Reviewing the U.S. stock version of the child: Unpacking Leopold Aschenbrenner's holding logic

Original Video Title: Forget NVIDIA | This 24-Year-Old’s $4.5B Bet on AI’s Real Problem (Leopold Aschenbrenner)
Original Video Source: Limitless Podcast
Original Text Compilation: Deep潮 TechFlow

Key Takeaways

Recently, everyone has been talking about Leopold Aschenbrenner—24 years old, $5.5 billion AI hedge fund, the American stock market version of a prodigy. But most discussions stay at the level of “he’s amazing” and “he made a lot of money,” with few deep dives into his holding logic.

Two months ago, Limitless Podcast did an episode analyzing his 13F report item by item:

Why he sold out of NVIDIA, why he allocated 20% of his portfolio into a fuel cell company, why he bought a bunch of Bitcoin miners, why he shorted Infosys. At that time, the episode hardly drew any attention. Looking back now, many of those judgments have been validated, worth revisiting.

Summary of Notable Insights

On Leopold Aschenbrenner’s Investment Performance

· “Last year, he managed $1 billion… today, just one year later, that $1 billion has grown to $5.5 billion.”

· “His fund was established at the end of 2024 with an initial size of $255 million. In just six months, his fund outperformed the S&P 500 by 8 times.”

· “He wrote a 165-page long article titled ‘Situational Awareness.’ In it, he basically predicts that we will reach Artificial General Intelligence (AGI) by 2027.”

Shift in Investment Paradigm: From Chips to Infrastructure

· “He sold off NVIDIA, Broadcom, TSMC, Micron. These are all major AI infrastructure companies.”

· “By the end of 2025 or early 2026, he believes the market will have fully priced in the value of GPUs.”

· “He shifted his focus to the main bottlenecks that investors have yet to fully address—energy and infrastructure.”

· “The existing power grid is designed for humans, not to meet the enormous AI demands we face today. That’s where his current investments are focused.”

Core Holding: Bloom Energy

· “Bloom Energy is his largest current investment, accounting for 20% of his entire portfolio… He has built a huge position in this company, amounting to $855 million.”

· “Bloom Energy develops devices called oxide fuel cells… capable of directly converting natural gas into electricity usable by data centers. It’s modular and can be deployed quickly.”

· “Their backlog orders total up to $20 billion. Their revenue grew about 34% in 2025, and they expect another 40% growth in 2026.”

· “If you use products like Bloom Energy’s natural gas turbines, you don’t need to rely on the power grid at all. Just install it next to your AI data center.”

Infrastructure and Bitcoin Mining as a ‘Shortcut’

· “Leopold heavily invested in CoreWeave. He made the largest leveraged investment in core GPU infrastructure and energy supply.”

· “He invested in many Bitcoin mining companies… because these companies possess two key elements needed for building AI infrastructure: land and electricity.”

· “He acquired these companies to obtain their licenses and grid access. Usually, getting these licenses takes months or even years.”

· “It’s like taking over a bar that already has a liquor license, instead of applying for a new one and waiting years—an extremely clever ‘shortcut.’”

Short Selling Logic and the End of IT Outsourcing

· “He held a short position in a specific company, Infosys… Their business model relies entirely on providing cheaper labor than Western countries.”

· “He realized these models are now powerful enough to automate simple tasks and handle some critical IT processes, so he shorted this company heavily.”

Investment Philosophy: Returning to the Physical World

· “Companies relying solely on software will become very difficult in the future. This shift isn’t just about architecture but about investing in the physical world—manufacturing, factories, energy, and infrastructure.”

· “These are fields that AI cannot build; they require manpower, licenses, and legislation—hard infrastructure and hardware.”

· “Energy is the only resource that everyone cannot get enough of… All of this revolves around one core: powering the future.”

Young Investment Prodigy Leopold Aschenbrenner

Josh Kale: There’s a guy named Leopold Aschenbrenner, he’s 24 years old. We covered him in a previous episode when he was 23, managing $1 billion, focusing on emerging frontier AI concepts and technologies. But today, just a year later, that $1 billion has grown to $5.5 billion.

This kid, much younger than us, just delivered a groundbreaking performance, earning more in AI than any other fund in the world. More importantly, AI is the hottest market right now, which means fierce competition. So it’s clear that Leopold is doing something different.

Last week, his new 13F report was released, giving us a glimpse into his recent trades. So next, we’ll analyze these documents carefully to see what he did to turn $1 billion into $5.5 billion.

