AI hedge fund collapsed: Lost 67% in one month

From thousand-percent returns to emergency sale

Few hedge funds have traced a more dramatic arc than Situational Awareness. When Leopold Aschenbrenner — a former researcher at OpenAI — launched the fund in July 2024, the starting point was a $225 million investment and a clear thesis: the development of artificial intelligence would require a massive buildout of physical infrastructure, from semiconductors and memory chips to computing power and energy production.

The strategy delivered impressive results. According to Nasdaq Markets, the fund's returns exceeded 1,000 percent since inception, and in the first half of 2026 alone the net result was 439 percent. At its peak, Situational Awareness managed an estimated $20 to $45 billion in assets under management.

Being right about a long-term megatrend is not enough — risk management determines whether you survive the correction.

Leverage and concentration: A dangerous combination

Behind the staggering numbers lay a high-risk structure. The fund leveraged positions at up to four times equity, according to research material cited by Nasdaq Markets. The portfolio consisted of around 42 stocks, but was heavily concentrated in a handful of AI infrastructure companies: SK Hynix, Micron, SanDisk, CoreWeave, Nebius, and Bloom Energy.

In addition, the fund held short positions in software companies such as Adobe, based on a thesis that AI would erode their business model. Those positions moved in the wrong direction as the stocks continued to rise.

July 2026: The AI sector cracks

July 2026 marked a turning point for AI stocks globally. A broad sector correction hit Situational Awareness directly. Key positions including Nebius, SanDisk, Micron, and CoreWeave fell between 20 and more than 35 percent during the month, according to the research material.

With four times leverage, the losses quickly became catastrophic. Margin calls poured in from the prime brokers — Bank of America, Goldman Sachs, and JPMorgan Chase. Under acute liquidity pressure, the fund was forced to divest the bulk of its listed equity portfolio in an emergency transaction to Citadel, estimated at around $16 billion in assets.

A symptom of something larger

The collapse of Situational Awareness is not merely a story about one fund's missteps. According to expert commentary cited by Nasdaq Markets, it mirrors a broader systemic vulnerability: when too much capital flows into the same positions and is financed with borrowed money, the investment horizon can collapse overnight.

Ray Dalio, chairman of Bridgewater Associates, has compared the current AI investment climate to the speculative waves that preceded the market crashes of 1929 and 2000. His warning that the AI bubble could burst appears more relevant in the wake of July's events.

For Norwegian investors and asset managers, this serves as a reminder that exposure to the AI infrastructure sector — whether through direct equities or funds — carries concentration and leverage risks that are not always visible in quarterly return figures.

Right thesis, wrong positioning

The paradox of Situational Awareness is that Aschenbrenner's underlying analysis is likely correct: the AI revolution demands enormous quantities of semiconductors, computing power, and energy infrastructure. But being right about the trend is not the same as being right about the timing and size of the positions.

The collapse has been described as a cautionary tale for the AI boom — a case in which superior thematic insight was undermined by aggressive leverage, a lack of diversification, and a failure of risk management. Wall Street will be studying this case for a long time to come.