Key Highlights
- Situational Awareness sold its public equities portfolio to Citadel after banks issued margin calls on a leveraged book of AI-related stock positions
- The fund's holdings included chipmaker Sandisk, which has fallen sharply over the past month alongside a broader decline in AI infrastructure stocks
- Founder Leopold Aschenbrenner said the fund remains up roughly 80% for the year and still holds unleveraged public stocks and a stake in Anthropic
- The episode has drawn comparisons to broader risks facing AI Data Center developers, who are financing large-scale buildouts with substantial debt
Leveraged AI Bet Turns Into Forced Sale
Situational Awareness, the hedge fund founded by former OpenAI researcher Leopold Aschenbrenner, sold most of its publicly traded stock holdings to rival Investment firm Citadel after banks called in loans backing a leveraged portfolio of AI-related positions. The fund had built its strategy around a thesis that certain AI infrastructure and semiconductor stocks would continue climbing, a view that had driven strong returns earlier in the year.
When those stocks instead declined sharply over the past month, the fund's use of Debt to amplify its bets turned a market downturn into a Liquidity crunch. Of the 29 US-listed stocks the fund reported holding at the end of March, the average share price has fallen roughly 21% over the past month, with large holding Sandisk down about 45%.
Fund Retains Unleveraged Positions and Anthropic Stake
Aschenbrenner has said the fund remains up approximately 80% for the year despite the forced sale. Situational Awareness still holds unleveraged public stocks and a stake in AI lab Anthropic, which has not yet gone public, giving the fund the ability to weather further market swings without needing to liquidate those holdings.
The episode illustrates a broader risk facing highly leveraged bets on the AI buildout: directional views on the technology's growth can prove correct while poorly timed Leverage still produces losses. Builders of AI data centers, who are expected to invest heavily in infrastructure through the end of the decade, face a similar mismatch between long-term financing commitments and shorter-term swings in asset values.
Episode Highlights Risks in AI Infrastructure Financing
Potential risks facing highly leveraged AI infrastructure investments include equipment becoming obsolete before financing is repaid, efficiency gains reducing Demand for some data center capacity, and competition from lower-cost AI models gaining traction. While the broader view that AI will see continued adoption may prove accurate, the scale of returns, which specific companies benefit, and the timing could vary considerably, creating risk for investors who have taken on debt to finance their positions.






Please wait processing your request...