
Prodigy Fund Manager's $45B Empire Collapses to $10B — But His Anthropic Bet Remains Intact
Key takeaways
- Situational Awareness sold most of its public equity holdings to Citadel after AI infrastructure stocks declined over 30%, shrinking fund assets from $45B to roughly $10B
- The fund retains a $5 billion private stake in Anthropic, which could go public as early as October at a valuation exceeding its last $65B round
- Founder Leopold Aschenbrenner, 25, had no prior trading experience before launching the fund in 2024 following his dismissal from OpenAI
Leopold Aschenbrenner, a 25-year-old former OpenAI researcher with no prior trading background, launched Situational Awareness in 2024 on the strength of a compelling and highly public investment thesis: that the scaling of artificial intelligence would demand enormous investments in semiconductors, computing infrastructure, memory chips, and clean energy. The fund attracted serious institutional backers, including quant-trading giant Jane Street, Stripe co-founders Patrick and John Collison, and tech executives Daniel Gross and Nat Friedman of Meta. At its peak, the fund reportedly swelled to $45 billion in assets under management — a staggering figure for a debut fund run by someone who had never traded professionally before.
Aschenbrenner's background is anything but ordinary. He enrolled at Columbia University at age 15 and graduated as valedictorian at 19, subsequently joining OpenAI's experimental 'superalignment' team in 2023. That team, originally led by co-founder Ilya Sutskever and researcher Jan Leike, was tasked with ensuring AI systems remained aligned with human values at scale. Aschenbrenner was dismissed from OpenAI in 2024, with the company citing an improper disclosure of internal information. The fallout scattered that team across the AI landscape — Sutskever went on to found Safe Superintelligence, Leike joined Anthropic, and Aschenbrenner launched his hedge fund.
For most of its first year, Situational Awareness appeared to be a runaway success. The fund reportedly returned 439% through June 2025, according to the Financial Times, riding a wave of enthusiasm around AI infrastructure stocks. But conditions shifted sharply as public market investors grew skeptical that the massive capital expenditures pouring into AI hardware were generating near-term revenue. Stocks at the core of the fund's thesis — including memory chip makers SK Hynix and Sandisk, clean energy provider Bloom Energy, and neocloud operator Nebius Group — each dropped by more than 30% over a single month. The fund's use of leverage amplified those losses considerably.
Rather than quietly reducing exposure, Aschenbrenner doubled down in a July 24 investor letter, calling the downturn one of the best buying opportunities since early 2024 and soliciting fresh capital commitments beginning August 1. According to Bloomberg, that appeal failed to generate the inflows he had anticipated. The fund ultimately offloaded the bulk of its public equity positions to Ken Griffin's Citadel — a firm well-known for acquiring distressed or leveraged holdings at opportunistic prices — after which Situational Awareness' total assets fell from approximately $20 billion to around $10 billion.
Critically, the fire sale did not include the fund's private holdings. Situational Awareness retains a stake in Anthropic currently valued at approximately $5 billion, a position many analysts view as the fund's most significant remaining asset. Anthropic was last valued at $65 billion following a Series H funding round in May, with reports suggesting the company could pursue an IPO as early as October at an even higher valuation. If that offering materializes at favorable terms, the resulting windfall could substantially offset the losses Situational Awareness incurred in public markets. The fund also holds private positions in chipmaker MatX and AI data center startup Fluidstack, which was reportedly exploring a new funding round at an $18 billion valuation earlier this year.
The bigger picture
The Situational Awareness story is a cautionary tale wrapped inside a legitimately interesting investment thesis. Aschenbrenner was not wrong about the structural importance of AI infrastructure — the buildout of compute, memory, and energy capacity remains central to the ambitions of every major AI lab. What the fund underestimated, or perhaps just ran out of time to wait on, was the gap between long-term structural demand and short-term market patience. Public equity investors are notoriously fickle about infrastructure spending that doesn't convert quickly to earnings, and leveraged bets against that sentiment can go wrong fast.
The Citadel angle here is quietly revealing. Ken Griffin's firm has long operated as a kind of financial shock absorber — stepping in to acquire distressed positions from overleveraged players, often at a discount, and then waiting for the thesis to play out. The fact that Citadel already held similar AI infrastructure positions before acquiring Situational Awareness' book suggests Griffin agrees with the underlying bet. In effect, the same trade is continuing — just with different capitalization and a much longer time horizon. That's a meaningful signal for anyone watching the AI infrastructure sector.
What readers should watch closely is the Anthropic IPO timeline and what it means for Aschenbrenner personally and professionally. A successful Anthropic public offering — especially at a valuation north of $65 billion — would validate his private-market instincts even as his public-market execution is being scrutinized. It also raises a broader question about whether first-time fund managers with exceptional intellectual credentials but no trading experience should be managing tens of billions of dollars. Situational Awareness may survive this episode, but the story of how it manages its private portfolio from here will define Aschenbrenner's long-term reputation in finance.
We're covering this story because it sits at the intersection of two conversations we've been tracking closely at LagPing: the long-term viability of the AI infrastructure investment thesis and the rise of AI-native voices entering financial markets. Leopold Aschenbrenner isn't just a hedge fund manager — he's become one of the most prominent public intellectuals arguing that AI development will reshape capital allocation globally, and that makes his fund's performance a kind of real-money referendum on those ideas. The dramatic swing from $45 billion in assets to $10 billion within a matter of weeks is the kind of event that will shape how institutional investors approach AI-adjacent funds going forward. We also think the Anthropic subplot is genuinely underappreciated — a $5 billion private stake in a company that may IPO within months is not a footnote, it's potentially the whole story. As always, we'll keep following this as it develops.
As an Amazon Associate, LagPing earns from qualifying purchases. Product links are affiliate links.
You might also like

Musk's Empire Loops Tighter: SpaceX Pours $329M Into Tesla Megapacks for AI Data Centers
3d ago

Intelligence lands $7.9M seed funding as its human-taste platform hits $60M ARR
4d ago

AI-Focused Hedge Fund Founded by Ex-OpenAI Wunderkind Dumps Entire Stock Portfolio
Jul 31

Evil Empire and Konami Team Up to Resurrect Classic Castlevania Side-Scrolling Action
Jul 17