, and how to protect himself with hedges, that would be even more formidable. We’re already starting to see the early signs of that ability. That $9 billion short position wasn’t actually a direct $9 billion cash short; it was achieved through options and leverage, not a one-to-one naked short. Either way, this is something well worth continuing to watch.

Energy is the Core Bet

Josh Kale: If you had to pick one stock from his entire portfolio that you’d most want to buy yourself, which would it be?

My own answer is energy stocks. I’ve always been bullish on energy, because even if AI demand slows down, energy itself remains a global necessity, and that demand will only increase. Even if you completely ignore AI, we still need more energy and more electricity. Companies like Bloom Energy, which can enhance power supply and transmission capacity, are the ones I’m most excited about, because they most resemble a hedge-style bet. The one single trend that will continue to rise regardless of the scenario is our demand for energy, electricity, and power, and these companies are the ones I’m most willing to be long on for the long term.

Ejaaz Ahamadeen:

My answer is a bit of a cheat.What I most want to follow are the companies that Jensen is investing in, which also intersect with Leopold’s logic. The stock I’m closest to mirroring right now is Marvell. Although it’s not a publicly disclosed holding of Leopold’s, it fits very well with his bet on fiber optics and power, and Jensen has already put $1.5 billion of real money into it.

I’ve observed a phenomenon: whenever Jensen invests in a company through NVIDIA, whether it’s Intel, CoreWeave, or others, it basically keeps going up afterward. So my position is roughly around here right now. I also hold some CoreWeave myself, because both Jensen and Leopold are extremely bullish on it.

Josh Kale:

Marvell has already risen 270% in the past 6 months. This might really be a good rule of thumb: when people like Jensen, or even figures with enormous influence like Trump, publicly say to buy a certain stock, many times you probably should take a serious look.

It has been proven multiple times in the past that such signals often have a very large realization potential. Whether it’s Intel or Marvell, these cases all show that, on one hand, they truly understand what they’re talking about, and on the other hand, they also have the ability to influence the outcomes for these companies. So this market move has been truly insane.

I hope it continues. From the current look of things, it very likely will. At least for now, we are all still leaning bullish, still optimistic, and will continue to make judgments day by day as things change.

Josh Kale: Do you have any final thoughts to add about Leopold’s portfolio update?

Ejaaz Ahamadeen:

I’d actually really like to hear what the skeptics think. If after listening to our analysis just now, you think we’re completely wrong, or that we’ve misunderstood something, please feel free to point it out directly.

Yesterday I stared at that news about NVIDIA’s $25 billion financing for a long time, originally intending to find fault with it. But if you look purely at the financial logic, it actually does make sense. Why not borrow this kind of almost risk-free cheap money? Using other people’s money to expand is clearly more rational than selling your own equity, because that way you can retain more of the future upside.

DeepSeek is recognized by the top-level domain ai.com

Popular Articles