Aschenbrenner's AI fund loses 67% in July, sells holdings to Citadel

Situational Awareness, the AI-focused hedge fund founded by former OpenAI researcher Leopold Aschenbrenner, is down about 67 percent in July after heavy losses on AI stocks, according to a letter the firm sent investors Thursday. The deep losses prompted the firm to sell the bulk of its stockholdings to Ken Griffin's Citadel as it raced to gather cash to cover margin calls from lenders.

"We let you down this month," Aschenbrenner wrote in the investor letter, the contents of which were described to The Wall Street Journal.

The fund reached a deal late Wednesday to sell $3.5 billion of its stake in privately held AI company Anthropic to a consortium led by Greenoaks and Sequoia Capital, but backed out Thursday morning. Despite July's losses, the fund remains up about 80 percent on the year, the letter said. Through the end of May, Situational had gained about 270 percent. The fund had grown to as much as $45 billion before the AI sell-off, after gaining roughly 439 percent after fees through the end of June.

Situational had amassed well over $20 billion in assets under management since its founding roughly two years ago, making it one of the fastest-growing firms in recent years. Aschenbrenner, who had no professional investing experience when he started with a few hundred million dollars, was seen by some as an AI oracle whose movements other investors tracked closely.

From 439% Gain to Margin Calls

Aschenbrenner partially blamed short sellers who targeted the firm's positions for worsening the fund's losses, the letter said. He compared Situational's experience to a bank run and told investors the firm had removed all leverage from the portfolio.

The fund's publicly traded AI infrastructure holdings — including Bloom Energy, CoreWeave, Nebius, Sandisk and Micron — were hit hard in the current rout. The selloff coincided with the U.S. IPO of SK Hynix, one of the fund's largest holdings, while a broader unwind of leveraged positions triggered a steep decline in South Korean stocks.

The fund's Q1 filings showed it took a new $43.9 million position in T1 Energy, grew its Bitdeer Technologies stake by 92.4 percent, acquired 3.39 million shares of HIVE Digital Technologies, and trimmed its Solaris Energy Infrastructure position by 40.8 percent. Those moves illustrate the breadth of AI-adjacent infrastructure bets that have now come under pressure.

The AI Infrastructure Selloff

The fund's collapse comes as investors have been spooked by rising AI costs at big technology firms even as they report strong earnings. Negative cash-flow reports from Alphabet and Tesla last week triggered a bout of selling in AI-linked stocks, with chip stocks also coming under pressure as investors questioned high valuations. The tech-heavy Nasdaq-100 dropped 10 percent from its early June peak on Wednesday.

Microsoft rose 9 percent in premarket trading Thursday after the company forecast current-quarter sales and cloud growth that beat expectations, giving a capital expenditure outlook below Wall Street estimates. The rebound in tech stocks Thursday helped ease some concerns about heavy corporate AI spending, and many of the stocks in Situational's portfolio rebounded sharply as the news of the Citadel sale broke.

In a July 24 letter to investors, the fund described the tech rout as a buying opportunity and pointed to a potential Anthropic IPO as a forward driver for the sector. "At times we call out opportunities that seem like a particularly good time to add funds, if you have been waiting for one," the letter closed.

Aschenbrenner graduated as Columbia University's valedictorian at 19 before joining OpenAI's Superalignment team. He was fired in 2024 after OpenAI said he improperly disclosed internal information, a characterization he disputes. He is married to Avital Balwit, chief of staff for Anthropic CEO Dario Amodei.

The firm still holds a portfolio of investments in private companies, including the Anthropic stake. The aborted $3.5 billion sale highlights the liquidity stress in private AI marketplaces as well as public markets. The fund's trajectory — from a few hundred million dollars to more than $20 billion in two years, then a 67 percent monthly drawdown — illustrates the extreme concentration risk in AI-focused investing. As the sector's most visible leveraged player unwinds, other funds with similar positions may face pressure to reduce exposure, potentially extending the selloff in AI infrastructure names.

This article is for informational purposes only and does not constitute investment advice.