Texas Hedge Strategy Amplified Situational Awareness's $35 Billion Loss

Situational Awareness, the AI hedge fund that swelled to $45 billion before collapsing, built its entire book on a "Texas hedge" — a concentrated one-way bet that Wall Street treats as the opposite of risk management.

Leopold Aschenbrenner's Situational Awareness hedge fund, which peaked at $45 billion, built its book on a "Texas hedge" — a concentrated one-way AI bet that amplified July losses to roughly $35 billion, according to the Wall Street Journal.

"When valuations are elevated, and positioning is concentrated, the difference between being on the right and wrong side of a trade can become unusually large — and very quickly," said Bruno Schneller, managing partner at Erlen Capital Management.

The fund ran leverage of up to 400 percent on a concentrated book of long positions in Nebius, SanDisk, CoreWeave and Korea-listed SK Hynix, with shorts including Adobe that moved against it. Nebius fell 17.98 percent between July 1 and Aug. 7. Lenders called collateral, and Aschenbrenner sold most of his public-equity portfolio to Ken Griffin's Citadel at roughly a 10 percent discount, cutting assets from $45 billion to about $10 billion.

The blowup rippled through markets. Jane Street, a major investor, posted a $15 billion loss in July, its first monthly loss in about a decade, while technology hedge funds suffered their worst month since 2008, dropping 7 percent, according to Hedge Fund Research.

A Bet That Was Never a Hedge

The "Texas hedge" traces to ranchers who owned herds of cattle and bought futures contracts betting their livestock would rise in price — the opposite of taking a short position to limit risk. For Aschenbrenner, the 24-year-old former OpenAI researcher, that meant a single, massive wager that artificial-intelligence stocks would keep climbing. The strategy worked spectacularly in the first half of the year, when the fund returned more than 1,000 percent since its July 2024 launch and drew backing from Stripe co-founders Patrick and John Collison, former GitHub chief executive Nat Friedman and investor Daniel Gross.

The reversal came fast. AI-linked equities sold off sharply in late July, with the Nasdaq 100 falling more than 5 percent. Memory and semiconductor names such as SK Hynix and CoreWeave dropped around 50 percent during the month, according to Jane Street's internal communications reviewed by Reuters. With leverage amplifying every move, margin calls forced Aschenbrenner to unwind the bulk of his public stock portfolio at a discount.

Contagion Across the Crowded Trade

The collapse exposed how concentrated the AI trade had become. Technology hedge funds dropped 7 percent in July, the worst month since January 2008, according to Hedge Fund Research. Altimeter Capital Management lost 11 percent, Marshall Wace's Eureka fund fell 6.9 percent, and Balyasny Asset Management and Verition Fund Management each slipped more than 1 percent. Citadel, by contrast, bought the discounted shares and saw its equities fund rise about 14 percent in July.

Jane Street, which had scaled up its stake as the fund performed, acknowledged in an internal memo that July was a "bad month" and said it had closed a significant portion of the risk in the areas where it lost money. The firm's year-to-date trading revenue still exceeds $40 billion, according to people familiar with the matter. Aschenbrenner, who preserved private stakes in Anthropic, Fluidstack and MatX, has told Silicon Valley backers he is not accepting fresh capital for now, even as some line up to offer it.

The episode is a warning about leverage in crowded trades. When one fund's forced selling cascades through shared prime brokers and overlapping positions, yesterday's winners can become today's most concentrated source of risk — a dynamic that turned a "Texas hedge" into a $35 billion loss.

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