At the end of July 2026, the most dramatic collapse ever recorded in an AI-focused hedge fund took place. Situational Awareness, founded by Leopold Aschenbrenner, a 25-year-old former OpenAI researcher, had grown to manage peak assets of approximately $45 billion and boasted stellar returns: +439% from the start of 2026 through the end of June. However, a bearish sell-off in AI “picks and shovels” stocks, ranging from memory chips to data center infrastructure, triggered the fund’s leverage to spike to around 4x, and within a few trading sessions, it wiped out 67% of the equity portfolio, triggering a series of margin calls. In an extraordinary transaction, nearly the entire public portfolio, valued at approximately $16 billion, was sold in a single block to Citadel.
Aschenbrenner admitted the failure (“We let you down this month”) and promised to eliminate leverage, leaving investors in the dark about the next steps. For example, after the crash, the fund was still formally in the black on a year-to-date basis (+80% YTD), a sign of the magnitude of its previous performance. The collapse, amplified by algorithmic selling and automatic stop-loss orders, shook the entire AI sector: in the weeks that followed, semiconductor and cloud stocks experienced high volatility, while investors assessed the potential spillover to other specialized funds and the impact on AI startups and suppliers.
IA Situational Awareness Fund Profile
Situational Awareness LP was a hedge fund specializing in artificial intelligence, launched in the summer of 2024 by former OpenAI researcher Leopold Aschenbrenner. The name derives from one of his controversial essays on the exponential development of AI. Aschenbrenner, a 25-year-old German citizen and former OpenAI employee, enjoyed a following among certain venture capital funds and tech leaders. Among the initial backers were brothers Patrick and John Collison (founders of Stripe), executive Nat Friedman (former CEO of GitHub), and investor Daniel Gross. Operationally, the fund adopted a long/short strategy heavily focused on securities tied to the physical infrastructure of AI, such as chip manufacturing, RAM/flash memory, data centers, electricity, and related investments.
At the height of its rise, in July 2026, Situational Awareness managed approximately $45 billion in assets from institutional investors and family offices. In the first six months of 2026 alone, the fund had posted a net gain of +439%, thanks to more than doubling returns on its key technical bets: the SEC Form 13F filing as of the end of March 2026 showed a highly concentrated portfolio: Bloom Energy (22.8% of the portfolio), SanDisk (18.8%), CoreWeave (14.4%), IREN and Core Scientific (both ~10%), with holdings large enough to control 4–8% of the outstanding shares of these small- and mid-cap companies. At the same time, the fund held large options contracts, such as significant put exposures in the semiconductor sector: ~$5.0 billion in notional value on the SMH ETF and $3.6 billion on Nvidia, according to unofficial data. This AI infrastructure strategy was designed to capitalize on any acceleration in the semiconductor, memory, and cloud computing cycles, assuming that the adoption of next-generation AI would require unprecedented investment.
The fund’s strategy focused on the so-called “picks-and-shovels” of artificial intelligence, favoring providers of the infrastructure, chips, and energy necessary for AI development, while taking bearish positions on the major software giants, which the market already considered overvalued. For about eighteen months, this approach proved successful, fueling the fund’s extraordinary growth. Before the crash, Situational Awareness managed approximately $45 billion, posting an exceptional return of +439% in the first half of 2026, compared with an average hedge fund performance of between 7% and 8%. This growth was driven by leverage of about 4 times the capital, well above the industry average, and a heavy concentration in artificial intelligence-related companies, including Nvidia, AMD, Intel, SanDisk, CoreWeave, and Bloom Energy. By the end of July, however, the fund had lost about 67% of its value, a much steeper decline than the losses recorded during the same period by global hedge funds and the technology sector.
