Aschenbrenner's AI Hedge Fund Collapses From $45B to $10B
TL;DR
- CoreWeave, IREN, and SanDisk recovered 27-41% within three sessions of the margin-call date, confirming the collapse was leverage-driven, not thesis-driven.
- SMH-ETF and Nvidia put hedges failed because longs and supposed hedges compressed together in the correlated July AI selloff, providing no real protection.
- Twelve people managing $45B had no risk infrastructure proportionate to a position size larger than most multi-decade multi-strategy firms accumulate.
There is a specific version of this story that keeps showing up in AI finance and it is worth naming plainly: a young researcher with an unusually clear thesis raises capital fast, runs concentrated positions with leverage, prints headline returns, and then the same concentration takes it apart. Forbes reports that Leopold Aschenbrenner's Situational Awareness fund fell from a peak near $45 billion in assets to about $10 billion in a matter of days after highly leveraged bets on AI infrastructure names including Nebius, SanDisk and CoreWeave unwound.
The mechanics matter here. Reporting puts the fund's leverage at up to 400% and describes cumulative returns above 1,000% since its 2024 launch, with 439% gains in the first half of 2026 alone. When July's AI-stock drawdown hit, CNBC reported that margin calls from Goldman Sachs, JPMorgan and Bank of America forced the 25-year-old manager to sell his entire roughly $16 billion public equity book to Ken Griffin's Citadel at what is described as about a 10% discount.
What is left, and this is the part that reframes the 'implosion' headline, is roughly $10 billion in mostly illiquid private stakes anchored by a position in Anthropic reportedly valued around $5 billion. The backer list reportedly included Stripe co-founders Patrick and John Collison, former GitHub CEO Nat Friedman, investor Daniel Gross and quant firm Jane Street. Those are not casual capital sources and they got in early.
One caution on the numbers: the specifics on leverage, on the identity of every backer, and on the exact price Citadel paid are largely sourced to people familiar with the trade rather than to the fund itself, so treat them as reported rather than settled. Still unexplained is how leverage that high on a book that concentrated cleared risk-committee approval at the prime brokers in the first place.
The forward read is less about Aschenbrenner personally and more about the shape of the AI picks-and-shovels long trade. If a $45 billion book got margin-called out of Nebius, SanDisk, CoreWeave and SK Hynix in a single move, everyone else holding similar exposure at any real leverage should assume the exit is narrower than they thought. Citadel reportedly posted its best month in years buying the other side of that trade, which is the cleanest tell about who was actually positioned for this.
What others are reporting
-
Bloomberg Read →
Bloomberg's July 30 breaking confirmation of the $10B floor and Citadel deal, filed six days before Forbes' August 5 post-mortem; first to frame it as an asset-level collapse.
-
CNBC Read →
CNBC's fire-sale dispatch; a companion piece published the same day argued the collapse was internally generated, not driven by broader market conditions.
-
TechCrunch Read →
First to name all three surviving private positions (Anthropic, MatX, Fluidstack) and tie Anthropic's October IPO to a recovery path for the restructured private vehicle.
The fund's losses were amplified by leverage, a common hedge fund strategy of using borrowed money to buy stocks.
-
FinanceFeeds Read →
Documents that CoreWeave, IREN, and SanDisk recovered 27-41% within three sessions of the block sale, making the case that the thesis survived while the leverage structure didn't.
The fastest fortune in hedge fund history did not die of a bad thesis; it died of leverage.
-
SpotGamma Read →
Technical post-mortem: SMH/Nvidia put hedges failed under correlated stress; a 20-27% same-day price snapback confirms the liquidation discount cleared the moment Citadel absorbed the block.
One brutal month in AI infrastructure stocks and leverage reportedly running as high as 4x turned the best-performing large fund in the world into a forced seller of everything.
-
Crypto Briefing Read →
Quantifies the staffing paradox: 12 people managing $45B, more capital than most multi-decade multi-strategy firms, with no risk infrastructure to match.
A fund managed by a team of about 12 people was overseeing more capital than most established multi-strategy firms accumulate over decades.
-
Alternatives Watch Read →
Institutional framing: positions the collapse as a due-diligence stress test for allocators building AI-infrastructure exposure, not just a personal failure story.
Situational Awareness is no longer just an oxymoron; it is a test case fueling a new call for risk management.
Originally reported by forbes.com
Read the original article →Original headline: Aschenbrenner's Situational Awareness Fund Craters From $45B to $10B, Citadel Buys Positions at 10% Discount