AI Funds Take a July Beating: Whale Rock Capital Loses 21.7% in a Month

Another AI-focused fund has blown up. According to Chinese tech outlet QbitAI, Boston-based Whale Rock Capital fell 21.7% in July, wiping out half a year of gains in a single month.

The fund had been on a tear, posting a 72.5% return in the first half of the year; after July, its year-to-date return was cut to 35.1%. Founder Alex Sacerdote is a value investor who runs a concentrated book, going all-in on tech stocks with little diversification.

The damage traces back to July's AI selloff. U.S. chip stocks tumbled as worries about the return on AI infrastructure investment intensified, and the selloff swept through global tech stocks.

AI hardware names bore the brunt: SK hynix fell nearly 40% from its June high, while SanDisk, Micron and CoreWeave each dropped more than 30% during the month. Even Amazon and Google slipped slightly. Software stocks, by contrast, rebounded, with Adobe and others climbing.

Other AI-linked funds suffered too. Turion, which goes long and short AI hardware and semiconductor companies, fell 11.4% in July; Coatue's hedge fund lost 8.3%, its worst month in over a year, leaving full-year gains at 14.3%; Eureka dropped 6.9%, taking its annual return to 11.6%.

This is no longer a question of one fund's bad luck, the report argues. For years, markets have helped tech companies pile on leverage in pursuit of AGI. Google, Amazon, Meta and Microsoft alone plan to spend $720 billion this year, most of it on data centers — yet returns have been slow to materialize.

The release of open-source models such as K3 and V4 Flash has sharpened the doubts: if capable open models keep arriving from China, was the enormous compute buildout necessary? Investors are suddenly questioning a narrative that underpinned years of AI infrastructure spending.

On the funding side, OpenAI, SpaceX and Anthropic are all seeking market financing and racing toward IPOs. Whether AI capex translates into revenue — and how fast open-source models improve — will determine whether funds that went all-in on AI can stop the bleeding.