Leopold Aschenbrenner lost 30 billion dollars by betting too

Imagine being 24 years old and managing a fund that ballooned from $1.5 billion to over $450 billion—hitting a peak of nearly $100 billion with leverage—only to watch 67% of it vanish in a single month. That is the absolute carnage Leopold Aschenbrenner just experienced. He was the "golden boy" of the AI prediction world, a former OpenAI researcher who wrote a massive 165-page manifesto on "Situational Awareness," claiming AGI was imminent and that the world was about to shift from a software problem to a physical infrastructure problem (chips, power, data centers).

He didn't just write about it; he bet the house on it. He started a fund called Situational Awareness and went on a shopping spree for AI infrastructure stocks like Micron and SanDisk, while simultaneously shorting traditional software companies like Adobe and Figma. For a while, he looked like a genius. The "AI infrastructure" trade was the hottest thing on Wall Street, and banks like Goldman Sachs and JPMorgan were practically shoving leverage down his throat. At one point, he was borrowing $3 for every $1 of his own capital.

But here is where the "AI genius" part collided with the "reality of finance" part. Aschenbrenner essentially executed a "Texas Hedge." He thought he was diversified because he was long on hardware and short on software, but both bets relied on the exact same premise: that AI would evolve exactly as he predicted and the market would price it exactly as he imagined.

When the market shifted in July, he got hit from both sides. His long positions in SanDisk and Micron tanked, while the software companies he shorted started rebounding. In a leveraged portfolio, that's a death spiral. The margin calls started hitting, and because his positions were so concentrated and well-known, the rest of the market smelled blood. Traders saw the liquidity shifts and realized a giant was being forced to liquidate.

The desperation peaked around July 29. He started offloading Anthropic shares at a 20% discount with a 12-hour deadline for buyers—the financial equivalent of a "fire sale" at 3 AM. He ended up dumping a massive amount of his leveraged positions to Citadel just to keep his head above water.

It is a brutal lesson in AI workflow and market psychology. Being right about the technology (AGI coming) doesn't mean you're right about the timing or the pricing of the assets. He tried to use a technical deep dive to master the stock market, but he forgot that leverage doesn't care about your 165-page thesis. He went from being the prophet of the AI era to a cautionary tale about what happens when you treat a hedge fund like a science experiment.