On May 18, 2026, Situational Awareness LP filed its 13F for the first quarter of 2026.
The fund’s notional exposure to US stocks and options expanded from $5.52 billion at the end of 2025 to $13.677 billion, a quarterly increase of 148%.
But what has attracted market attention is not the scale, but the structure:More than 60% of the new notional exposure was placed on put options in the chip sector.
What did Q1 do?
The put options cover nine underlying assets:VanEck Semiconductor ETF (SMH), Nvidia, Broadcom, Oracle, AMD, Micron, TSMC, ASML, Intel.
SMH had the largest notional put size at $2.04 billion, followed by Nvidia at $1.56 billion. Micron and TSMC held both call and put options, betting on volatility in both directions rather than shorting in one direction.
It should be noted that the 13F filing only discloses the notional value of the options and cannot directly determine the size of the net short position.These PUT positions may be actively shorting, or they may be hedged while holding long positions.
The full intent cannot be reconstructed from the documents alone.
In terms of underlying stocks, funds continue to increase their investment in computing power infrastructure.
CoreWeave increased its holdings from 6.1 million shares to 7.18 million shares; IREN and Applied Digital also increased their holdings.
The expansion was most evident in the mining sector, with Bitfarms (now renamed Keel Infrastructure) increasing its shares from 6.9 million to 19.88 million, CleanSpark from 1.64 million to 12.28 million, and Riot Platforms from 6.17 million to 11.5 million.
Bloom Energy sold 3.59 million shares, but still holds approximately $879 million in market value, while retaining 408,500 call options. This is profit-taking, not a change of direction.
The exit actions are concentrated in the direction of optical communication.
Lumentum and Coherent were completely liquidated. Last quarter, Lumentum accounted for as much as 8.68% of the portfolio, but this quarter it has dropped to zero.
Intel's actions deserve special attention:Last quarter, we held approximately 20 million call options, which we liquidated this quarter, while simultaneously establishing new put option positions.
It's not closing out a position; it's a complete reversal of direction, from bullish to bearish.
Where the bottleneck is, that's where the money is.
The logic behind this 13F is a specific supply and demand assessment:The constraints on the expansion of AI are shifting.
Over the past two years, the core constraint on the scale of AI has been the shortage of GPUs, leading to continuous trading in Nvidia, HBM memory, advanced process technologies, and optical communications. During this period, the chip sector as a whole has enjoyed a significant premium.
However, as computing clusters grow to hundreds of thousands or even millions of cards, new constraints are emerging.
In the United States, grid connection applications are currently backlogged at over 2TW, with an average waiting period of more than five years; transformer production capacity is limited, and the construction cycle for new data centers is measured in years; chip production can continue to expand, but the power, land, and construction capacity required to support chip operation cannot keep up.
Under this judgmentThe logic behind shorting the chip sector is not that AI will fail, but that the valuation of chips has already reflected expectations, and value is shifting to more downstream physical infrastructure.
Buying SMH and Nvidia PUTs is a hedge against a possible valuation correction in the chip sector; continuing to hold CoreWeave, mining companies undergoing transformation, and Bloom Energy is a bet on the power and data center capacity that are truly becoming bottlenecks.
CoreWeave's actions also confirm this idea: call options were reduced from 10.81 million to 1.81 million, while common stock was increased from 6.1 million to 7.18 million.
The direction hasn't changed.The only difference was that the highly leveraged option positions were replaced with underlying stocks, reducing the impact of volatility on the portfolio.
From 225 million to 13.677 billion
This fund was founded in September 2024, and its first 13F filing revealed approximately $225 million in U.S. stock exposure.By the end of 2025, this figure had grown to $5.52 billion; and as of March 31, 2026, the nominal exposure reached $13.677 billion.
In the first half of 2025, the fund achieved a return of approximately 47%, while the S&P 500 rose by only about 6% during the same period, outperforming the S&P 500 by about 12.5 percentage points for the whole year.
Before founding the fund, the 24-year-old German published a 165-page book, "Situational Awareness: The Decade Ahead," outlining his prediction that the AGI timeline, power, and computing infrastructure would be the biggest bottlenecks. Early funding for the fund came from Nat Friedman, Daniel Gross, and Stripe co-founders Patrick and John Collison.
The significance of this quarterly report lies in,It has transformed a judgment that was previously more at the narrative level into a concrete position structure.
Chips are merely the entry point for expansion. What truly determines the speed of AI expansion is whether it can be connected to electricity in the real world, whether a data center can be built, and whether grid access approval can be obtained within five years.
If this assessment holds true, the key words for AI investment over the past two years have been GPUs and models;The next few years will likely involve electricity, land, and construction time.