Governance was extremely concentrated under the master-feeder model, with a master fund registered in Delaware. Governance was highly concentrated: Leopold Aschenbrenner and a few key partners held operational control, without resorting to formal lock-up or redemption restriction mechanisms, despite the high level of leverage. Thanks to agreements with major prime brokers, including Goldman Sachs, JPMorgan, and Bank of America, the fund was able to operate with a significant level of debt secured by its own assets. Capital came exclusively from accredited investors, with no public offering, while transparency toward the market was limited to mandatory filings submitted to the SEC, such as Form 13F for U.S. equity holdings and Form 13D for significant stakes. Among these was an 8.2% stake in Core Scientific, disclosed in August 2025. In terms of communication, the fund maintained an extremely low profile: official information was scarce, and the first detailed indications of its 2026 performance emerged only through a letter to investors released in late July and subsequently reported by the Financial Times and Reuters.
Timeline of the Implosion
Situational Awareness’s collapse unfolded over the course of just a few days. After posting a return of +439% in the first half of 2026, the fund began showing the first signs of trouble on July 24, when Leopold Aschenbrenner reassured investors by describing the market correction as simply a buying opportunity. A few days later, however, the sharp decline in semiconductor and artificial intelligence stocks, combined with increased margin calls from prime brokers, forced the fund to liquidate a large portion of its positions.
On July 30, Situational Awareness sold approximately $16 billion in assets to Citadel, drastically reducing its leverage and seeing its value plummet by 67%. The following day, Aschenbrenner publicly took responsibility for the losses, announcing that the fund would abandon its highly leveraged strategy while confirming that it would continue operations. The speed with which the crisis unfolded surprised observers. In just a few days, the fund went from record returns to an extraordinary liquidation of its portfolio. Citadel’s intervention as the sole buyer prevented a disorderly market sell-off and helped contain panic, leading to a sharp rebound as early as August 1 in many of the artificial intelligence-related stocks hardest hit by the sell-off.
The collapse of Situational Awareness is part of a broader wave of nervousness surrounding tech and AI stocks. In the preceding months, global hedge funds had celebrated another record half-year thanks to AI: according to Goldman Sachs, the average return for hedge funds in the first half of 2026 was approximately +7%, well above the historical average of 4%, driven by a boom in allocations to AI “picks-and-shovels”: managers systematically shifted capital from semiconductors to energy companies and data centers, and finally to memory stocks.
However, in July, part of this excessive positioning was unwound. The Nasdaq 100 index lost over 7% last month, while tech stocks in Asian markets posted their worst-ever declines. JPMorgan noted that the losses stemmed from crowded positions: when prices turned, many were forced to exit at an unfavorable price. A geopolitical factor further exacerbated the situation: the war in the Middle East caused a sharp rise in oil prices during the final weeks of the month. The rise in global uncertainty and energy prices also triggered a sell-off in traditional stocks such as chips, which are well-known and significant energy consumers, and reinforced investors’ preference for low-risk investments; all of this combined helped intensify the pressure on the long positions in semiconductors and data centers held by Situational Awareness.
Regulatory and Legal Issues
At this time, no investigations have been officially announced by regulatory authorities, and the documents filed with the SEC confirm the fund’s major holdings in publicly traded companies, including a position of approximately $213 million in Core Scientific, disclosed in a filing for the fourth quarter of 2025. There is no public evidence that the fund has exceeded regulatory limits; however, it is plausible that the SEC is monitoring the situation, as, in similar cases, regulators investigate possible anomalous trading in conjunction with key public announcements. Some observers have noted unusual trading volumes in AI-sector stocks in the hours leading up to the trade with Citadel, although no concrete allegations have been made. Furthermore, the major prime brokers involved (Goldman, JPM, BoA) must cover margin calls: a review of their procedures will be inevitable.
The primary source of information on performance has remained the letter to investors; there is no obligation for a private hedge fund to publish real-time loss statements. In the coming months, however, the credit ratings of counterparties that had exposure to the fund could be downgraded, although this will depend on the collateral provided.
Risk of Contagion and Impact on the AI Sector
The rapid sell-off of Situational Awareness has raised fears of contagion among AI and tech investors. The good news is that Citadel’s intervention, with a single counterparty ready to “take on” the entire book,limited the fire-sale of stocks like Core Scientific, which accounted for 8% of the market supply: a fragmented sell-off could have triggered days of price-insensitive selling in already weak stocks. On the other hand, the crisis has already triggered a phase of symmetrical deleveraging among other managers who held similar positions. In the days that followed, in fact, various AI-driven long/short funds reduced their exposure to chips and cloud stocks, following a shared trajectory as described by JPMorgan. Furthermore, institutional investors temporarily shifted capital out of the tech sector in search of safer assets, as suggested by the spike in Treasury yields.
