Big Tech AI stocks slide as investors demand proof capex will pay off

Investors are pulling back from Big Tech’s AI stocks and pressing the largest platforms to prove that a projected $7.6 trillion AI infrastructure buildout will pay off. That estimate, from Goldman Sachs, covers cumulative spending on compute, data centers and power between 2026 and 2031, led by hyperscalers including Alphabet, Amazon, Meta, Microsoft and Oracle.

The move broadens a rout that began earlier in the week with a chip-stock selloff into a wider reckoning over AI capital spending. Where the immediate question was chipmakers’ valuations, the pressure has spread to the companies writing the biggest infrastructure checks.

Kate Brennan of the AI Now Institute, a policy research group, said the returns ‘are not coming in’ and that efficiency and productivity claims ‘are not netting out.’ Ed Yardeni of Yardeni Research warned that the AI ecosystem ‘falls apart’ if end-user demand fails to materialize or prices fall sharply.

The skeptics have data to point to. A Gartner study in May 2026 found that businesses replacing workers with AI agents frequently failed to generate returns. Public sentiment is cool, too: a Pew Research survey found only 16% of US adults expect AI to help society over the next two decades, against 40% who expect harm.

None of this proves a bust. Spending on this scale can lead revenue by years, and the largest platforms are funding it largely from cash flow. The bear case is not that demand has failed. It is that demand has not arrived fast enough to satisfy the market.

What changes the story is evidence of paying customers at scale. Without it, each earnings season becomes a referendum on whether the buildout is an investment or an overhang.