The timing almost sounds fictional.
Leopold Aschenbrenner was reportedly in Carmel, California, preparing for a multi-day wedding celebration, complete with a pre-wedding colloquium on big ideas, just as his AI-focused hedge fund was unraveling. The 24-year-old had built a firm that at one point managed roughly $45 billion, riding the AI boom with heavily leveraged bets on AI-related stocks.
The irony is hard to ignore.
His fund was called Situational Awareness.
Yet the biggest thing it appears to have underestimated wasn’t AI.
It was leverage.
According to reports, the fund fell about 67% in July, triggering what Aschenbrenner himself reportedly compared to a “bank run.” To survive, billions of dollars in public equities were sold, much of them to Citadel at discounted prices, while a planned sale of a $3.5 billion Anthropic stake was ultimately avoided. He later told investors the firm’s risk management would be overhauled.
That’s the statistic people should focus on.
Not 67%.
Leverage.
For months, many investors assumed the biggest risk in AI was whether the technology would live up to the hype.
This episode suggests a different risk.
What happens when the people financing the AI boom are themselves financed with borrowed money?
The mechanism is straightforward.
AI stocks fall.
Lenders demand more collateral.
Funds are forced to sell.
Those sales push prices lower.
Lower prices trigger more margin calls.
Suddenly, it isn’t fundamentals driving the market.
It’s liquidity.
That’s how bubbles usually crack.
Not because everyone suddenly changes their mind.
Because somebody is forced to sell.
This also connects with a pattern that’s becoming harder to ignore.
Long-term Treasury yields are sitting near their highest levels since 2007.
Private credit is showing rising stress.
Capital is becoming more expensive.
IPO windows are narrowing.
Now one of the highest-profile AI investment firms has discovered that being right about AI doesn’t protect you from being wrong about financing.
That’s the part I keep coming back to.
Aschenbrenner may still prove right about AI over the next decade.
The market has a long history of destroying investors who were directionally correct but financially overextended.
Being early with leverage often looks exactly like being wrong.
The bigger warning isn’t about one hedge fund.
It’s that the AI boom is entering the stage where the cost of money matters just as much as the quality of the technology.
I’ve said all along OPEN AI is the weak link in AI.
They are the AOL of this tech cycle:
First to market and big but garbage. https://t.co/8LFqdJEcVB
— QE Infinity (@StealthQE4) August 1, 2026