For the second straight quarter, Nvidia’s data center division—its growth engine—missed expectations despite rising 56% YoY. Inventories and receivables jumped, pointing to possible overbuild as hyperscalers front-load orders. The pattern is familiar: the dot-com fiber boom, the shale surge—periods when spending raced ahead of sustainable demand.
‍

That risk is magnified by the sheer scale of current commitments: tech’s megacaps are set to spend roughly $320B on AI and data centers this year, with about half of that flowing directly to Nvidia. Wall Street’s expectations now demand near-perfection. After two years of triple-digit growth, Nvidia is valued at $4T, makes up 7.5% of the S&P 500, and every earnings print dictates sentiment on the entire AI trade. Even OpenAI’s Sam Altman has warned investors may be “overexcited,” calling parts of AI a potential bubble phase.
‍

CEO Jensen Huang insists demand for Blackwell is “extraordinary,” and long-term the trajectory is intact. But in the near term, investors fear an AI digestion phase—a pause before the next leap. Those worries are amplified by competition: Alibaba’s new inference chip, announced Friday (BABA +13%), shows how quickly rivals are moving to fill the U.S. export void and chip away at Nvidia’s dominance.
‍

The bigger takeaway: the AI trade may be losing its safe-haven status. For months, investors leaned on Nvidia and its peers as the reliable “big return” hedge against macro volatility. But with growth expectations stretched and signs of spending fatigue emerging, that reliability is being questioned—just as the spotlight shifts to the Fed’s own fragility.