AI’s Biggest Beneficiary: The Rise of Leopold, the New Stock God on Wall Street

Leopold Aschenbrenner’s holdings have surged again. As a rising star in the hedge fund world, his investment logic is being validated by market trends in real-time.

Over the past few days, multiple stocks in Leopold’s Situational Awareness LP public portfolio have risen collectively: Bloom Energy, Cipher Mining, Intel, Applied Digital, SanDisk, IREN, and other targets saw single-day gains of over 10% at one point. This has prompted the market to revisit his 13F filing from late last year, trying to understand why this former OpenAI researcher bet early on the AI infrastructure track.

What makes him noteworthy isn't the flashy labels of "young" or "overnight millionaire," but rather the framework he offers, which diverges from mainstream AI trading. Most equate AI investment with Nvidia, Microsoft, OpenAI, and model capabilities, yet Leopold’s portfolio bypasses the most crowded star assets, pivoting instead towards Bloom Energy, CoreWeave, Core Scientific, Lumentum, Intel, Bitcoin miners, and power-related companies.

The AI narrative is shifting from "whose model is stronger" to "who can sustain further model expansion." Training and inference require GPUs, GPUs require data centers, and data centers need power, land, cooling, fiber optics, permits, and long-term energy supply contracts. Leopold is betting on the physical bottlenecks that AI’s continued growth must navigate. Fortune summed up his latest holdings: this former OpenAI researcher is translating his AGI thesis into billions of dollars in bets on electricity, AI infrastructure, and crypto miners.

In early March, 动察Beating conducted an in-depth analysis of Leopold, his fund holdings, and investment logic, sharing his vision of the future of AI competition. All of this is now being confirmed in reality: the AI narrative is retreating from the model on the screen back to the land and power grid beneath our feet. The most valuable things in the future may not be algorithms, but the physical world that supports the continued expansion of algorithms.

The following is the original content from 动察Beating:

In February 2026, the hedge fund Situational Awareness LP submitted its quarterly holdings report, showing that as of the end of the fourth quarter of 2025, the total market value of the fund's U.S. stock holdings was $5.517 billion.

Wall Street manages trillions of dollars in assets; $5.5 billion is a drop in the bucket. However, just 12 months prior, this fund had less than $400 million under management. Moreover, its founder and Chief Investment Officer is a young man born in 1999.

His name is Leopold Aschenbrenner. He is 27 years old.

In 12 months, he grew the fund from $383 million to $5.517 billion, an increase of over 14 times. During the same period, the S&P 500 posted single-digit gains.

What's even more astonishing is his portfolio. Opening the quarterly holdings report, you won't find any of the AI star companies you typically see on financial news headlines. Instead, you'll find a fuel cell company, Bitcoin miners just climbing back from the brink of bankruptcy, and a chip giant being abandoned by the broader market.

He claims his fund invests in AI, but this doesn't look like an AI fund's portfolio at all. It looks more like a madman's shopping list.

But this "madman" happens to be one of the earliest and deepest thinkers on how AI will transform the world. Before joining Wall Street, he was a researcher at OpenAI, tasked with contemplating how to ensure AI doesn't spiral out of control when it becomes smarter than humans. Later, he was ousted for saying things he shouldn't have, wrote a 165-page manifesto predicting a future most find absurd, and then went all-in with his entire net worth.

Deconstructing $5.5 Billion: What Exactly Did He Buy?

To understand Leopold Aschenbrenner's genius in investing, the most direct way is to open his holdings report and read it line by line.

His top holding is Bloom Energy, with a market value of $876 million, accounting for 15.87% of the total portfolio.

This company makes fuel cells. More precisely, it produces something called "Solid Oxide Fuel Cells" that convert natural gas directly into electricity with high efficiency. Founder KR Sridhar, a former NASA engineer on the Mars exploration program, was called "one of the top five futurists shaping the future today" by Fortune magazine.

An AI fund places its biggest bet on a power generation company.

According to Gartner's forecast, global electricity consumption by AI-optimized servers will skyrocket from 93 terawatt-hours in 2025 to 432 terawatt-hours in 2030, nearly a fivefold increase in five years. Grid power demand for U.S. data centers is expected to nearly triple by 2030, reaching 134.4 gigawatts. The average age of the U.S. power infrastructure exceeds 25 years, with many components aged between 40 and 70 years, far exceeding their design life.

In other words, AI needs more electricity than the entire grid can provide. And the grid itself is aging and about to fall apart.

The scarcest resource in the AI era isn't chips; it's electricity.

Bloom Energy's fuel cells happen to bypass this bottleneck. They don't need to connect to the grid; they generate power directly on-site at data centers, 24/7. In 2025, Bloom Energy secured a contract from CoreWeave to provide fuel cells for its AI data center in Illinois.

Speaking of CoreWeave, this is Leopold's second-largest holding.

He holds $774 million in CoreWeave call options, plus $437 million in common stock, totaling over $1.2 billion, or 22% of his portfolio. CoreWeave is a GPU cloud service provider that pivoted from cryptocurrency mining.