Insights from the 13F Report

Ejaaz Ahamadeen: He achieved all this within 12 months. His fund was established at the end of 2024 with an initial size of $255 million. In just six months, his fund outperformed the S&P 500 by 8 times, reaching $2 billion. Since we last discussed his Q3 fund report on the show, his fund has grown another $1.5 billion. So now, he’s in a period of explosive growth.

He’s very young, made a significant shift, but all of this aligns with his so-called “Bible”—a 165-page article titled “Situational Awareness.”

In this article, he basically predicts we will reach AGI by 2027. He describes in detail his view of how the AI revolution will unfold. His predictions have been almost entirely correct; he foresaw the GPU infrastructure boom, and now he’s proposing a very important shift, which we will explore further.

From Chips to Infrastructure Shift

Josh Kale: I think the entire investment philosophy is shifting from chips to infrastructure. What we see on the screen now is very interesting. He used Claude to create a document that will guide us through the changes from last year to this year. Perhaps we can start with the assets he sold, because these positions are quite large, including NVIDIA, for which he sold $300 million worth of put options in a single quarter.

Ejaaz Ahamadeen: You’ll see many of the stocks he sold are very popular companies that many are investing in now. So the question is, why did he sell $1 billion worth of stocks in these companies? He sold NVIDIA, Broadcom, TSMC, Micron. These are all major AI infrastructure companies.

His sale of NVIDIA stock actually made him money; he held $300 million worth of put options, which means that as NVIDIA’s stock price declined over the past few months, he likely profited from it. So why did he do this?

His 165-page paper mentions that by late 2025 or early 2026, he believes the market will have fully priced in GPU value. This value mainly comes from companies like NVIDIA and Broadcom, which produce these chips and stack them for AI labs like OpenAI and Anthropic to train models.

Now, he’s shifting focus to the main bottlenecks that investors have yet to fully address—energy and infrastructure. Currently, many AI labs face a major problem: First, they have too many GPUs; second, the existing power grid is designed for humans, not to meet the enormous AI demands today. That’s where his current investments are concentrated.

Selling NVIDIA Put Options

Josh Kale: Seeing him sell NVIDIA puts and completely exit NVIDIA’s investment is quite interesting. Because when I talk to friends or Wall Street folks, NVIDIA is always the most discussed stock and the biggest investment target.

And seeing him move away from NVIDIA again proves he’s always ahead, able to foresee future trends rather than sticking to past hot spots. To him, the future focus is infrastructure, shifting from chips to information infrastructure.

This could be an area for us to explore further—these are the stocks you should pay attention to. These are the assets he currently holds and believes will grow in the future. If his judgment is correct, we should see quite substantial returns. So, what new investments did he make this quarter?

Ejaaz Ahamadeen: Here’s a very neat chart categorizing Leopold Aschenbrenner’s entire portfolio by AI tech stack. We see investments divided into categories like power generation, real estate and facilities, compute and hosting, connectivity, storage and memory, chips and silicon, etc.

Actually, I want to add a point I mentioned earlier: I noticed he made a very clever trade involving Intel. He sold his stock but still held a large long position. This way, he freed liquidity and redirected funds into other companies.

The main company he invested heavily in is in the power generation sector—Bloom Energy. This company was almost unknown three months ago but specializes in making power turbines for AI data centers.

He built a huge position in this company, totaling up to $855 million. Although the report shows $865 million, the official figure is $855 million.

Bloom Energy: Power Innovator

Josh Kale: Bloom Energy is his largest current investment, accounting for 20% of his portfolio. It’s completely unrelated to chips—a totally different direction. I looked into their business, and it’s quite fascinating.

Bloom Energy developed a device called oxide fuel cells, an advanced technology that generates power directly from natural gas. Usually, natural gas is transported to data centers and then heated and cooled via turbines—a very clunky energy process.

Bloom’s “fuel cell” can directly convert natural gas into electricity usable by data centers. It’s modular, deployable quickly, and apparently has no supply shortages. As far as I know, they plan to produce 2 gigawatts of power this year.

This is a very interesting energy play. I’ve been looking for an “NVIDIA of energy”—a chip manufacturer in the energy sector. I haven’t found a perfect match yet, but Bloom Energy might become such a company.

Ejaaz Ahamadeen: I also checked their latest financial report since they are a listed company. Their backlog orders total $20 billion. Their revenue grew about 34% in 2025, and they expect another 40% growth in 2026. Clearly, their demand exceeds supply.

You mentioned oxide fuel cells. Their natural gas turbines are especially attractive because they don’t rely on the existing power grid. As I mentioned earlier, the current grid is under huge pressure because humans need energy, and AI data centers also need energy, which drives up energy prices in those regions.

