CalPERS cut its direct holdings in the seven largest U.S. Technology companies by 26% during the first half of 2026, redirecting capital toward private equity as investors reassessed lofty valuations and the scale of artificial-intelligence spending. The group includes Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta Platforms and Tesla. Deputy chief investment officer Anton Orlich said public-equity valuations appeared expensive and that the technology leaders could underperform other investments.
The shift comes as CalPERS seeks broader exposure across hundreds of companies rather than relying as heavily on a small group of market leaders. Its move coincided with mixed performance among the stocks: the Roundhill Magnificent Seven ETF was up 7% for the year, while Apple had gained 34% and Nvidia 12%; Meta, Alphabet and Tesla were lower. The reallocation therefore reflects a portfolio decision to reduce concentration even while some of the largest technology shares continued to rise.
Pressure was sharper among smaller companies that build or operate AI-focused data centers. He attributed SpaceX’s access to investment-grade credit and revenue from Starlink and rocket launches, while saying debt markets had nearly closed to companies without substantial revenue outside AI data centers.
Neoclouds typically borrow to buy specialized computing hardware and rent that capacity to AI developers and other customers, leaving them exposed to both financing costs and uncertainty about future demand. CoreWeave, Cipher Digital and IREN fell in Monday pre-market trading as that concern spread, while broader market commentary described investors as demanding stronger evidence that the AI buildout will produce returns large enough to support its borrowing.
The contrast between CalPERS’ diversification and SpaceX’s borrowing highlights a market increasingly separating established companies from highly leveraged AI infrastructure providers. For neocloud operators, tighter credit could limit expansion, encourage mergers or leave more of the AI infrastructure market to companies with stronger balance sheets; those outcomes remain possibilities rather than confirmed results.