Hedge Fund Returns Stumble in July as AI Stock Rally Cools

Hedge Funds Take Big Hit in July After Bruising AI Selloff

For investors, the message is less about predicting the next AI move than recognizing how quickly leverage and concentration can turn a winning position into forced selling.

The cushion got thinner

Briefs Finance reports that Verition Fund Management fell 1.1% in July, leaving it up 4.5% for 2026. ExodusPoint Capital Management lost 0.9%, reducing its year-to-date gain to 3.5%. Millennium Management suffered a sharper 2.1% decline.

Those losses arrived after a remarkably strong opening stretch. PivotalPath figures cited by Briefs Finance put average hedge-fund returns for January through June at 7.4%, the best start to a year since 2009. Then AI stocks began to slide, and several of the funds that had benefited from the rally discovered the usual market truth: yesterday’s edge becomes today’s exposure when everyone owns the same thing.

The firms remain ahead for the year, including Balyasny, Verition and ExodusPoint. But the margin for error is smaller now. A strong first half can create the illusion that risk has been tamed. It has not. It has merely been paid for—so far.

The trade was crowded, and leverage supplied the accelerant

The concern behind the selloff was straightforward: investors began questioning whether the enormous spending flowing into artificial-intelligence infrastructure could continue rising at the same pace. When that assumption weakened, the stocks leading the rally fell fastest.

This is where the hedge-fund machinery matters. Large multistrategy funds spread positions across equities, bonds, currencies and other markets, but diversification on paper does not guarantee diversification in reality. If several strategies depend on the same AI-linked momentum, the portfolio can become one large trade wearing different jackets.

Borrowed money makes the structure more fragile. Leverage magnifies returns when positions work and magnifies losses when they do not. A falling position can trigger a margin call—a demand from a broker for more cash—and leave the manager with little control over the timing of a sale.

July offered a clean example of that asymmetry. Citadel’s main Wellington fund gained 5.9% after buying assets from Situational Awareness in an emergency sale. One manager’s forced exit became another manager’s discount. Markets are not sentimental. They transfer leverage from the distressed to the prepared.

What investors should actually check

The practical lesson is not to declare AI finished, nor to pretend that a single bad month proves the market has reached a lasting turning point. The useful question is narrower: how much of a portfolio’s recent performance depends on one sector, one factor or one narrative?

Look at concentration, not just the number of holdings. A portfolio can own many securities and still be exposed to the same AI-related trade. Check whether gains came from broad earnings strength or simply from rising valuations and momentum. And understand whether borrowed money—or any structure that creates a forced selling point—sits underneath the position.

Most investors will never receive a margin call. They can still experience the same economic problem: a crowded trade turns, liquidity disappears, and the exit becomes more expensive than the entry.

I watched versions of this movie in 2008. Sophisticated investors did not lack spreadsheets; they lacked room to wait. July’s AI selloff delivered the same old warning in a newer costume: when conviction becomes crowded, leverage does not create courage. It creates a deadline.