There is a certain kind of arrogance that comes with real intelligence, and Leopold Aschenbrenner has always had a lot of both. He graduated from Columbia as valedictorian when he was 19. He worked at one of the world’s most secret AI labs by the time he was 21. He wrote essays about the coming of artificial general intelligence when he was 23 years old, with all the confidence of someone who had already seen the whole map. All of that moved so quickly, and then July 2026 stopped everything in an instant.
At the beginning of this month, Aschenbrenner’s hedge fund, Situational Awareness LP, had grown to about $45 billion. The fund was based on a belief that, as AI got stronger, it would need a lot more chips, memory, data centers, and energy infrastructure. This belief felt more like a worldview than a financial plan. It was an interesting thought. Some very important people, like Patrick and John Collison, Daniel Gross, and Nat Friedman, backed it. The numbers seemed to match up for a while.
Then the places started moving in the wrong direction. Some stocks in AI infrastructure, like SK Hynix, went down. Short bets against software companies, like Adobe, went against the fund very quickly. When prices went up, leverage made the gains bigger. On the way down, it did what leverage always does: it bucked the trend. Bank of America, Goldman Sachs, and JPMorgan Chase were among the top brokers who had to sell the fund’s holdings before the markets opened in order to meet margin requirements. Citadel, led by Ken Griffin, stepped in and bought up a lot of the publicly traded portfolio. It didn’t really feel like an acquisition, more like a controlled evacuation.
Aschenbrenner was fired from OpenAI in April 2024. The official reason given was that he improperly shared internal information, which he denies. His story is based on a non-confidential planning document that was shared with some outside researchers to get their thoughts. At most companies, this wouldn’t be a big deal. From the outside, it’s harder to judge because he had written a memo to OpenAI’s board a few months before, warning about the risk of foreign espionage and saying the company’s security was dangerously weak. He says that memo really made things tense. OpenAI says that firing him had nothing to do with it. It’s possible for both to feel true while still leaving questions.
The essay came next. “Situational Awareness: The Decade Ahead” was 165 pages long and written with a sense of urgency that can make you think of it as either bold or scary, depending on your point of view. Aschenbrenner said that AI systems would be doing their own AI research by the year 2027. Then, hundreds of millions of AGIs could pack ten years’ worth of algorithmic progress into less than a year. In his writing, he said that the US should be very worried about China and Russia getting these abilities. People in the tech industry, the government, and the media all read the essay. It was also the first document that his investment firm used.
The exact size of the losses and the amount of new money the fund was trying to raise are still not clear. People with knowledge of the situation have said that it is still changing as this is written. Reports said that a stake in Anthropic was being shopped around. This is interesting because Aschenbrenner is engaged to Avital Balwit, who is the chief of staff to Anthropic CEO Dario Amodei. The wedding is said to be this weekend. To say the least, the timing is tricky.
The author is 25 years old. In the event that the fund fails, it will not change what the thesis was trying to say. The building of infrastructure that he said would happen is still happening. Chips are still being ordered, data centers are still being built, and money is still coming in. It’s possible that the market moved faster than he thought it would, or that the risk of the trade was too high for the thesis to pay off in time. Investors have been right about direction but wrong about when to buy and sell. Now the question isn’t just whether he was right about AGI. The question is whether the fund will last long enough to find out.