If you use products like Bloom Energy’s natural gas turbines, you are completely independent of the grid. Just install it next to your AI data center, and you can get power efficiently for training or inference.

Companies like Broadcom and CoreWeave will need this kind of energy, especially large cloud providers and AI labs. It reminds me of the game Civilization—have you played it? It’s like moving infrastructure and energy production facilities to your own small settlement to boost its development—the scenario is very similar.

Josh Kale: It’s obvious that energy shortages are not the issue; the question is who can produce the most energy. They do have a huge backlog, but the real question is whether they can produce enough to meet demand.

Manufacturing capacity becomes a key issue here. Many such investments are entering a “atomic” world—where manufacturing truly becomes critical. I’d love to explore whether they really have the capacity for large-scale production. For now, it’s undoubtedly a very important sector, making up 20% of his portfolio. So, what other notable positions are in his new portfolio?

Ejaaz Ahamadeen: He also increased his investment in CoreWeave by about $300 million. Imagine an AI lab needing GPUs. Buying GPUs from NVIDIA is just part of the job.

Deploying these GPUs into racks, providing power, engineering support, and maintaining cooling systems—this is a whole other story. You can outsource these tasks to a “new cloud provider” like CoreWeave, which specializes in handling these.

Broadcom also offers similar services, but CoreWeave is smaller, initially focused on gaming GPUs, now transformed into an AI-focused company. Leopold has made a big bet on CoreWeave.

In our previous discussion of Q3, he invested $500 million, and now he added another $300 million. His total investment in CoreWeave might be around $800 million. But there’s more: he also holds about 10% of Core Scientific, one of CoreWeave’s main suppliers, which provides energy grid services.

If you consider his investment strategy, Leopold is making the biggest leveraged bets in core GPU infrastructure (like CoreWeave’s new cloud services) and energy supply (like Bloom Energy)—these are his two main holdings.

Bitcoin Mining

Josh Kale: I find it interesting that he’s already holding enough shares in these companies to become an activist investor, capable of influencing their decisions. That’s very intriguing. When I studied his portfolio, besides power generation, I noticed he added the most positions related to real estate—about 10 of them—and these are connected to Bitcoin mining.

What we see now is that he’s invested in many Bitcoin mining companies. It seems odd, even illogical, since the crypto market isn’t doing well, and Bitcoin’s performance isn’t great. Why buy into these Bitcoin miners? Because these companies possess two key elements needed for building AI infrastructure: land and electricity.

What does Bitcoin mining need? Massive energy and enough space for GPU racks. Although Bitcoin mining isn’t in full decline, these companies’ real estate and power assets could offer better risk-adjusted returns. It looks like he’s betting these Bitcoin miners will either sell their land rights and licenses or pivot directly into AI data centers.

Ejaaz Ahamadeen: To be clear, his interest in these companies isn’t for mining; he’s acquiring them to get their licenses and grid access. Usually, obtaining these licenses takes months or years.

That’s why we see companies like Meta, Microsoft, and OpenAI announcing $1.4 trillion in compute partnerships, but these collaborations haven’t fully translated into their models yet. That’s also why GPU supply always lags behind the latest generation—licenses take too long.

Leopold is acquiring small companies that already have licenses, bypassing the entire licensing process. He’s divested from their crypto services and repurposed them for training AI models, becoming an infrastructure provider for AI labs.

It’s like taking over a bar that already has a liquor license instead of applying for a new one and waiting years—a very clever “shortcut.”

AGI and Market Trends

Josh Kale: One of the things I admire most about his investment philosophy, and how it’s been validated over the past year, is its simplicity and efficiency. For example, Bitcoin mining companies obviously hold licenses and energy, and every AI company needs these resources.

So why isn’t everyone buying these companies? I think it’s because these ideas are too simple, and many are blocked from investing. But time and again, his simple ideas prove correct.

Will Leopold’s prediction of achieving AGI by 2027 also be correct? Will we really reach AGI in 2027?

Ejaaz Ahamadeen: To test this prediction, we set up a prediction market on Polymarket, asking whether OpenAI will announce AGI before 2027. Currently, when Leopold proposed this fund, many didn’t believe his forecast, but now the market’s probability is only 13%. So, it seems quite distant. His investment thesis might be right, but the timeline could be slightly off.

That probability is indeed small. But I must say, he was initially criticized for that paper, with many thinking his views were too outlandish and unrealistic. About 50% believe AGI will be achieved within the next few months, while others think it will be around 2030. Leopold is the only one who predicted 2027 and is closest to the mark so far.

He predicted the importance of GPUs before the GPU boom exploded. Now, he’s making a prediction before the energy infrastructure boom hits. So I think he’s still ahead in this regard.

