How Situational Awareness fell 67% in a month while still up 80% on the year: liquidity and leverage explained

The puzzle: If Situational Awareness was still up around 80% for 2026, why did it have to liquidate the bulk of its public book? And how does a fund drop 67% in a single month when its underlying positions did not fall anywhere near that much? The answer is the interaction of leverage and liquidity — the two forces that turned a bad month into a forced unwind.

The maths of the drawdown: Picture the fund as $100 of investor capital at the start of the year. A 439% first-half return grows it to roughly $539 by 30 June. A 67% July drawdown then cuts it to about $178 — still up ~80% on the original $100, yet a brutal one-month fall. (This stylised example sets aside inflows and outflows, which disclosures suggest were relatively small; later investors did not capture the full 439% run-up.)

Why the fall was so fast: The fund was reportedly levered 4–5x. On ~$539 of equity that supports roughly $2,150–$2,700 of gross positions, plus options. A $361 loss is 67% of the fund's capital but only 13–17% of its leveraged exposure — so a mid-teens move in the book translated into a two-thirds hit to equity. Long positions such as SanDiskDealroom has a profile for this one. Try Dealroom → (down ~37%) and Bloom Energy (down ~31%) fell hard, while some software names the fund was short rose against it — leverage multiplying the pain on both sides.

Then the margin calls: Prime brokers had lent the fund billions against its portfolio. As positions fell, the equity the banks were lending against collapsed, and they demanded more collateral. Aschenbrenner tried to raise fresh capital and sold shares gradually to meet the calls — but rivals who knew the fund's positions began betting against them, making an orderly exit harder without pushing prices down further.

The block trade: To stop the spiral, the fund sold the vast majority of its public equities to CitadelDealroom has a profile for this one. Try Dealroom → in a single block at a discount of more than 10%, used the proceeds to repay lenders, closed every short and eliminated all leverage. It kept its roughly $3.5bn AnthropicDealroom has a profile for this one. Try Dealroom → stake, its other private holdings and a much smaller, fully paid-for public portfolio.

Where that ~80% actually comes from: The headline year-to-date figure is a blend, not a broad-based gain. With the public book novated to Citadel to satisfy the margin debt, LPs are left with essentially zero in public equities — that slice, roughly three-quarters of NAV, is marked at about –100%. The number is held up almost entirely by the private side: the AnthropicDealroom has a profile for this one. Try Dealroom → stake, up around +620% year-to-date and about a quarter of NAV. Blend a +620% quarter of the book with a –100% three-quarters and you land near +80% — a figure that says far more about one private mark than about how the fund actually traded.

The catch for late money: An LP who allocated on 1 July fares worse still. They took the full –100% on the public book, almost certainly captured none of the private gain (their entry mark sits at or above today's Anthropic valuation), and are on the hook for a tax bill — the forced July liquidations to meet margin calls crystallised capital gains. Net of tax, that is a loss that can exceed 100% of the money put in.

The takeaway: The fund was not wound down because it was losing money for the year — it was still up ~80%. It had to sell most of its public book because its remaining liquid collateral was no longer enough to safely support the money it had borrowed. As Steve Cohen has put it, “leverage, concentration and illiquidity are the three things that can kill you” — Situational Awareness managed to check all three boxes at once.

The old warning still holds: Warren Buffett, quoting his late partner Charlie Munger, made the same point more memorably: “My partner Charlie says there is only three ways a smart person can go broke: liquor, ladies and leverage. Now the truth is — the first two he just added because they started with L — it's leverage.” A fund up 439% in six months, run by one of the sharpest minds in the AI trade, still got margin-called out of its own book. Leverage, as ever, was the one that mattered.

Sources: mechanics explainer by Sheel Mohnot (@pitdesi), drawing on WSJ reporting; blended-NAV and late-allocator point via Geoff Karren (@geoffreykarren); Munger quote via Warren Buffett, CNBC, 2018.