“When the tide goes out, you find out who’s been swimming naked.” It’s a popular Warren Buffett quote that sums up this week’s biggest news. We talked a few weeks ago about the astonishing run of the KOSPI over the last eighteen months, an index dragged skyward almost single-handedly by two memory-chip giants, SK Hynix and Samsung. This was, by any measure, a wild week for that market. Two limit-down days, a peak-to-trough decline of more than 40%, and then a Friday rally of about 17% that left everyone dizzy and no one particularly reassured.
But that is not really the focus this week. We are going to look at something tangential to the chaos, a story as old as portfolio management itself. A hot young manager posts numbers that seem impossible. Assets pour in. An almost cult-like following forms, the kind where investors sit refreshing the SEC website waiting to see what he bought and sold in his latest 13-F. And then one morning you open your preferred news website and read that the whole thing (almost) has been liquidated, sold off in a single block trade, because of a phone call that has been ending careers since before I was born: the dreaded margin call.
The Kid Who Saw the Future
To understand this week, you have to understand Leopold Aschenbrenner, and to understand him you have to accept a few facts that sound more like a movie script than real life. He is only now in his mid-20s, born in Germany, schooled in Berlin, and graduated valedictorian of Columbia at nineteen with degrees in economics and mathematical statistics.
He began his working life at the philanthropy arm of FTX. Then he joined OpenAI’s Superalignment team, the group tasked with making sure a superintelligent AI machine does not decide humanity is superfluous.
After OpenAI fired him in the spring of 2024, citing an information leak (which he denied), he did the thing that made him famous. He published a 165-page essay called “Situational Awareness: The Decade Ahead,” a sweeping dissertation that artificial general intelligence would arrive around 2027, that the world was nowhere near ready, and that the companies building the physical guts of it all, the chips, the power plants, the gigawatt-scale data centers, were the most mispriced assets on earth.
From Manifesto to Money
Most people who write a thesis like that never have to back it. You can debate an essay in the abstract forever and never face a consequence for being wrong. Aschenbrenner did the opposite. He took the manifesto and built a hedge fund around it, one where every single position was, in effect, a wager that the essay was correct. He named the fund after the essay, which tells you something about his relationship with confidence.
The fund launched in late 2024 with roughly $225 million. What happened next is what built the cult. Through the first half of 2026, Situational Awareness reportedly returned about 439% net of fees. Assets ballooned toward $20 billion, and at one fleeting moment in July, reportedly as high as $45 billion. The whispers started, as they always do. He was the next Druckenmiller or the next Soros or Buffett. Pick your investing legend.
The Mechanics of Conviction
But here is where we need to slow down and talk about the tool that built this fund, because margin is one of those things everyone has heard of but far fewer people actually understand.
Buying on margin means buying with borrowed money. Your broker lends you cash against the securities in your account, and you use it to buy more securities than your own capital would allow. If you have a dollar and borrow four more, you now control five dollars of stock. That is “5x leverage,” and it is roughly where Situational Awareness was reportedly running its book (somewhere in the 4x to 5x range). The appeal is obvious and seductive: when your positions rise 10%, you do not make 10%, you make something closer to 50%, because you are collecting the gains on the borrowed dollars too. A 439% return does not happen without this kind of amplification. Leverage is what turns a very good year into a mythical one.
The catch, and there is always a catch, is that the broker who lent you the money is not a passive partner. They want their collateral to stay worth more than the loan. So your account carries a maintenance requirement, a line below which the value of your collateral cannot fall. Cross it, and the phone rings. That call is the margin call, and it comes with a demand. Post more cash immediately, or we sell your positions until the math works again. Notably, they do not ask when it would be convenient. They do not care about your thesis, your essay, or your 2027 timeline. The person on the other end of the line is the least sentimental in finance.
Why It Breaks in a Storm
In a calm market, leverage is a quiet servant. The trouble is that leverage is symmetric. The same multiplier that turns a 10% gain into a 50% windfall turns a 10% loss into a 50% wound, and it does so precisely when you can least afford it.
But most people already live with enormous leverage and never feel a tremor, because they hold it in the one place designed to hide it, their house. Put $100,000 down on a $500,000 home, and you are running the same 5x leverage that would terrify you in a brokerage account. If that house quietly loses 20% of its value, your entire $100,000 of equity is gone, completely wiped out. And yet nothing happens; you don’t get a phone call. No one demands a check by the end of the day. You keep making the mortgage payment, the kids keep going to the same school, and you may never even know the paper loss occurred. Real estate leverage feels safe for one reason and one reason only. And that is because no one marks your house to market every morning. You hold the asset for years, through cycles, and the absence of a daily price is what lets the leverage sit there peacefully.
