Leopold Aschenbrenner runs $5.5B fund

The real story here is not just that Leopold Aschenbrenner runs a very large fund now. It’s that the fund looks like a tradable version of his AI worldview. Situational Awareness LP showed $5.52 billion in reported 13F holdings for Q4 2025 after reporting about $255 million a year earlier, and third-party trackers now show firm AUM above that. That is an absurdly fast jump, even before you get to the stock picks. ### What actually is this fund? Situational Awareness LP is the investment firm Aschenbrenner launched after leaving OpenAI in 2024. The SEC trail is real — the manager appears under CIK 0002045724, and the firm has filed quarterly 13Fs since late 2024. By the Q4 2025 filing, the disclosed public portfolio had reached $5,516,758,345. ### Why are people suddenly talking about it? Because the portfolio got big fast, and because the bets are unusually concentrated around AI infrastructure. (13f.info) WhaleWisdom shows two clients, roughly $9.28 billion in discretionary AUM from an April 27, 2026 Form ADV, and a top-10 concentration near 86%. That’s not a broad hedge-fund buffet. It’s a conviction book. ### Why does Bloom Energy keep coming up? (sec.gov) Bloom was the fund’s largest disclosed holding at year-end 2025. WhaleWisdom lists 10,076,022 shares. Other portfolio trackers peg the position around $875 million at the filing date, or roughly 10% of the public book. The basic idea is simple: if AI data centers can’t get enough grid power fast enough, on-site generation becomes valuable. Bloom sells exactly that kind of workaround. (whalewisdom.com) ### Was the “1,420% gain” real? That specific number looks like a social-media performance note, not something you can verify cleanly from the SEC filing alone. A 13F tells you what a fund held at quarter-end, not the exact internal return on each line item. What is verifiable is the position size, the accumulation over time, and the fact that Bloom became the flagship disclosed equity bet as AI-power demand surged. So the big-picture claim holds. The exact percentage should be treated cautiously unless the fund itself publishes it. (whalewisdom.com) ### What else is in the book? The pattern is the interesting part. Public trackers show big exposure to Bloom, CoreWeave, Intel calls, Lumentum, Core Scientific, IREN, Applied Digital, and other names tied to compute, networking, power, and data-center buildout. In plain English — not “which chatbot wins,” but “who sells the picks, shovels, electricity, and fiber.” ### Why does that fit Aschenbrenner so well? (sec.gov) Because it mirrors the thesis he became known for. His “Situational Awareness” essay argued that advanced AI would force a massive buildout in chips, energy, and industrial capacity. The fund looks like someone took that memo and bought the bottlenecks. That’s why people are fascinated — it feels less like stock picking and more like a map of where AGI would break the physical world first. (13f.info) ### What’s the catch? Two catches. First, 13Fs are delayed and incomplete — they miss shorts, foreign holdings in many cases, and plenty of private exposure. Second, a concentrated infrastructure book can look genius in an AI capex boom and brutal if that capex cycle cools. Even some public trackers already disagree on exact current weights and performance because they reconstruct returns from stale filings. (oksana-meier.medium.com) ### Bottom line? The headline isn’t just “ex-OpenAI guy runs $5.5 billion.” It’s that one of the most closely watched new AI investors is making a very specific bet: the money may sit upstream from the models, in power, semis, optics, and data-center plumbing. If that thesis is right, Bloom Energy is not a quirky side position. It’s the point. (whalewisdom.com) (13f.info)