Inside Leopold Aschenbrenner's Q1 2026 13F: The Most Anticipated Filing of the Quarter

It is rare for a 13F filing, a quarterly snapshot disclosure that hedge funds have been making for forty-plus years, to generate genuine market suspense. Aschenbrenner managed it. On Friday May 15, his firm Situational Awareness LP submitted its Q1 2026 Form 13F-HR to SEC EDGAR right against the late deadline. The SEC did not post it to the public site that evening, so the entire investing internet had to wait until Monday morning, May 18, 2026, to see what the 24-year-old former OpenAI researcher actually owned.

When the filing finally hit the wire, it justified the wait. The fund disclosed $13.68 billion in 13F-reportable securities, more than double the $5.52 billion shown three months earlier. The top 10 positions accounted for 72.66 percent of the book. And in a sharp turn that surprised many of his followers, Aschenbrenner had loaded up on put options against essentially every major AI semiconductor name. Nvidia. Broadcom. AMD. Oracle. Micron. ASML. Intel. Taiwan Semiconductor. Corning. Plus the VanEck Semiconductor ETF as a wrapper trade on the whole sector.

This is the same Leopold Aschenbrenner who, in June 2024, published a 165-page essay arguing that artificial general intelligence is coming faster than anyone realizes and that the AI capital expenditure wave will be measured in trillions. He has not changed his mind. What changed in Q1 2026 is how he is positioned to capture it.

Who Is Leopold Aschenbrenner?

If you have only encountered Aschenbrenner's name through hedge fund headlines this year, you might not realize how short his on-ramp to running billions of dollars actually was. He is twenty-four years old. Five years ago, he was an undergraduate.

The arc that matters most for understanding the fund is the OpenAI exit. During his time on the Superalignment team, Aschenbrenner wrote a memo to OpenAI's board flagging what he described as serious security risks around the company's research and information handling, particularly with respect to potential foreign-state industrial espionage. He was fired in April 2024 over what OpenAI characterized as an information leak, a framing Aschenbrenner has publicly disputed. The memo has been reported as a factor in the dismissal. Two months later, in June 2024, he published the document that turned him into a Wall Street figure.

The "Situational Awareness" Manifesto

Published in June 2024, Situational Awareness: The Decade Ahead is a roughly 165-page essay. It makes three claims that are central to understanding everything that has followed:

The essay went viral well outside investing circles. Among those who praised it publicly was Ivanka Trump. Among those who quietly funded the next chapter were Patrick and John Collison (the brothers behind Stripe), Nat Friedman (now product lead on Meta's AI team), and Daniel Gross (now co-lead of Meta Compute). They seeded Situational Awareness LP with roughly $225 million in late 2024. Aschenbrenner reportedly put nearly his entire personal net worth into the fund.

The Fund: $225 Million to $13.68 Billion in Eighteen Months

The fund's growth, traced through its publicly filed 13F reports, is the kind of trajectory that hedge fund consultants normally describe as career-defining for someone on year fifteen, not year one. Aschenbrenner has done it in his first eighteen months.

That trajectory is partly performance and partly inflows. According to the firm's own breakdown, net new external capital accounted for roughly 32.5 percent of the Q1 2026 growth, with mark-to-market appreciation handling the rest. Names like Bloom Energy ran up 176 percent from his Q4 2025 entry. SanDisk reportedly delivered north of 800 percent on his cost basis. Those are right-tail outcomes that pulled the rest of the book to bigger numbers, and pulled new LP capital in behind them.

This is also the moment to flag what 13F numbers mean and do not mean. Under SEC reporting rules, the market value of options on a 13F is the notional value of the underlying shares, not the premium the fund actually paid. So the $13.68 billion headline mixes long stock at full mark-to-market with options where the cash outlay can be a small fraction of the displayed notional. That mechanical detail becomes critical when you read what comes next.

Why This Filing Was the Most Anticipated of the Quarter

There were three reasons the investing world was glued to this particular 13F.

First, the source. Aschenbrenner is not just running a fast-growing book. He is the most public spokesman for the bull case on AI infrastructure scaling. His thesis, his fund, and his name have become a kind of consensus reference point. One Fortune profile this year quoted an analyst at a major tech fund who described a moment of market stress when his colleagues held off dumping shares after another analyst said, "Leopold says it's fine." When the AI prophet rebalances his book, the rest of the AI-adjacent world wants to see what he saw.

Second, the timing. Nvidia reports Q1 fiscal 2027 earnings on Wednesday May 20, two days after the filing dropped. Any 13F revealing serious bearish positioning on Nvidia immediately before its earnings is going to get oversized attention. The market spent the weekend modeling scenarios in which Aschenbrenner was about to call a top.

