From $225 Million to $45 Billion to a Fire Sale: What Happened to Leopold Aschenbrenner’s AI Fund?
A former OpenAI researcher built one of the fastest-growing hedge funds in history. Then leverage destroyed it in six days. Here is what happened, and what it tells you about borrowed money.
In late 2024, Leopold Aschenbrenner launched a hedge fund called Situational Awareness with approximately $225 million. By the end of June 2026, it had grown to $45 billion — a return of roughly 1,000% in under two years, making it one of the best-performing large funds in history.
By July 30, 2026, it was gone. Goldman Sachs, JPMorgan Chase, and Bank of America issued margin calls on the fund’s leveraged positions. The entire public book was sold in a single block trade to Ken Griffin’s Citadel at distressed prices. Assets fell from $45 billion to around $10 billion in days. Aschenbrenner, who was getting married that weekend in California, watched the fund he had built unwind in less than a week.
The story is remarkable on its own terms. But it is also one of the clearest illustrations available of how leverage works and why it is the single most dangerous tool in investing.
Who Is Leopold Aschenbrenner?
Aschenbrenner is a 25-year-old former researcher at the OpenAI Foundation. He became publicly known in 2024 when he published a long document arguing that artificial general intelligence was closer than most people assumed and that the geopolitical stakes of AI development were being dangerously underestimated. The document circulated widely in technology and finance circles.
After leaving OpenAI, he launched Situational Awareness as an investment fund built around a single conviction: that the AI infrastructure boom was real, durable, and being underpriced by markets. He put that conviction into practice through concentrated, leveraged bets on semiconductor companies, AI infrastructure providers, and related names. For a year and a half, the bet paid off spectacularly.
What Leverage Is
To understand what happened, you need to understand leverage. It is the practice of borrowing money to invest, amplifying both gains and losses.
A simple example. Suppose you have €10,000 and invest it in a stock that rises 10%. You make €1,000, a 10% return. Now suppose you borrow an additional €30,000 and invest €40,000 in the same stock. The same 10% rise now produces €4,000, a 40% return on your original €10,000.
That is why leverage is attractive. It magnifies returns when the trade goes right.
The same mechanism works in reverse. If the stock falls 25%, an unlevered investor loses €2,500. An investor leveraged four times over loses €10,000, their entire original capital, and still owes the borrowed €30,000 to whoever lent it to them.
Situational Awareness was reportedly running leverage of approximately four times. That means for every €1 of investor capital, the fund had borrowed roughly €3 more. While AI stocks were rising, this produced extraordinary returns. When they fell, the losses were catastrophic.
What a Margin Call Is
When an investor borrows money to invest through a broker, the broker requires that the account maintain a minimum level of collateral — a cushion of real assets relative to the borrowed amount. If the portfolio falls in value and that cushion shrinks below the required level, the broker issues a margin call: a demand to deposit more cash or sell assets immediately to restore the required ratio.
Margin calls do not wait. They are not negotiable. If the investor cannot meet the call, the broker sells their positions, often at whatever price is available in the market at that moment.
Aschenbrenner’s prime brokers issued margin calls on July 30 after AI infrastructure stocks fell and the fund’s cushion was exhausted. Unable to meet the calls, the fund sold its entire public book in one transaction to Citadel at distressed prices. That is what you get when you are a forced seller.
Why Conviction Is Not Enough
This is the part of the story worth sitting with. Aschenbrenner was not obviously wrong about AI. The technology is real. The infrastructure buildout is real. The companies he invested in may well be worth significantly more in five years than they are today. His original thesis — the one that produced a 1,000% return in eighteen months — has not been disproven.
What destroyed the fund was not the thesis. It was the leverage. A concentrated, leveraged bet does not give the investor time to be right. It requires being right, at the right scale, on the right timeline. When AI stocks fell in July 2026, it did not matter whether the long-term case was correct. The margin calls came immediately, and forced selling does not wait for the market to recover.
This is a pattern that repeats throughout financial history. Long-Term Capital Management, the famous hedge fund staffed with Nobel Prize winners, collapsed in 1998 not because its models were wrong but because leverage amplified losses beyond what it could survive. The AI boom’s greatest beneficiary just added itself to the same list.
What This Means for Ordinary Investors
The Aschenbrenner story is extreme — most investors do not run four times leverage on a $45 billion portfolio. But the underlying lesson applies at any scale.
Leverage is not inherently wrong. Mortgages are leverage. Business loans are leverage. The question is always whether the asset being financed can generate returns that comfortably exceed the cost of borrowing, and whether the investor can survive a period of adverse prices without being forced to sell.
The answer to the second question depends on not being forced. Margin calls force selling. Over-leveraged positions force selling. Any structure where a temporary fall in price triggers an automatic, non-negotiable exit removes the investor’s ability to wait out a bad period, even when they are ultimately right.
Aschenbrenner built one of the most remarkable investment records in recent history and then lost most of it in six days. The AI stocks he owned may yet recover. He will not benefit when they do.
Key Takeaways
- Leopold Aschenbrenner’s fund Situational Awareness grew from $225 million to $45 billion in under two years, then collapsed to around $10 billion in days after margin calls forced a distressed sale.
- Leverage amplifies both gains and losses. The fund was reportedly running approximately four times leverage — for every €1 of investor capital, roughly €3 was borrowed.
- A margin call is a demand from a broker to deposit more capital or sell assets immediately when a portfolio falls below a required collateral threshold. They are not negotiable.
- The fund’s thesis about AI infrastructure may still be correct. Leverage removed the ability to wait for that thesis to play out.
- Borrowed money in investing introduces a forced exit that patience and conviction cannot override.