Sardor Akhmedov
August 4, 2026
7
min. read
and updated on:
August 4, 2026
Quick Answer
Leopold Aschenbrenner went from a junior researcher on OpenAI's Superalignment team, to a fired whistleblower, to the author of a viral AI manifesto, to the 24-year-old manager of a $45 billion hedge fund named after that manifesto — and then, within a single week in late July 2026, watched roughly three-quarters of that fund's value evaporate in forced margin calls, culminating in a discounted sale of his leveraged stock positions to Ken Griffin's Citadel (CNBC). It is one of the fastest rise-and-fall arcs the AI boom has produced, and it holds a lesson well beyond hedge funds.
Key Facts
- Aschenbrenner joined OpenAI's Superalignment team in 2023, working under Ilya Sutskever.
- OpenAI fired him in 2024 after he wrote a board memo warning about security vulnerabilities, including risk from China; OpenAI said he improperly shared confidential information, a characterization he disputed.
- In June 2024, he published "Situational Awareness," a 165-page essay arguing that AGI could arrive within years and that governments were underestimating the pace of progress.
- In July 2024, he launched the AI-focused hedge fund Situational Awareness with $225 million in seed capital from Stripe co-founders Patrick and John Collison, former GitHub CEO Nat Friedman, and investor Daniel Gross.
- The fund reportedly returned more than 1,000% since inception and peaked at $45 billion in assets in July 2026, reportedly using leverage as high as 400%.
- A late-July 2026 semiconductor sell-off triggered margin calls that forced Aschenbrenner to liquidate leveraged public positions, including SK Hynix and CoreWeave, to Citadel at a discount, dropping the fund to roughly $10 billion.
Act I: The Whistleblower
Aschenbrenner's path to the center of the AI conversation started conventionally enough for a prodigy: he skipped grades in Germany, graduated high school at 15, and by his time at Columbia University was co-founding its Effective Altruism chapter and drawing attention for an academic paper called "Existential Risk and Growth." After Columbia, he spent time at the Future Fund, the now-defunct philanthropic arm of FTX, working under Sam Bankman-Fried before FTX's collapse (CNBC).
In 2023 he joined OpenAI's Superalignment team under Ilya Sutskever, working on the problem of keeping increasingly capable AI systems aligned with human interests. After a hacker breached OpenAI's internal systems, Aschenbrenner wrote a memo to the company's board arguing that its security practices were not strong enough to prevent foreign espionage — naming China specifically as a risk. In 2024, OpenAI fired him, saying he had improperly shared confidential information. Aschenbrenner disputed that framing, maintaining he was raising legitimate security concerns. Former OpenAI safety researcher Scott Aaronson later said he believed Aschenbrenner may have been trying to do the right thing (CNBC).
Act II: The Manifesto
Weeks after leaving OpenAI, in June 2024, Aschenbrenner published "Situational Awareness," a sprawling 165-page essay arguing that artificial general intelligence could arrive within just a few years and that governments were badly underestimating the speed of progress. He wrote that only a few hundred people in the AI world truly understood what was coming, and that if they were even approximately right, "we are in for a wild ride." The essay became required reading across Silicon Valley almost overnight — admired by some as prophetic, dismissed by others as overconfident about both AI timelines and its author's own certainty. Either way, it made Aschenbrenner one of the most talked-about voices in AI, at just over 20 years old.
Act III: The Fund
Aschenbrenner didn't just cash in on the essay's fame with speaking fees or a book deal — he turned it into a hedge fund, literally named Situational Awareness, launched in July 2024 with $225 million in seed capital from Stripe co-founders Patrick and John Collison, former GitHub CEO Nat Friedman, and investor Daniel Gross. He had no prior professional money-management experience. That didn't stop the fund from posting reported gains north of 1,000% since inception, according to the Wall Street Journal, as it rode the broader AI trade. By earlier this July, assets had swelled to a peak of $45 billion. About two-thirds of that was in long and short public-equity bets; the rest was private holdings dominated by a multibillion-dollar stake in Anthropic. Behind the scenes, the fund was reportedly running leverage as high as 400% — a level former bank traders described as making a blow-up a matter of "not if, but when" (CNBC).
Act IV: The Collapse
The reckoning came fast. A sell-off in semiconductor stocks in late July 2026 triggered mounting margin calls, and Aschenbrenner was forced to unwind his entire leveraged public-equity book — including bruised positions in SK Hynix and CoreWeave — selling them to Ken Griffin's Citadel at a discount. In a matter of days, the fund's holdings fell from a $45 billion peak to roughly $10 billion. It became, as CNBC put it, the most high-profile casualty yet of the volatility riding alongside the AI boom. In a detail almost too on-the-nose for fiction, the near-collapse played out the same weekend as Aschenbrenner's wedding to Avital Balwit, chief of staff to Anthropic CEO Dario Amodei (CNBC).
The Unique Takeaway
Most of the commentary on this story lands on the obvious moral: too much leverage is dangerous, and a 24-year-old with no fund-management track record probably shouldn't have been running 400% margin on an already volatile sector bet. That's true, but it's not the interesting part.
The interesting part is that Aschenbrenner's original thesis was, by the fund's own returns, directionally correct. AI capital expenditure exploded roughly the way "Situational Awareness" predicted; the fund's 1,000%+ run before the crash is proof his read on the industry's trajectory wasn't wrong. He saw the future clearly enough to build a $45 billion position on it. What broke him wasn't a bad forecast — it was a fragile structure wrapped around a good one. Being right about where the world is going and being able to survive the volatility of getting there are two entirely different skills, and the market only pays out to people who have both.
That's the takeaway worth carrying into any high-conviction bet, whether it's a fund, a product roadmap, or a company: conviction is not a risk-management strategy. If your structure can't absorb a bad week, it doesn't matter how right your thesis is about the next five years — you won't be at the table to collect on it. The founders and investors who compound over decades aren't necessarily the ones with the boldest predictions; they're the ones who size their bets so a single bad week can't end the story.
Sources
