Situational Awareness LP, the investment firm run by Leopold Aschenbrenner, reported a $20.24 billion U.S. equity and options portfolio at the end of the second quarter, with more than half of its disclosed holdings concentrated in Sandisk and Micron. The June 30 Form 13F also shows the firm had removed a large set of semiconductor-related downside hedges that appeared in the previous quarter, leaving the reported portfolio heavily exposed to a narrow group of AI infrastructure stocks.
Sandisk represented approximately $5.67 billion of the portfolio, or 28.0%, according to Situational Awareness’ filing with the U.S. Securities and Exchange Commission. Micron accounted for another $5.57 billion, or 27.5%. Together, the two storage-chip companies made up roughly $11.25 billion, equal to 55.5% of the firm’s reported 13F holdings.
That concentration placed a substantial part of the disclosed book behind a single market view: that demand for high-performance memory, storage and computing infrastructure would continue to rise as artificial intelligence companies build and expand data-center capacity. Memory chips have become a closely watched part of the AI supply chain because large-language-model training and inference require vast amounts of high-bandwidth memory and storage.
AI infrastructure names filled the rest of the largest positions
The remaining major positions extended the fund’s exposure across the physical infrastructure supporting AI workloads. Bloom Energy, a fuel-cell and power company, was listed at about $1.90 billion. Taiwan Semiconductor Manufacturing’s U.S.-traded ADRs were valued at about $1.27 billion, while AI cloud company Nebius accounted for approximately $1.23 billion.
Situational Awareness also reported positions in CoreWeave, Core Scientific, Applied Digital and IREN. Those companies operate in or around the data-center, cloud-computing, power-generation and high-performance-computing markets. Their inclusion alongside memory-chip holdings suggests the portfolio was designed around the capital-intensive buildout needed to support AI systems, rather than around software developers alone.
The filing listed 26 positions in total. Such a relatively small number of holdings can amplify gains when a chosen theme performs well, but it also gives individual stock movements greater influence over the portfolio’s reported value. In this case, the two largest positions alone carried more weight than the other 24 holdings combined.
Put-option hedges disappeared during the second quarter
The most consequential portfolio change in the filing was the removal of put options tied to major semiconductor and technology names. Put options generally increase in value when the underlying security declines, allowing a fund to reduce losses from a market downturn or from weakness in a specific industry.
In the first quarter, Situational Awareness had reported more than $8 billion in notional exposure through puts tied to chip-sector instruments and companies including the VanEck Semiconductor ETF, Nvidia, Oracle, Broadcom, AMD, Taiwan Semiconductor and ASML, according to the prior 13F disclosure.
The June filing no longer listed those positions. The reported changes included reductions in puts tied to the semiconductor ETF, a Nvidia bear ETF, Oracle, Broadcom, AMD, Taiwan Semiconductor and ASML. The firm also reduced options exposure linked to Micron and Sandisk while building larger direct equity positions.
The shift changed the character of the reported portfolio. A hedged book can retain a bullish view on a sector while limiting some downside from a sharp reversal. By June 30, Situational Awareness’ disclosed holdings were more directly tied to the performance of AI infrastructure equities, particularly storage semiconductors.
A 13F filing cannot show a manager’s complete financial position. It does not include cash, short positions, many foreign securities, private holdings, swaps, futures or the terms of any borrowing. The values assigned to listed options also generally reflect the market value of the underlying shares represented by the contracts, rather than the premium paid for the options. As a result, the filing offers a useful snapshot of long U.S.-listed holdings but not a full measure of the fund’s risk, leverage or daily trading activity.
Additions reinforced the same investment theme
Situational Awareness added or expanded several positions during the quarter, with Taiwan Semiconductor’s ADRs increasing by 6.19% of the reported portfolio. Nebius, STMicroelectronics, SharonAI and Keel Infrastructure appeared as new or increased holdings, based on the filing’s reported changes.
Taiwan Semiconductor is central to the global manufacture of leading-edge processors, including chips designed by many major AI companies. STMicroelectronics adds semiconductor exposure from a different part of the industry, while the other additions fit the filing’s focus on computing capacity, power and infrastructure.
The portfolio structure means the fund’s reported results would have been especially sensitive to a change in sentiment toward AI spending, data-center construction costs, chip supply, memory pricing or the financing conditions facing compute providers. These businesses operate at different points in the technology supply chain, but their valuations can move together when traders reassess expected returns from AI capital expenditure.
Filings arrive after the quarter has ended
The disclosure arrived under the SEC’s Form 13F regime, which requires institutional investment managers with at least $100 million in qualifying assets to report certain U.S.-listed securities within 45 days of each calendar quarter’s end. The filings are backward-looking and do not reveal whether positions were reduced, closed or hedged after June 30.
That timing is especially relevant for portfolios concentrated in volatile technology shares. A position disclosed at the quarter-end may have changed substantially by the time the document becomes public, while price movements after the reporting date can alter its economic importance.
Situational Awareness’ filing offers a clear record of how aggressively its reported public-equity exposure had tilted toward AI infrastructure by late June: a $20.24 billion book dominated by memory-chip leaders, supported by power, foundry, cloud and data-center names, and no longer carrying the broad semiconductor downside options that had appeared three months earlier.
Explore how AI-driven strategies shape crypto markets in 2026 with Toobit’s insightful guide on web3, AI, and crypto.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
