The Smartest Man in AI Lost $35 Billion in 30 Days

Leopold Aschenbrenner was 25 years old, had never managed a dollar of other people's money before 2024, and ran a hedge fund worth $45 billion. Then July happened. What followed is not just a story about one spectacular collapse. It is a story about what happens when an entire market borrows to believe, and what the wreckage reveals about where we are right now.

But before we get to that, let's take a quick look at the markets and what matters...

3 Movers in 3 Minutes

1. Palantir's "otherworldly" quarter. Palantir Technologies (PLTR) surged 29.5% on Tuesday after posting 93% revenue growth to $1.94 billion, with U.S. commercial revenue up 149%. CEO Alex Karp called demand for AI sovereignty "otherworldly" and raised full-year revenue guidance to $8.15 billion. It was the stock's best single day in over two years.

1. Caterpillar crosses $20 billion for the first time. Caterpillar (CAT) reported record quarterly revenue of $20.5 billion, up 24% year-over-year, with adjusted EPS of $8.17 crushing the $6.19 consensus. The backlog surged 92% to $72 billion, driven by North American construction demand and data center power generation projects. Shares rose roughly 6%.

1. Oil crashes for a second straight day on Iran deal hopes. WTI crude plunged to $75 as U.S. Treasury Secretary Scott Bessent said a deal to reopen the Strait of Hormuz could come "today or tomorrow." Qatar confirmed a de-escalation proposal was being circulated. Oil is now down over 10% in two sessions, its steepest two-day drop since December 2025. Iran, however, denied any direct talks with the U.S.

And with that out of the way, let's get to today's big story: the spectacular collapse of the AI era's most famous hedge fund.

The Sip

The Essay That Became a Balance Sheet

In June 2024, a 23-year-old former OpenAI researcher named Leopold Aschenbrenner published a 165-page essay titled "Situational Awareness: The Decade Ahead." It argued, with the urgency of a Cold War dispatch, that artificial general intelligence was arriving faster than anyone in government or finance understood. It became required reading in Silicon Valley within weeks.

One month later, Aschenbrenner did something unusual for a 23-year-old essayist. He launched a hedge fund. He named it Situational Awareness, after the essay, and started with roughly $225 million in capital. His backers read like a tech aristocracy: Patrick and John Collison of Stripe, former GitHub CEO Nat Friedman, and, notably, quantitative trading giant Jane Street.

The thesis was simple. If superintelligence was coming, the companies supplying its physical backbone, semiconductors, memory chips, data centers, power infrastructure, would become the most valuable enterprises on Earth. Aschenbrenner bet accordingly. Long on SK Hynix. Long on CoreWeave. Long on Nebius Group. Long on Micron Technology (MU). Short on legacy software names like Adobe (ADBE) that he believed AI would render obsolete.

And for two years, he was spectacularly right. It reported cumulative returns of 1,550% since inception. By the first half of 2026, the fund was up 439%. Assets under management ballooned to $45 billion by early July, making Aschenbrenner, at 25, one of the most successful new fund managers in a generation.

"A lot of people saw this blow-up as a matter of not if, but when."

Then July happened.

The Month That Ate $35 Billion

The AI infrastructure trade, which had powered Situational Awareness to extraordinary heights, violently reversed in mid-July. The trigger was familiar by now: a Chinese AI lab called Moonshot released a frontier model that performed near the level of the best American systems at a fraction of the training cost. Semiconductor stocks, already stretched by 18 months of relentless buying, cracked.

Nebius Group, Sandisk, and Micron, three of the fund's top disclosed positions, each shed more than 35% during July. CoreWeave, the GPU cloud computing company, dropped alongside them. Simultaneously, the short book went wrong. Adobe rallied. Legacy software, far from dying, was finding new life integrating AI into existing products rather than being replaced by them.

But what turned a painful drawdown into an existential crisis was not the thesis. It was the leverage.

According to multiple reports, Aschenbrenner was running some positions with up to 400% leverage, meaning for every dollar of capital, four dollars were deployed. When the underlying stocks moved 20-30% against him, the losses were multiplied four-fold. Prime brokers Bank of America (BAC), Goldman Sachs (GS), and JPMorgan Chase (JPM) issued margin calls in rapid succession. The fund needed cash it did not have.

On July 30, Situational Awareness sold the bulk of its public stock portfolio to Ken Griffin's Citadel at a roughly 10% discount. The portfolio was valued at approximately $16 billion. The fund's assets crashed from $45 billion to roughly $10 billion. A 67% loss in a single month.

Aschenbrenner, who had reportedly invested almost his entire personal net worth in his own fund, was getting married the same weekend the fire sale closed. His fiancee is the chief of staff to Anthropic CEO Dario Amodei. The wedding, as one person close to the situation put it, went ahead as planned.

The Canary, Not the Mine

Here is where the story gets uncomfortable. Because Aschenbrenner's collapse was not unique. It was just loud.

FINRA data shows that outstanding margin debt at U.S. brokerages surged 77% between April 2025 and June 2026, reaching an all-time high of $1.502 trillion. To put that in context, every previous time outstanding margin debt rose by 65% or more in a compressed timeframe, it preceded a major market event. An 80% spike between March 1999 and March 2000 preceded the dot-com crash. A similar surge between 2006 and 2007 preceded the financial crisis. And a sharp run-up in 2020-2021 preceded the 2022 bear market that wiped trillions from growth stocks.

Aschenbrenner was not an outlier. He was a concentrated, hyper-visible version of what thousands of funds and millions of retail investors are doing right now: borrowing money to own the AI trade at record valuations, while the S&P 500 sits at all-time highs.

The irony is striking. Aschenbrenner's thesis about AI may prove correct over the next decade. The companies he owned may indeed become some of the most valuable in history. But in the public markets, being right about the destination does not help if the leverage kills you before you arrive.

As one former Wall Street trader told CNBC: "Maybe his views on AI are correct in the long run, but in the public markets, you have to be prepared for the short-term."

There is another detail worth noting. Before Aschenbrenner launched his hedge fund, he worked at FTX, where he helped Sam Bankman-Fried run a philanthropic fund out of a Bahamas penthouse. That fund, too, was built on leverage and conviction. That, too, ended in ruins. One blow-up can be bad luck. Two starts to look like a pattern.

The MarketSips Takeaway

Until then, sip slowly!

The Market Sip Desk