However, his portfolio isn’t just long positions; he also holds a short position in Infosys, a company focused on IT outsourcing, mainly in India. Their business model relies entirely on providing cheaper labor than Western countries—simply, “outsourcing all your administrative IT work to us.”

I think his bet here is based on observing a trend. He saw the rise of products like Claude Code and GPT Codex 5.3, and realized these models are now powerful enough to automate simple tasks and handle some critical IT processes. So he shorted this company heavily.

This is one of his deeper investments, aligning with current trends, showing he’s willing to put real money behind his views.

Bull and Bear Markets

Josh Kale: Let’s discuss reasons for a bull market and a bear market. When you look at such a portfolio, what are the criticisms or cautions? First, this investor is only 24—I’m not sure if he has the experience many others do. That could be an advantage, but at some point, this advantage might collapse.

Another concern is that his investment thesis is somewhat like a single-theme bet. If the growth of AI infrastructure and related spending slows down, or macroeconomic conditions change, each position in this portfolio could face downside risks. There’s hardly any hedging. So, this strategy has potential vulnerabilities, but all signals currently suggest his fund will keep performing well.

Ejaaz Ahamadeen: If you look at some of the most successful investors today, their success isn’t about making money in a single year or quarter, but about achieving steady returns year after year, decade after decade, and compounding.

Leopold’s start has been stunning; his performance far exceeds the average of hedge funds in any industry, not just AI. But he still needs to prove himself over a longer horizon—time will tell.

I just want to say, this guy, once fired by OpenAI, has deep insights into AI’s future and has made the boldest predictions. He’s the only one whose predictions have been almost entirely correct so far. He poured a lot of effort into that 165-page paper, confident in his views, and it’s paying off.

Will things change in the future? Possibly. But you can see these reports and investments as real-time tracking tools for bottlenecks in the AI race. I want to emphasize that. His initial fund focused on GPUs, believing they’d be the demand hotspot, and the market underestimated this opportunity.

Now, he believes the next bottleneck is energy infrastructure.

Look at Elon Musk—he’s launching data centers into space. Why?

Because the sun provides more energy. Also, companies like Google, Meta, Broadcom, and NVIDIA are investing in data centers or infrastructure to access the power grid. He’s just putting money where the demand is, which I think is very smart.

Josh Kale: I recently read a great article by Naval, which argues that companies relying solely on software will become very difficult in the future because developing and generating custom software is now very easy. I see his shift as not just about architecture but about investing in the physical world—manufacturing, factories, energy, and infrastructure.

These are fields that AI cannot build; they require manpower, licenses, and legislation—hard infrastructure and hardware. I believe this is the future direction.

Energy is the only resource that everyone can’t get enough of. Whether in power generation or real estate, it all revolves around one core idea: powering the future. In the last earnings season, just Google, Amazon, and NVIDIA committed over $650 billion in capital expenditure, showing huge investments aimed at solving this problem. His portfolio is clearly positioned to capture this upside.

Ejaaz Ahamadeen: Yes, he’s made some high-risk bets, like Bloom Energy, which many might not have heard of unless they’re very familiar with energy infrastructure.

But this company can be seen as a top-tier energy firm, especially in portable energy. He connected the dots, believing the grid can’t support current demand, so he invested heavily. We’re talking about nearly one-fifth of his entire portfolio in this single asset.

This is a highly concentrated, high-risk, high-conviction approach. But if it works, that’s how his portfolio could return 4.5 to 5 times in a year and a half. We must respect his achievement—turning $1 billion into $5.5 billion in a year is incredible.

Leopold’s Investment Future

Josh Kale: Overall, it’s astonishing what he’s achieved, and his latest shift from hardware to infrastructure and energy looks very promising. The outlook is very optimistic. If you believe in his portfolio, it might be a good opportunity to watch. Of course, this isn’t investment advice—just an analysis of his holdings. But it does look very hopeful and could perform extremely well this year.

Josh Kale: I’m also curious about what our listeners think. Do you think our analysis is professional enough, reaching Leopold’s level? Or do you think we’re completely wrong, missing some obvious stories?

Ejaaz Ahamadeen: You know what I want? I want to know what you think is the best stock this year.

Josh Kale: Yes, Leopold is betting on Bloom Energy. I want to know—what’s your Bloom Energy? Did we miss something that could again deliver 5x growth this year?

Original Video Link

Click to learn more about Rhythm BlockBeats job openings

Join the Rhythm BlockBeats Official Community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Twitter Official Account: https://twitter.com/BlockBeatsAsia