On May 18, 2026, Situational Awareness LP filed its first-quarter 2026 13F report, revealing a dramatic escalation in nominal exposure to U.S. equities and derivatives. The fund's total nominal exposure surged from $5.52 billion at the end of 2025 to $13.677 billion, marking a 148% quarterly increase. While the sheer scale of capital deployment drew immediate attention, the structural composition of this expansion defined the market narrative: more than 60% of the incremental nominal exposure was directed toward put options targeting the semiconductor sector. These bearish instruments covered nine specific underlying assets, including the VanEck Semiconductor ETF (SMH), Nvidia, Broadcom, Oracle, AMD, Micron, TSMC, ASML, and Intel. The SMH put options represented the largest single position with a nominal value of $2.04 billion, followed closely by Nvidia at $1.56 billion.

Notably, positions in Micron and TSMC included both call and put options, suggesting a strategy focused on capturing volatility rather than a directional short bet. It is critical to note that 13F filings disclose only nominal values, preventing a precise calculation of net short exposure; these positions could function as active short-selling mechanisms or as hedges against existing long holdings, a distinction the report does not explicitly clarify.

Concurrently with the aggressive hedging in the chip sector, the fund significantly expanded its direct equity holdings in companies providing computing infrastructure and energy solutions. The stake in CoreWeave grew from 6.1 million shares to 7.18 million shares, accompanied by increased positions in IREN and Applied Digital. The most pronounced accumulation occurred within the cryptocurrency mining sector, which is increasingly viewed as a proxy for flexible power capacity. Bitfarms, recently rebranded as Keel Infrastructure, saw its holding jump from 6.9 million to 19.88 million shares. Similarly, CleanSpark holdings rose from 1.64 million to 12.28 million shares, and Riot Platforms increased from 6.17 million to 11.5 million shares. In contrast, the fund reduced its Bloom Energy stake by 3.59 million shares, yet retained a position valued at approximately $879 million along with 408,500 call options. This partial liquidation appears designed to realize profits rather than signal a strategic departure, as the fund maintained significant upside exposure through retained options. Data compiled by Woofun AI indicates that such rebalancing acts often precede broader sector rotations where capital moves from high-valuation hardware to undervalued utility assets.

The fund's exit strategy was concentrated heavily within the optical communications sector, marking a decisive rotation away from components that previously drove AI infrastructure growth. Lumentum and Coherent were completely liquidated; Lumentum, which constituted 8.68% of the portfolio in the prior quarter, was reduced to zero holdings. Intel's activity warrants specific scrutiny as it represents a complete reversal of sentiment. The fund held approximately 20 million call options on Intel in the previous quarter but sold them entirely during this period, simultaneously establishing new put option positions. This maneuver signifies a fundamental shift from a bullish to a bearish outlook on the manufacturer, moving beyond simple position closing to a strategic realignment. Woofun AI notes that such a complete inversion of optionality on a legacy chip giant underscores a belief that the valuation premium for traditional semiconductor manufacturing has peaked relative to emerging infrastructure bottlenecks.

The underlying logic driving this 13F filing rests on a specific assessment of shifting supply-demand dynamics within the AI ecosystem. For the past two years, the primary constraint on AI scaling was the scarcity of GPUs, fueling massive valuations for Nvidia, HBM memory, advanced process nodes, and optical interconnects.

However, as computing clusters approach scales of 100,000 to 1 million cards, the limiting factors are transitioning from silicon availability to physical infrastructure. Current data reveals a backlog of over 2 terawatts of grid connection applications in the United States, with average wait times exceeding five years. Transformer production capacity remains constrained, and the construction timeline for new data centers spans several years. While semiconductor fabrication can theoretically be expanded, the requisite power generation, land acquisition, and grid infrastructure cannot be deployed with comparable speed. Woofun AI analysis suggests that this divergence between chip supply elasticity and power infrastructure rigidity is the primary catalyst for the fund's bearish stance on semiconductors.

Consequently, the strategy of shorting the chip sector does not imply a failure of AI development but rather a recognition that current valuations have already priced in growth expectations, necessitating a rotation toward downstream physical assets. The purchase of put options on SMH and Nvidia serves as a hedge against potential valuation corrections in the semiconductor space, while the accumulation of CoreWeave, transforming mining firms, and Bloom Energy reflects a conviction that power and data center capacity will become the definitive bottlenecks. CoreWeave's own positioning supports this thesis; the fund reduced its call option holdings from 10.81 million to 1.81 million shares while increasing its common stock stake from 6.1 million to 7.18 million shares. This adjustment indicates a preference for direct equity ownership over highly leveraged derivatives to mitigate volatility risks while maintaining exposure to the core infrastructure play.

The fund's trajectory since its inception in September 2024 highlights the rapid scaling of this thesis. Its initial 13F report disclosed nominal exposure of approximately $225 million, which grew to $5.52 billion by the end of 2025 and reached $13.677 billion by March 31, 2026. Performance metrics underscore the efficacy of this approach; in the first half of 2025, the fund generated returns of approximately 47%, vastly outpacing the S&P 500's 6% gain. For the full year, the fund outperformed the benchmark by roughly 12.5 percentage points. The fund's manager, a 24-year-old German author, previously detailed this outlook in a 165-page book titled "Situational Awareness: The Decade Ahead," predicting that AGI timelines would be dictated by power and computing infrastructure constraints. Initial capital was secured from prominent figures including Nat Friedman, Daniel Gross, and Stripe co-founders Patrick and John Collison.

The significance of this quarterly report lies in its translation of theoretical constraints into concrete capital allocation. Chips serve merely as the entry point for expansion, but the true determinant of AI velocity is the ability to secure power, construct data centers, and obtain grid connections within a five-year horizon. If this assessment holds, the investment drivers of the past two years—GPUs and models—will be superseded by power availability, land rights, and construction timelines. The market is witnessing a pivotal moment where the narrative shifts from software and silicon dominance to the hard realities of energy and physical logistics, a transition that Situational Awareness LP is aggressively positioning to exploit.