In 2017, Mike Intrator and Brian Venturo got together to mine Bitcoin. The crypto market crashed in 2018, making mining unviable. But they had a bunch of GPUs. In 2019, they had a eureka moment: GPUs aren't just for mining; they can also run AI.

So the company transformed from a mining farm to an AI computing arms dealer. On March 27, 2025, CoreWeave IPO'd on the Nasdaq, raising $1.5 billion at $40 per share. A company born from a mining farm became a core supplier of AI infrastructure.

Leopold was attracted by CoreWeave's vast GPU inventory and its deep relationship with Nvidia. In an era where computing power is productivity, whoever holds the GPUs is king.

But the truly puzzling holding is his third-largest: Intel. The portfolio value is $747 million, all in call options, accounting for 13.54% of the total.

In 2025, Intel was one of Wall Street's most out-of-favor companies. Its stock price halved from its 2024 highs, market share was eroded by AMD and Nvidia, and CEOs were replaced one after another. Almost every analyst declared Intel was finished.

Yet Leopold chose precisely this time to make a heavy bet using call options. This is an extremely aggressive move, promising a moonshot if right and total loss if wrong.

What is he betting on? Simply put: foundry services.

In November 2024, the U.S. Department of Commerce announced that Intel would receive up to $7.86 billion in direct funding under the CHIPS and Science Act. The sole purpose of this money is to make Intel a domestic chip foundry for the U.S., competing with TSMC.

Against the backdrop of U.S.-China tech decoupling, America needs a "homegrown" champion to manufacture chips. Although Intel is behind, it's the only option. Leopold isn't betting on Intel's technology; he's betting on the national will of the United States.

The subsequent holdings become even more interesting. Core Scientific at $419 million; IREN at $329 million; Cipher Mining at $155 million; Riot Platforms at $78 million; Hut 8 at $39.5 million.

These companies share a common trait: they are all Bitcoin mining companies.

Why would an AI fund invest in a bunch of Bitcoin miners?

Simple: Bitcoin mining companies possess the cheapest electricity and the largest available data center sites in the United States.

Core Scientific holds over 1,300 megawatts of power capacity. IREN plans to expand 1.6 gigawatts of capacity in Oklahoma. To survive intense hash rate competition, these miners have long secured the cheapest power resources globally, signing long-term Power Purchase Agreements.

Now, the scarcest resources for AI data centers are precisely power and land.

In 2022, Core Scientific filed for bankruptcy due to the crypto crash. It completed restructuring in January 2024, shed about $1 billion in debt, and relisted on the Nasdaq. Then, it signed a 12-year contract worth over $10.2 billion with CoreWeave, converting its mining farms into AI data centers. To fully commit to this transformation, Core Scientific even plans to sell off all its Bitcoin holdings.

IREN (formerly Iris Energy) signed a $9.7 billion AI contract with Microsoft, receiving $1.9 billion in prepayments. Cipher Mining signed a 15-year lease agreement with Amazon. Riot Platforms inked a 10-year, $311 million contract with AMD.

Overnight, Bitcoin miners became the landlords of the AI era.

Now, let's put this puzzle together.

Bloom Energy provides electricity, CoreWeave provides GPU computing power, Bitcoin miners provide land and cheap power, and Intel provides domestic chip manufacturing capabilities. Add the fourth-largest holding, Lumentum ($479 million, optical components crucial for interconnecting AI data centers), the ninth-largest, SanDisk ($250 million, data storage), and the eleventh-largest, EQT Corp ($133 million, natural gas producer providing fuel for fuel cells).

This forms a complete AI infrastructure supply chain.

From power generation, to transmission, to chip manufacturing, to GPU computing, to data storage, to fiber optic interconnection – he has bought into every link.

And another move he made clarifies this logic further. During the fourth quarter of 2025, he completely liquidated his positions in Nvidia, Broadcom, and Vistra. These three companies were precisely the star performers with the highest gains in the 2024 AI rally.

He also shorted Infosys, one of India's largest IT outsourcing companies.

Sell the hottest AI chip stocks, buy unwanted power plants and mining farms. Short traditional IT outsourcing, because AI coding tools are making programmers more efficient, compressing the demand for outsourcing.

Every trade points to the same judgment: AI's bottleneck is not in software, but in hardware; not in algorithms, but in electricity; not in cloud models, but in the physical world.

So the question arises: How did a 27-year-old develop this framework?

From the Son of East German Doctors to OpenAI's Rebel

Leopold Aschenbrenner was born in Germany to parents who are both doctors. His mother grew up in East Germany, his father in West Germany, and they met after the fall of the Berlin Wall. This family background itself carries an imprint of historical rupture – Cold War, division, reunion. The seeds of his later obsession with geopolitical competition can perhaps be traced here.

