09 Sep Situationally Unaware
Posted at 17:38h
in INSIGHTS
Two years ago, Leopold Aschenbrenner was Wall Street’s newest prophet. Then, in his early twenties, having been recently fired from OpenAI over a disputed leak, he published a 165-page essay titled “Situational Awareness: The Decade Ahead”. In it, he forecasts that AI can likely surpass human cognitive abilities across all tasks by 2027. This bold vision inspired him to launch a hedge fund, aptly named Situational Awareness LP, which grew rapidly to $45 billion in assets, backed by the likes of Patrick and John Collison (Founders of Stripe). Then, over the course of a single month, the fund plummeted back down to $10 billion, with leverage having compounded unfavourable market movements.
The trade behind both the rise and the collapse was straightforward on paper. Aschenbrenner ran a classic AI-infrastructure play: long the physical backbone of the AI build-out with a concentrated handful of retail investor favourites such as SK Hynix, SanDisk, Micron and CoreWeave, and funded against short positions in “legacy” software he believed AI would hollow out, Adobe chief among them. Boost the potential returns with leverage, and ordinary moves became extraordinary ones. For most of 2026, that worked in his favour.
In July, the strategy began to unravel. Doubts about AI capex payback sent the infrastructure basket sharply lower; SK Hynix and SanDisk halved from their June peaks, and Micron was not far behind, while Adobe, the short leg meant to offset that risk, rallied instead. Traders call this a “Texas hedge”: a position dressed up as protection that in fact doubles the bet, because both legs move against you together rather than offsetting. With nowhere to hide, the fund’s cushion vanished, assets fell back towards $10 billion within weeks, and margin calls arrived from Goldman Sachs, JPMorgan and Bank of America. Facing forced liquidation, Aschenbrenner sold the public book to Ken Griffin’s Citadel at fire-sale prices.
The irony is clear, but the lesson is not in what you name your hedge fund, but rather it is something much more familiar. One that Long-Term Capital Management learnt during the 1998 Russian debt default, Leopold Aschenbrenner learnt this July and many others in between. Margin calls do not care that the portfolio would have been fine had it not been forced to sell at depressed prices. When working with leverage, not only are your returns amplified but also your risks. And that means timing matters a lot more than for long-only funds like those managed by High Street, which can weather times of turbulence far better. A feature that would have proved useful to Aschenbrenner given the stock rally that followed immediately after the fund was forced out of its public market positions.