Leopold Aschenbrenner's AI-focused hedge fund cratered 67% in a single month, then sold its entire portfolio to Citadel. The collapse is a stark reminder that even the most sophisticated machine-learning strategies can't escape the need for old-fashioned risk controls.
A single month of losses
The fund, which had been betting heavily on AI-driven trading models, saw its value evaporate over the course of just 30 days. Aschenbrenner, a former OpenAI researcher turned hedge fund manager, had built a reputation on using deep learning to find market inefficiencies. But the rapid drawdown wiped out more than two-thirds of the fund's capital.
Details on what exactly triggered the losses remain scarce. The fund did not disclose specific positions or the nature of the trades that went wrong. What is clear is that the decline was swift and severe, leaving little time for a recovery.
The Citadel fire sale
After the losses, Aschenbrenner's fund sold all remaining assets to Citadel, the massive Chicago-based hedge fund and market maker. The sale was a complete liquidation, meaning Citadel now holds whatever positions the AI fund had left. Terms of the deal were not disclosed.
For Citadel, the acquisition is a bet that the distressed assets can be turned around or unwound profitably. For Aschenbrenner's investors, it marks the end of a once-promising experiment in AI-driven investing.
Why risk management matters for AI funds
The episode underscores a critical lesson for the growing number of hedge funds that rely on artificial intelligence. AI models can identify patterns and execute trades at speeds humans can't match, but they can also amplify losses when markets move against them. Without robust risk management and liquidity strategies, a single bad month can be fatal.
Liquidity is especially important. If a fund's positions are hard to sell quickly, a sudden margin call or redemption wave can force a fire sale at the worst possible prices. That appears to be what happened here: the fund had to sell everything to Citadel, likely at a discount.
Other AI-focused funds are now reviewing their own risk frameworks. The collapse of Aschenbrenner's fund is a cautionary tale that no amount of algorithmic sophistication can replace basic safeguards like position limits, stress testing, and cash reserves.
The question hanging over the industry is whether the next AI fund to hit a rough patch will have those safeguards in place before it's too late.
