Nobody Was Hunting Him

Before dawn on Thursday, July 30, a twenty-four-year-old sold his entire stock portfolio to Ken Griffin in a single trade.

Not a position. The book. Every long, every short, negotiated overnight and closed shortly after sunrise. Leopold Aschenbrenner’s fund, Situational Awareness, had been up 439 percent on the year through June 30. July was going to print at minus 67.

Six days before that block trade, his fund sent investors a letter. It reported the 439 percent. It ended with a postscript: at times we call out opportunities that seem like a particularly good time to add funds, if you have been waiting for one.

Six days. That is the whole distance between asking for more money and having none of your own left to lose.

What the tape did while that was happening

Wednesday, July 29 was an FOMC day. The Fed held at 3.50 to 3.75 percent on a 9 to 3 vote, and the market read the hold as hawkish. The S&P fell 1.5 percent to 7,316.15, its low for the move. The Nasdaq 100 closed 11 percent below its June 2 high, which is the technical definition of a correction and the emotional definition of something worse.

So there was your explanation, ready to use. The Fed is behind on inflation. The AI trade is over. I have written before about trading through a Fed pivot and how quickly a rate decision becomes a story about everything.

Underneath the index, that same afternoon, the tape did something that did not fit.

Nebius closed at $148.22, down 46 percent from its June 30 price of $276.17. CoreWeave closed at $60.82. Adobe, which a lot of people were short as a proxy for software getting eaten by AI, closed at $263.43, up 28 percent in a month.

The AI infrastructure names bottomed and the software short topped in the same session. That is not what a macro repricing looks like. That is what one leveraged seller looks like on his last day of selling.

By Tuesday of this week, Nebius was $225.74. Up 52 percent in four sessions. CoreWeave was $91.90, up 51 percent. The S&P closed at 7,736.52, a record, its first in two months, which brings its own problem: at all-time highs, FOMO is the bigger risk. Nothing about any of those companies changed in four days. What changed is that the seller was gone.

Then he told a story

On July 31 Aschenbrenner wrote his investors again. It is a better letter than most people manage on the worst week of their life, and it opens with a sentence I respect: we let you down this month. Later he takes full responsibility in plain words.

But in between, he explains what happened to him:

As these moves proceeded, we started to see increasingly adverse trading in names publicly associated with us. These dynamics are essentially similar to a bank run: vulnerability begetting more vulnerability.

Blood in the water. Sharks. The market found the wounded animal and finished it. Half of finance repeated that version within the hour, because it is a good story and because it flatters everyone in it, including the man who lost.

Then somebody checked.

Bloomberg took the claim to S3 Partners, which measures short interest for a living. Bob Sloan looked at the fund’s top holdings and found no meaningful increase in short selling. Half of the top ten saw short interest flat or falling.

There was no raid. His positions were crowded, his size was enormous, and his leverage ran near four times capital. When those three things sit in one portfolio you do not need a predator. You need a bad week and a margin desk with a calculator.

Be fair to him on one point, because precision matters more than the dunk: he said adverse trading, not short attack, and adverse trading in a name everyone knows you own is real. But the version that traveled was the shark, and the shark is not in the data.

Why the shark is always the first thing we reach for

The mind will not accept a loss without an agent. It needs somebody to have done this.

Psychologists have measured the reflex for fifty years and given it a dull name, the self-serving attribution bias. We credit our wins to skill and our losses to circumstance, and under stress the sorting gets faster and cruder, because the amygdala cannot read your trading plan and it is not waiting for you to finish thinking. Market makers ran my stop. The algos hunted the level. Somebody knew. Every one of those sentences does the same work: it moves the cause of the pain from inside the account to outside it.

It works. That is the problem. It works well enough to get you to tomorrow without changing anything.

And notice the price. If the market hunted you, there is no rep to run. If your size was wrong, there is one, and it starts Monday. The story you tell about a loss becomes the position size on the next one. That is the real cost of a comfortable explanation.

The part that should sting

While a twenty-four-year-old was learning this at a scale most of us cannot picture, so were you.

Citadel Securities publishes what its desk sees. For the week of the collapse, its market intelligence team reported that retail investors were net sellers every single day, four consecutive sessions, the longest selling streak of the year. In technology, retail sold more notional than in any week in their data going back to January 2019, beating the previous record by more than 80 percent. Two of the three largest retail technology selling days ever recorded on that platform happened inside that one week.

The selling was heaviest in the semiconductor and memory names that retail had bought most aggressively in May and June.

And then this, from the same report: on Thursday’s bounce, retail continued to sell.

Read that again. The forced seller finished, the tape turned, the names that would be up 50 percent by Tuesday started to lift, and retail sold into it. The 11:47 stranger did not need four times leverage or a prime broker to arrive at the same outcome as a $45 billion fund. He got there with a phone and a feeling.

He had a story too. His story was that the Fed had broken something.

He was right, and it did not save him

The detail that makes this more than a story about somebody else’s bad week is that his call was good.

Microsoft reported 43 percent Azure growth in the middle of the collapse and guided higher. His largest names are up violently since the liquidation. Run the same book at moderate size and he is having one of the best years of anyone’s career, right now, this week, on the same tape that ruined him.

Leverage takes away the one thing a correct idea requires, which is time. Being right is a claim on the future. Margin is a claim on you, and it comes due first. This is the whole reason smart traders still blow up, and why position sizing keeps turning out to be a psychology problem rather than a math one.

Steve Cohen has a version of this he has said for years. Liquidity, leverage, concentration. Any one of them is a business. Two is a hard year. All three in one book is a countdown with a number on it. Long-Term Capital Management ran the same experiment in 1998 with two Nobel laureates on the masthead, and it ended the same way.

Aschenbrenner knew. In 2024, months before he launched, he said it out loud on a podcast: obviously not blowing up is sort of like task number one and two, or whatever.

He knew the rule. He could not, in the moment, be the person who followed it. Call that gap what you want, but do not call it stupidity, because he has more raw intelligence than I do and it bought him nothing that week. The gap is the whole subject of this publication, and it does not close by reading. It closes when you size for the version of yourself who turns out to be wrong for three weeks straight.

One more thing

That weekend, while the wires ran the worst headlines of his life, Leopold Aschenbrenner got married.

The Citadel trade closed early Thursday morning, and one of the things it bought him was the removal of all leverage from the fund. Which means that on Saturday, standing in front of the people who love him, his phone did not vibrate.

That is what he bought with the most expensive week of his year. A phone that does not buzz, and the ability to be where he was standing.

You can buy the same thing on Monday morning. It is available at any size, and the smaller you go, the cheaper it gets.

Sources: the fund’s July 24 and July 31 investor letters as reported by the Financial Times and Business Insider; Bloomberg on S3 Partners’ short-interest analysis; Citadel Securities Global Market Intelligence, “August, After The Reset,” August 3, 2026; price data from the July 29 and August 4 closes.

If this essay was useful, the deeper work lives in my book, Trade Calm, available wherever books are sold. The companion piece to this one is The Winning Streak Is the Setup, on why the run before the blow-up is part of the blow-up. New essays publish first on Trader Intelligence: tradequillo.substack.com. Subscribing is free.