The impact, however, extends beyond Situational Awareness and could reach far beyond the fate of the fund itself. As for the AI startup ecosystem, the impact appears limited at this time: the fund was not, in fact, a direct source of funding for hundreds of young companies, and venture capital fundraising figures remain strong. However, the episode could make institutional investors more selective in new funding rounds and prompt some tech companies to postpone potential IPOs. The reaction from the semiconductor and cloud computing sectors was more immediate, as in the hours following the crash, stocks such as Nvidia, Intel, and AMD came under heavy selling pressure, dragging down sector-specific ETFs as well. However, the recovery seen in late July has reinforced the belief among many analysts that this was a technical correction rather than a structural shift. Demand for chips used in artificial intelligence continues to be driven by the ongoing expansion of data centers, while major cloud providers, such as Microsoft and Alphabet, are keeping their investment plans unchanged. More broadly, the Situational Awareness case serves above all as a wake-up call regarding the risks associated with highly leveraged strategies. Although this crisis is limited to a single hedge fund, many managers are already reviewing their risk limits, and investors are showing greater caution toward the higher valuations in the technology sector. The incident could accelerate a consolidation process already underway, without, however, compromising the long-term prospects for artificial intelligence, which continue to be supported by steadily growing structural demand.
Market and Analyst Reactions
The collapse of Situational Awareness has sparked comments from many quarters: many analysts point out that it was only a matter of time, given the fund’s structure. Well-known hedge fund manager Jerry Diao points out that “many saw this collapse as a matter of when, not if.” In fact, the extreme use of leverage and over-concentration had been fraying the nerves of several financial observers: it was likened to the “hot air of an AI bubble” colliding with market reality.
On the political and regulatory front, the affair has also drawn the attention of some lawmakers: while the fund was celebrated in tech publications through July, by early August calls for greater oversight were emerging. Questions have been raised in the U.S. Senate regarding the role of prime brokers and leverage limits in next-generation funds. However, at least so far, regulators have not issued any public sanctions: the story remains marked as a case of extremely poor risk management, rather than a full-blown financial crime.
For institutional investors, the lesson is one of rigorous risk management; some industry managers have warned that there will be repercussions on future asset allocations: in a draft report, Goldman Sachs advised carefully evaluating the use of options hedging, which did not function as a “buffer.” Others are talking about a price reset: as evidenced by the observed phenomenon of sharp rebounds, it is clear that, until now, the multiples of some AI stocks were based on growth expectations rather than current fundamentals. The herd mentality dissipated when the major seller disappeared from the market.
Conclusions
The outlook for the artificial intelligence sector will depend on how the geopolitical and macroeconomic landscape evolves. The most likely scenario remains a gradual easing of international tensions, which would support a recovery in technology markets and allow for continued investment in data centers, semiconductors, and AI startups, albeit with greater caution on the part of investors. A more favorable scenario could, however, materialize if tensions in the Middle East were to subside permanently and the Federal Reserve were to begin a cycle of interest rate cuts. In this context, leading AI-related stocks could once again drive market gains, supported by growing demand for technology infrastructure.
The most negative scenario remains that of a new geopolitical escalation accompanied by a deterioration in the global economic outlook. A sharp rise in oil prices, persistent inflation, and high interest rates could fuel a new correction in the technology sector, also curbing investment in AI startups.
In any case, the Situational Awareness case demonstrates that, even in a sector with extraordinary prospects such as artificial intelligence, exceptional returns cannot be achieved without careful risk management. The fund’s collapse serves as a warning to the entire industry: innovation and leverage can accelerate gains, but they can also amplify losses in a devastating way.
Unbiased Alpha builds macro data infrastructure and quantitative research tools for systematic traders. AuraStream provides standardised macro and market data through a single API; the kind of data that makes this sort of cross-asset analysis feasible. Learn more →
