The hedge fund that made Leopold Aschenbrenner the most famous 24-year-old in artificial intelligence has sold most of its public stock positions to Citadel, according to people familiar with the matter, ending a run that began with promises of a generational AI trade and ended with July’s debt-fueled reckoning.
Aschenbrenner’s fund, Situational Awareness, unloaded the bulk of its public holdings to the firm run by Ken Griffin, the people said. The fund kept its private stake in Anthropic, the AI lab where Aschenbrenner spent years as a researcher before leaving to raise money. The deal was reported earlier by The Wall Street Journal.
The sale caps a brutal quarter for the fund, which had grown to roughly $45 billion in assets at its peak, an extraordinary sum for a manager who turned 24 this year and whose firm is younger than most of its employees. The fund built its positions with borrowed money, and the July selloff in AI stocks turned a large drawdown into a forced exit.
The details of the unwind are still emerging. Business Insider canvassed Wall Street traders and analysts for their read on the Citadel purchase, and the consensus was that the sale was a deleveraging, not a change of conviction. Aschenbrenner kept his Anthropic stake, the argument runs, because he still believes in the long-term AI story; he simply could not survive the volatility of the public market with borrowed money on top.
The episode has become a case study in the perils of concentrated, levered bets on a single theme. Aschenbrenner built Situational Awareness around a thesis he laid out in a widely read essay: that AI would transform the global economy within a decade, and that the companies building it would be the greatest investment opportunity of the century. The thesis attracted capital from family offices and funds that wanted exposure to AI beyond the public megacaps.
What the thesis could not survive was timing. July’s selloff, triggered by fears that AI capital spending had outrun demand, hit the fund’s concentrated portfolio hard. The borrowed money, which had amplified the fund’s gains during the AI rally, amplified the losses on the way down, and the margin calls followed.
The Citadel purchase is being described by traders as a bargain for Griffin’s firm. Buying the fund’s positions at forced-sale prices, in a market that had already fallen sharply, gives Citadel exposure to the AI trade at a discount to where those stocks traded a month ago. If the AI selloff proves to have been an overreaction, Citadel will have been paid to take the other side.
Aschenbrenner’s backers, including investors who put money in at the fund’s early stages, are left with a familiar lesson in the mathematics of margin: a 40% drawdown on a book financed with borrowed money can wipe out gains that took years to build. The fund’s collapse is not a referendum on AI, most analysts said; it is a case study in what happens when conviction meets volatility without a margin of safety.
Aschenbrenner’s rise was among the fastest in hedge fund history. He left OpenAI in 2024 and published an essay, also called Situational Awareness, arguing that artificial general intelligence would arrive within a few years and that the world was unprepared. The essay made him a celebrity in AI circles, and the fund he named after it raised money at a pace that surprised even its backers, growing to roughly $45 billion in assets at its peak, according to people familiar with the matter.
The unwind itself is a study in how fast forced selling moves. The fund’s positions were sold in blocks to Citadel over a short period, according to the people familiar with the matter, and the sale prices reflected a market in motion. For the stocks involved, mostly large-cap AI and semiconductor names, the selling added to the downward pressure in July even as other investors were trying to gauge whether the selloff had gone too far.
The episode has also drawn attention to how much borrowed money sits behind the AI trade. Concentrated positions in a handful of large names, financed with margin, unwind quickly when prices fall, and the speed of the fund’s collapse offered a preview of that dynamic. The stocks involved have since recovered some of their losses, but the lesson for investors who borrowed to ride the AI rally is that the exits are narrower than the entrances.
For the AI capital complex, the episode is a marker of how quickly the money has moved. A fund that raised billions on the strength of a single thesis, with a founder young enough to be mistaken for an intern, grew to rival established hedge funds in assets and then unwound in weeks. The next cycle will produce another fund with the same ambition; whether it learns from this one’s use of borrowed money is the question its backers will be asking.
