AI Selloff Hits Balyasny and Verition as Citadel Collects the Fire Sale

July turned AI weakness into a flow-driven reset

The clearest signal from July was not just that some big funds lost money, but that Citadel's flagship Wellington fund gained while others were under pressure. Balyasny lost 1.5% in July, Verition fell 1.1%, and Citadel's Wellington rose 5.9% after buying the fire sale of Situational Awareness assets. That contrast suggests July was as much about liquidity and balance-sheet flexibility as it was about AI sentiment.

Balyasny's 1.5% July loss pared gains over the first seven months of the year to 1.2%. Verition's 1.1% decline left its year-to-date gain at 4.5%. The takeaway is not that one month settles the long-term AI debate, but that the selloff in AI-linked shares moved quickly through fund performance.

That outcome also lined up after hedge funds returned an average of 7.4% in the first half, their best six-month stretch since 2009. When positioning is rich after a strong run, the firms with cash and flexibility can benefit most when forced selling starts.

Millennium's loss showed the hit was broader than one fund

Balyasny and Verition were not alone. Millennium also fell 2.1% last month, which made it clearer that July was not just a problem tied to one book or one firm.

Leverage turned an AI drawdown into a funding problem

Situational Awareness had run its book at times up to four times the fund's own capital with exposure tightly focused on the AI buildout. When those positions weakened, leverage amplified the pressure. In that kind of setup, funding stress can hit before a longer-term thesis is fully disproved.

That helps explain why Citadel's gain and the buyer's advantage did not necessarily validate the AI thesis itself. It did show that a seller under margin pressure can turn a market decline into an immediate opportunity for a buyer with liquidity.

Quant losses and mixed results pointed to crowded risk

If the pain had stayed confined to one fund, this could have been treated as an isolated mistake. Instead, it spread. A China-linked quant fund lost more than 40% in just over three weeks, suggesting that concentration in AI-related trades and rapid deleveraging were affecting multiple books at once.

At the same time, Quantedge gained 7.8% in July and finished the year up 34.6%, while Millennium's decline trimmed its yearly gain to 8.2%. That split matters: the market was not simply punishing all systematic trading or all risk exposure. It was exposing crowded positioning that had to be cut under stress.

Higher oil and Treasury yields added pressure

July was not only an AI unwinding. The same report noted that surging oil prices stoked inflation fears and pushed 30-year Treasury yields higher. That can matter when markets are already digesting weaker AI multiples, because higher financing conditions can deepen selloffs before company fundamentals change meaningfully.

Watchpoints: - Broader multistrategy weakness would suggest forced selling was wider than one AI-focused book. - Continued strength in flexible platforms would support the idea that liquidity, not just stock selection, drove July returns. - A reset in oil and Treasury yields would remove one of the squeeze's amplifiers.

What July means for the next leg in AI stocks

July already offered a live read-through. In the same month that Balyasny lost 1.5%, Verition declined 1.1%, and Millennium lost 2.1% as a selloff in artificial intelligence stocks roiled markets, Citadel's Wellington still rose 5.9% after buying the fire sale of Situational Awareness assets. That contrast matters more than a simplistic debate over whether AI demand is real.

Not every fund was equally exposed

Not every fund hit in July had the same risk profile. Quantedge gained 7.8% in July and finished the year up 34.6%, showing that the month was not a clean verdict on all systematic or AI-linked investing. It was closer to a stress test on who could absorb supply and who had to sell it.

A useful way to frame the takeaway: - Concentrated AI-focused funds were more exposed to collateral stress and redemption pressure. - Public-market semiconductor and AI names still reflect longer-term buildout demand, even as their moves also mirror broader equity volatility and leadership concentration. - Flexible multi-asset platforms are better places to watch for signs that liquidity is stabilizing, because they can rotate across assets rather than remain stuck in one crowded book.

Stabilization would show up in buyers, not just in sentiment

Stabilization does not require an immediate AI rebound. It requires evidence that forced supply is being absorbed.

Watch for: - Flexible winners continuing to post gains while more crowded AI books lag. - Fewer large multistrategy drawdowns after July's hit across Balyasny, Verition, and Millennium. - Semiconductor leadership broadening beyond a small group of mega-cap names. - Less linkage between AI weakness and broader hedge-fund flow stress.

If current buyers stop absorbing forced AI selling and the same multistrategy losses spread again, July was likely only a pause rather than a true stabilization.