Two years ago, a 25-year-old who used to work at OpenAI started a hedge fund with about $225 million. By early July it was worth close to $45 billion.

That's a 439% gain, built on one bet, that AI keeps winning. Then, over six trading days in late July, his own banks made him sell almost all of it. Read that again.

What's left is a stake in Anthropic worth about $5 billion, and he can't sell that yet either. This isn't a story about a bad stock pick. It's a story about leverage, and today we show you the math behind it.

Leverage means borrowing against your own money to control a bigger position than your cash alone would buy.

Put up one dollar, borrow three more, and you're moving four dollars of stock. Only one of those dollars is yours.

His fund ran with about four times leverage. That $225 million, grown toward $45 billion, controlled roughly four times as much in AI and chip stocks.

Here's the number that matters. At four times leverage, a 30% drop in your holdings isn't a 30% loss to you. It's closer to a 120% loss against your own money, more than you started with.

That's close to what happened. Core holdings like SanDisk and CoreWeave fell 27% to 54% in July, as the AI trade cooled off.

He carried hedges too, bets meant to pay off if other trades went wrong, including $8.5 billion in chip options. They barely helped, because those hedges moved the same way as the rest that week.

The leverage in a fund like this comes from prime brokers. These are banks that lend cash and stock, so a fund can trade bigger than its money allows.

Goldman Sachs, JPMorgan, and Bank of America all played that role here.

When your collateral loses enough value, the bank issues a margin call. Pay up more cash right now, or they sell your bets to get their money back. It doesn't matter what you think the stock is worth next year. It only matters what the bank can get today.

Rather than dump the stock bit by bit into a falling market, the brokers arranged one big trade. They moved roughly $16 billion of public stock to Ken Griffin's Citadel before the market opened on July 30. That kept the fire sale off the open tape, but he had no say over the price by then.

Three places to watch how this plays out, framed as names to study, not names to buy.

The one piece he kept is that $5 billion Anthropic stake. It can't be priced for real until Anthropic goes public, which some reports say could happen as soon as October. Once that listing lands, it's the real scorecard on this whole story.

Aschenbrenner wasn't wrong about the trade. He was up 439% chasing the same AI buildout we've circled all year, and some of those names still work.

What ended his fund wasn't the thesis. It was the six trading days he didn't have room to survive.

Everything here is how we think about the mechanics, not a set of instructions. Nothing in leverage ever plays out exactly the way the math sketches it. So do your own homework, and treat borrowed money with real respect.

Stay Locked In
— The Lead Editor, Main Street Betz

DISCLAIMER

We are a bunch of apes who figured out how to use a Bloomberg Terminal and a group chat at the same time. That combination is either genius or a liability, and we honestly aren't sure which one yet.

Nothing here is financial advice. Seriously. Do your own research. Talk to an actual licensed professional before you YOLO your savings into something you read in a newsletter written by an ape in a hoodie.

Past performance doesn't guarantee future results. The market doesn't care about your feelings, your conviction, or your "diamond hands." It will humble you. It humbles us too. That's the game.

We may hold positions in securities mentioned. Trading and investing is risky. Trade at your own risk. Eat Cookiez responsibly.