From AI Wunderkind to Cautionary Tale: The Fund That Nearly Vanished in Days

Photo: Hanna Pad · Pexels

Leopold Aschenbrenner, a former researcher who worked on AI safety topics, became a well-known figure in investing circles after launching a fund, named Situational Awareness, that grew to manage $45 billion by betting heavily on the artificial intelligence boom. The thesis was ambitious: if AI transformed the economy as fast as some predict, certain positions could generate extraordinary returns.

The trouble with concentrated bets is that they work spectacularly well while the narrative holds, and unravel just as fast when something calls it into question. In a matter of days, the fund lost most of its value, a reminder that conviction, however solid it seems, is no substitute for spreading out risk.

The inevitable comparison to LTCM

It's no coincidence that several analysts have placed this episode alongside Long-Term Capital Management, the late-1990s fund that combined sophisticated mathematical models with extreme leverage and ultimately needed a coordinated rescue to avoid broader contagion in the financial system. The lesson then, and perhaps now, is that even well-reasoned theses can fail when executed with too much concentrated risk.

In this case, it was Ken Griffin, founder of Citadel, who stepped in with a deal to help stabilize the situation, echoing his role in other episodes of financial market stress. The exact outcome of this rescue and its terms remain under scrutiny, but the episode has already become a case study on the risks of betting everything on a single narrative, however promising it may look.