Citadel, one of the world's biggest hedge funds, bought most of the stock portfolio belonging to a smaller, AI-focused hedge fund called Situational Awareness. The sale happened in July 2026 after Situational Awareness suffered deep losses on risky AI stock bets. The portfolio Citadel acquired was worth about $16 billion (WSJ; FT).

Situational Awareness was founded by Leopold Aschenbrenner, a 24-year-old former researcher at OpenAI, the company behind ChatGPT. OpenAI fired him in April 2024 over alleged leaks (FT). He then turned a viral essay about artificial intelligence into a hedge fund, launching in late 2024 with about $225 million in assets (SpotGamma; Benzinga). The fund later raised $1.5 billion from investors, including Jane Street, a major trading firm (FT; WSJ).

The fund's growth was explosive. At its peak, it reached $20 billion in size (Benzinga). It returned 439% during the first half of 2026 alone (Forbes). As of March 31, 2026, the fund disclosed a $5.52 billion U.S. stock portfolio in a regulatory filing (Yahoo Finance).

The reversal came abruptly. The fund had used borrowed money to place big bets on AI stocks. Think of it like buying a house with a mortgage: if the house price rises, you profit on the full value, but if it falls, you still owe the bank. When AI stock prices dropped, the lenders demanded more collateral, which is called a margin call. When the fund could not come up with the cash, it was forced to sell off its positions quickly (SpotGamma; The Next Web). The sell-off and the sale to Citadel both happened in July 2026 (Forbes).

Following the losses, Aschenbrenner's fund was seeking to raise fresh capital (FT; Benzinga). This came after a broader decline in AI stock prices in mid-2026.

The formal structure of the fund places Aschenbrenner as Managing Partner of SAF AI GP LP, which serves as the general partner of Situational Awareness Partners LP, according to an SEC filing (SEC).

The trajectory from $225 million to $20 billion in under two years, followed by a forced sell-off, follows a pattern that has played out many times in financial history: borrowed money can produce extraordinary gains, but it also makes a fund fragile. When the trend reverses, the borrowed money magnifies losses instead of gains, and the lenders pull the plug. A fund that returns 439% in six months is almost certainly betting with far more money than it actually owns. The margin call is the predictable result.

Some market observers have interpreted the collapse and Citadel's acquisition of the portfolio as potentially marking a bottom in the AI trade, rather than a negative verdict on AI as an investment theme (The Next Web). The logic follows a well-worn pattern: forced selling by a fund that borrowed heavily transfers the risk to a buyer like Citadel, which has its own capital, lower borrowing costs, and a longer time horizon. Citadel, as a large firm with deep resources, is in a position to hold a big stock portfolio that a smaller, specialized fund could no longer afford to keep.

What makes this episode unusual is the speed of the scale-up. A fund that launched in late 2024 with $225 million and reached $20 billion at peak, returning 439% in the first half of 2026, was not growing that fast simply because investors were sending money. The gap between the $1.5 billion raised from investors and the $20 billion peak size means that borrowed money and paper gains accounted for the majority of the fund's apparent size. That matters to the banks that lend to hedge funds and to anyone trying to understand how much risk is concentrated in AI stocks across the hedge fund industry.

The $5.52 billion U.S. stock portfolio disclosed as of March 31, 2026, provides a snapshot of the fund's holdings at that date. But the approximately $16 billion portfolio value at the time of Citadel's purchase suggests either the holdings grew significantly between March and July, the fund had exposure through other financial instruments not shown in the filing, or both. The gap between what was disclosed and the final purchase price is worth noting for anyone tracking how AI-related risk flows through regulatory filings versus what is actually changing hands.

The broader question for anyone watching the markets is whether the Situational Awareness collapse was a one-off or the first sign of a wider problem across hedge funds that have borrowed heavily to bet on AI stocks. Citadel's willingness to buy the portfolio at scale suggests there are well-funded buyers still interested in AI stock exposure. Whether that interest holds up if more leveraged funds face the same kind of pressure is the question nobody can answer yet.