A leveraged fund is the same arrangement stripped of that mercy. Every position is priced every second the market is open, and the collateral is checked against the loan continuously. In the fund, a 20% drop stops being a paper loss you can ignore and becomes a demand for cash you may not have. This is why the same tool that safely finances half the houses in America can vaporize a $45 billion fund in a few weeks. It is not the leverage alone. It is leverage married to a market that reprices every single day.
Consider the negative feedback loop. Your leveraged positions fall. That drop eats into your collateral, which trips the maintenance line, which triggers the call. You must raise cash, so you sell. But you are selling into a falling market, often the same crowded, concentrated names that everyone else running the same strategy is also being forced to sell. Your selling pushes prices lower, which trips more accounts, which forces more selling. This is the feedback loop, and it is the reason a leveraged unwind rarely looks like a graceful exit and almost always looks like a stampede. When every position is a bet on the same idea, there is nothing in the portfolio to break the fall.
SK Hynix, one of the fund’s largest holdings, was at the center of this month’s AI-infrastructure selloff. In addition, AI darlings Nebius, Micron, CoreWeave, and SanDisk, names reportedly in the fund, fell somewhere between 27% and 54% on the month. On the other side, short bets against software companies like Adobe moved against the fund too, so it was bleeding from multiple wounds at once.
The brokers then began the uncomfortable work of marketing the fund’s holdings for sale. And by Thursday morning it was over. Situational Awareness sold its entire public equity book in a single enormous block trade to one buyer, later reported to be Ken Griffin’s Citadel. Assets that had touched $45 billion at the start of July were reportedly down around $10 billion.
The kicker, and it is a revealing one, is that Aschenbrenner had written to investors on July 24 calling the selloff “one of the best buying opportunities since early 2025” and inviting fresh capital starting August 1. The conviction never wavered, but the margin call came just the same.
Smart Is Not the Same as Seasoned
So here is the lesson buried inside the wreckage, and it is not the lazy one about a smart kid getting his comeuppance. Aschenbrenner is clearly brilliant. His essay made specific, dated predictions about the AI cycle, and so far the cycle has broadly tracked them. The thesis might even be right. The AI buildout is real, but that’s not really the question.
The question is whether being right about a story is the same as being able to manage a portfolio, and this week answered it once again. They are different skills, and the second one is mostly invisible until the tide goes out. A brilliant thesis tells you what to own. Risk management tells you how much, how to finance it, and what happens to you when you are right eventually but wrong right now.
This is the oldest pattern in this business, and it rhymes across every generation. A dazzling mind, a compelling story, a following that mistakes conviction for a track record, and leverage quietly doing the work underneath until the day it doesn’t. The names change, and the essays get more eloquent, but the broker stands ready to make the margin call the whole time, and they have never once been talked out of it.
The uncomfortable truth for anyone tempted to follow the next wunderkind into the next 13-F is that the most important number in a fund is not last year’s return. It is how the manager behaves in the week the borrowed money comes due. That number does not appear in the marketing deck. You only ever find it out the hard way, so make sure your swimsuit is on when the tide goes out.
Markets / Economy
- It was an absolutely wild week with big tech earnings and the blowup of the hottest fund of the last year. The S&P finished the week up 1.0%, the Nasdaq up 1.6%, and the small-cap Russell 2000 was flat.
- Core PCE prices increased by 0.1% from the previous month in June, less than market forecasts of a 0.2% rise. From the previous year, Core PCE rose as expected, up 3.3%, well above the Federal Reserve’s 2% target.
- The U.S. economy expanded an annualized 1.5% in Q2 2026, below 2.1% in Q1 and forecasts of 2.1%, the BEA’s advance estimate showed.
Stocks
- U.S. equities were in positive territory. Consumer Discretionary and Communication Services were the top performers, while Utilities and Real Estate lagged. Value stocks led growth stocks, and large caps beat small caps.
- International equities closed higher for the week. Developed markets fared better than emerging markets.
Bonds
- The 10-year Treasury bond yield increased six basis points to 4.75% during the week.
- U.S. bond markets were in negative territory this week while International bond markets were positive.
- High-yield bonds led for the week, followed by corporate bonds and government bonds.