Third, the suspense was self-inflicted. Situational Awareness filed its 13F-HR with SEC EDGAR at 6:33 p.m. Eastern on Friday May 15, right against the 45-day deadline. The SEC's public posting infrastructure did not push it out for general access until the following Monday morning at the open. The investing community, especially on financial X (formerly Twitter), spent the weekend openly speculating, often in real time, about whether the fund had simply requested confidentiality or whether something dramatic was being held back. The market did not see the answer until Monday May 18.

What the Q1 2026 13F Actually Shows

The biggest takeaway is structural, not directional. Aschenbrenner has built what professional traders call a barbell portfolio. On one end: high-conviction long bets on the physical infrastructure that powers AI compute. On the other end: large put-option hedges against the high-multiple semiconductor names that have already absorbed most of the consensus AI capex narrative. The middle of the AI stack, the place where you could go neutrally long the broad theme, has been hollowed out.

Side One: The $8.46 Billion Bear Hedge on Chips

This is the headline. Across eleven separate put-option positions, the fund disclosed roughly $8.46 billion in notional bearish exposure, with $8.45 billion of that concentrated in ten semiconductor and AI-hardware names and a small additional $6.76M position on Infosys (the only non-chip put). Roughly 62 percent of the entire 13F book by notional value is short the chip stack.

A few things to notice. SMH is the single largest put position, which makes it a portfolio-level hedge against the entire semiconductor sector rather than a one-name bet. The Nvidia and Broadcom puts are individual-name bearish positions on the two dominant AI chip beneficiaries. The Oracle put is interesting because Oracle is not a semiconductor company at all; it is on the list as the most aggressive infrastructure-priced cloud play in the AI buildout, and Aschenbrenner is implicitly saying its valuation already prices in too much.

One subtle move: the fund holds both puts and calls on Micron and on Taiwan Semiconductor. That is a two-way volatility bet, signaling that Aschenbrenner expects large moves on those names but is not committing direction. Memory cycles and Taiwan geopolitics both feed into that read.

Side Two: The Long Book - Bet on the Bottlenecks

The other end of the barbell is where the conviction trades sit. These are long equity positions in companies whose business model puts them physically downstream of the AI compute build-out: on-site power generation, repurposed cryptocurrency mining facilities pivoting to AI hosting, the GPU cloud layer, and storage.

The pattern is direct. Anything that supplies power, hosts compute, or stores the data that AI workloads need: Aschenbrenner is long it. Anything that sells the picks-and-shovels and is already richly valued: Aschenbrenner is hedged against it.

The Reshuffle: What He Exited Entirely

Q1 was a heavy turnover quarter. The fund opened 23 new positions, added to 9 existing ones, reduced 4, and fully exited 10. A few of those exits are worth flagging.

The Lumentum and Coherent exits are notable because they cut against the "Aschenbrenner is long everything physical in the AI buildout" simplification. Optical transceivers are physical AI infrastructure. He is out of them, presumably because their valuations got there before the underlying capex did. The Intel call unwind is the largest single sentiment change, taking what had been a high-leverage directional bullish bet on an Intel turnaround and replacing it with a small put position.

The Strategy in Plain English

If you collapse all of this down, Aschenbrenner is making the following set of bets at once:

Anna Coulling at Investing.com framed the architecture neatly: the fund is selling shovels and building substations to the AI miners, while buying insurance on the underlying metal itself. That captures the structure better than any single position does.

Timeline: How We Got Here

What This Means Going Forward

A few things to watch from here. First, this filing is a March 31 snapshot. Between quarter-end and today, the AI semiconductor names have rallied hard. Some of those puts are deeply underwater on premium-paid basis right now, though the notional protection structure remains in place. Aschenbrenner's actual current book may already look quite different from what is disclosed; the Q2 2026 13F (due by August 14, 2026) is what will tell us whether the hedge has been pressed, rolled, or unwound.

Second, the Nvidia print on May 20 is the immediate catalyst. If Nvidia delivers a clean beat and raise, the AI chip narrative gets fresh oxygen and the puts continue to lose value. If guidance underwhelms or the Blackwell ramp shows any cracks, the puts become very valuable very quickly and Aschenbrenner looks prescient.

Third, the long book is more durable than the put book in terms of investor signal. Even if the hedge gets unwound, the conviction trades on Bloom Energy, CoreWeave, SanDisk, T1 Energy, and the bitcoin-miner-to-AI-host basket are not the kind of positions you put on if your view is that AI capex is about to stall. He is still long the underlying thesis. He is hedged against the part of it that the consensus has already paid up to own.

The Complete Q1 2026 Portfolio: All 42 Active Positions + Exits

Below is the full picture of Situational Awareness LP's Q1 2026 13F holdings, sourced directly from the SEC filing (Accession 0002045724-26-000008) and organized by position type. All dollar values are at the filing date of March 31, 2026. Options are reported at notional value (the value of the underlying shares), not premium paid. The filing discloses 42 active positions plus 10 fully-exited names. Note that several companies appear on multiple lines because the fund holds both common stock and options on the same name (for example, AMD has both a $20M common stock position and a separate $969M put position).

Frequently Asked Questions