But Germany couldn't keep him. He later said in an interview: "I really wanted to leave Germany. If you are the most curious kid in the class and want to learn more, the teacher doesn't encourage you; they get jealous and try to suppress you."

He calls this phenomenon the "tall poppy syndrome" – whoever grows tall gets cut down.

At age 15, he convinced his parents, flew alone to the United States, and entered Columbia University.

Entering university at 15 makes one an anomaly anywhere. But Leopold's performance at Columbia turned "anomaly" into "legend." He double-majored in Economics and Mathematics-Statistics, winning nearly every available award, such as the Albert Asher Green Memorial Prize, the Romine Economics Prize, and membership in the Junior Phi Beta Kappa honor society.

At 17, he wrote a paper on economic growth and existential risk. Renowned economist Tyler Cowen, after reading it, said: "When I read it, I couldn't believe it was written by a 17-year-old. If it were an MIT doctoral dissertation, I would be impressed."

At 19, he graduated from Columbia University as Valedictorian, the highest undergraduate honor. In 2021, while the world was still under the shadow of the pandemic, a 19-year-old German boy stood at Columbia's graduation ceremony, delivering the valedictory address on behalf of all graduates.

Tyler Cowen gave him one piece of advice: don't pursue a Ph.D. in economics.

Cowen felt academia in economics had become somewhat "decadent" and encouraged him to do bigger things. Cowen also introduced him to the Silicon Valley "weird Twitter" culture, a circle fascinated by AI, effective altruism, and humanity's long-term destiny.

After graduation, Leopold first went to the Forethought Foundation, researching long-term economic growth and existential risk. He then joined the FTX Future Fund founded by SBF, working with key figures in the effective altruism movement like Nick Beckstead and William MacAskill. His title was "Economist Affiliated with the Global Priorities Institute at Oxford University."

This period was significant. It means that before entering the AI industry, Aschenbrenner had spent years systematically thinking about one question: What kind of events could fundamentally change the trajectory of human civilization?

Then, he joined OpenAI.

The exact timing is unclear, but he joined a specific team – the "Superalignment" team. Established on July 5, 2023, it was co-led by OpenAI co-founder Ilya Sutskever and alignment team leader Jan Leike. Its goal was to solve the superintelligence alignment problem within four years – ensuring an AI far smarter than humans would still obey them.

OpenAI had promised to dedicate 20% of its computing power to this team. But between the promise and reality lay a chasm.

Inside OpenAI, Leopold saw things that troubled him. He submitted a security memo to the board, warning that the company's safety measures were "severely deficient" in preventing foreign governments from stealing key algorithmic secrets. The company's reaction was unexpected. HR called him in, saying his concerns about espionage were "racist" and "unconstructive." Company lawyers grilled him on his views about AGI and the loyalty of his team.

In April 2024, OpenAI fired him, citing "leakage of confidential information."

The alleged "leak" involved sharing a brainstorming document on AGI safety measures with three external researchers. Leopold said the document contained no sensitive information and that sharing such documents internally for feedback was standard practice.

A month later, Ilya Sutskever left OpenAI. Three days after that, Jan Leike also left. The Superalignment team disbanded. The promised 20% computing power from OpenAI was never delivered.

A team researching "how to control superintelligence" was disbanded by the very company building it.

The irony of this situation cannot be overstated. But for Leopold, being fired became a liberation. He was no longer employed by anyone, no longer needing to carefully phrase internal memos. He could finally speak his true thoughts to the world.

On June 4, 2024, he published a 165-page article on a website called situational-awareness.ai. The title: "Situational Awareness: The Decade Ahead."

The 165-Page Prophecy

To understand Leopold's investment logic, you must first read this manifesto. Because that $5.5 billion portfolio is the financial translation of these 165 pages.

The core argument of the manifesto can be summarized in one sentence: There is a very high probability that AGI (Artificial General Intelligence) will be achieved by 2027.

This judgment sounded like madness in June 2024. But Leopold's method of argument is straightforward: count the orders of magnitude.

From GPT-2 to GPT-4, AI capabilities made a qualitative leap, evolving from a preschooler to a smart high school student. Behind this leap was approximately a 100,000-fold (5 orders of magnitude) increase in effective compute. This growth came from stacking physical compute, improving algorithmic efficiency, and the capability release from "de-bottlenecking" models.

His prediction is that by 2027, a similar magnitude of growth will occur again. In terms of physical compute, the resources used to train the most advanced models will be 100 times greater than for GPT-4. Algorithmic efficiency improves by about 0.5 orders of magnitude annually, accumulating roughly 100-fold over four years. Add the gains from "de-bottlenecking," turning AI from a chatbot into an agent capable of using tools and acting autonomously, which represents another order of magnitude leap.

Three 100-fold multipliers combine to yield another 100,000-fold increase, another qualitative leap. From a smart high schooler to surpassing human capabilities.

What makes this article truly unsettling is the series of consequences he deduces from this prediction.

The first consequence: Trillion-dollar compute clusters.

He writes that over the past year, Silicon Valley's