In short: Context, not a view: he thinks the anticipated IPO funding-flow "played into the implosion" — the fund "was hugely long some of the most liquid and best performing names," exactly the names managers must sell months in advance to fund mega-IPO allocations. "Anticipation of the upcoming flood of new issues became a pin that pricked at what was going on inside." Other things mattered too, "but there was more to it than that."
Situational Awareness was a hedge fund that blew up. Kedrosky offers it as evidence for the mechanism above rather than as a comment on the fund.
The fund was heavily long the most liquid, best-performing AI-linked names — which are precisely the shares every other big manager has to sell in order to fund allocations to the coming mega-IPOs. That selling begins quietly, months ahead, so the pressure arrived before there was any obvious news.
His phrase: the anticipated flood of new issues "became a pin" that pricked the position. Other things went wrong too, but the first leak of air was structural and, he says, predictable from the supply calendar alone.
26:11And in a weird way, and I was talking to one of the derivatives guys at JP Morgan about this, I think that played into the implosion at Situational Awareness, that they were hugely long some of the most liquid and best performing names for obvious reasons, but those things came under early pressure because they were going to become a source of funding inevitably for the upcoming spate of IPOs.
In short: The cautionary tale that frames his whole risk section: a $45B-exposure, ~4×-levered long-AI-infrastructure / short-software book run by a 24-year-old ex-OpenAI superalignment researcher (Columbia valedictorian at 19), seeded with ~$225M by the Collisons, Nat Friedman, Daniel Gross and Graham Duncan off a viral 165-page paper. It gave back 400%+ of gains and sold the public book to Citadel. "I think it's the heavy leverage that killed him… if he only leveraged 50% he'd still be running billions today."
A 24-year-old former OpenAI researcher wrote a widely-read 165-page paper about where AI was heading, raised money on the strength of it, and built a fund that was long AI infrastructure and short software — with roughly four times leverage on $45 billion of exposure. He was up more than 400% before giving it all back and selling his entire public portfolio to Citadel.
Singh's reading is that the analysis was fine and the leverage was fatal: "if he only leveraged 50% he'd still be running billions today." It is also why the software stocks he owned suddenly exploded higher — when a fund that size is forced to buy back its shorts, the prices go straight up regardless of fundamentals.
8:51I think he had given back all his gains on the years of like 400 plus percent on the public book and he was up 80% net just on Anthropic on the private book and then I think he's going to focus on privates now. Basically this guy was a former OpenAI super alignment researcher and back in 2024, he published this 165-page paper titled Situational Awareness: The Decade Ahead and he predicted a lot of the AGI and a lot of the rallies we've seen in memory and
In short: Fail of the week — the ~$20B AI fund run by the "prodigy" whose 2024 essay Situational Awareness: The Decade Ahead cast him as one of a chosen few who could "price all this in." He instead grew the fund fast, "levering up with a concentrated bet on a few bottleneck companies," posted meteoric first-half returns, then sought rescue capital and sold the bulk of its stock portfolio to Ken Griffin after deep losses. "Pride comes before the fall… This is investing 101."
The fail of the week is a cautionary tale rather than a stock. Leopold Aschenbrenner wrote a widely-shared 2024 essay called Situational Awareness: The Decade Ahead — at 22 years old — arguing that only a few hundred people on earth truly understood what AI was about to do, and that he was one of them. He then launched a hedge fund with the same name, grew it to roughly $20 billion, and made a concentrated bet on a handful of "bottleneck" AI companies, borrowing money to magnify the position.
Borrowing to invest (leverage) works both ways: it multiplies gains on the way up and losses on the way down. The fund posted spectacular returns in the first half of the year, then the bets went against it. It approached investors and lenders for rescue capital, and ultimately sold the bulk of its stock portfolio to Ken Griffin after deep losses. Carlson's takeaway is deliberately unglamorous — concentration plus leverage is a known way to blow up, and no amount of foresight exempts you: "Pride comes before the fall… This is investing 101."
25:00Barely anybody knows how to price all this in, but Leopold does. Well, it turns out that he didn't. It was a bit harder to accurately price all this in because what Leopold ended up doing was taking his fund, growing it to a massive extent, levering up with a concentrated bet on a few bottleneck companies, AI companies.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.