Capital Economics: AI bubble is late-stage, S&P 500 to fall 30%

Capital Economics has spent months building its case, and the warning is not only about the Fed. Reilly's September 10 report found consensus estimates for S&P 500 earnings growth matching dot-com-era peaks, concentrated overwhelmingly in technology and semiconductor names. The combined free cash flow of the four largest hyperscalers is projected to turn negative in 2027 as capital spending balloons, and their bond issuance has more than doubled over the past year. A fresh wave of equity issuance — the clearest signal in Reilly's view — is already visible in Anthropic's planned initial public offering this fall.

The market's own behavior has made the warning harder to parse. On July 30, Microsoft's market value rose by $450 billion in a single day, prompting Acadian Asset Management's Owen Lamont to compare it to '1.04 Houstons' in assessed property value. His dispersion index hit its third-highest reading in more than 2,850 trading days, trailing only 'vaccine Monday' in November 2020 and the DeepSeek shock of January 2025. Morgan Stanley's Lisa Shalett acknowledged the volatile September-October stretch but argued rising rates, oil-market stress and policy noise are already priced in 'clear-eyed fashion.'

What happens next may hinge most on Fed Chair Kevin Warsh. At Jackson Hole on August 28 he said he 'would be hard pressed to describe broad financial conditions as restrictive,' and UBS expects the vote to split 10-2 with Governors Christopher Waller and Michelle Bowman dissenting in favor of holding steady. UBS also expects the Fed to revise inflation projections down and rate-hike projections up in the same Summary of Economic Projections — something it says has never happened before. Capital Economics has pointed to the dot-com collapse as its template, when the Fed raised rates from 4.75% to 6.5% between June 1999 and May 2000 and the S&P 500 fell by nearly half. If Warsh hikes Wednesday and signals more ahead, it would mark the first meaningful tightening cycle since the AI trade took off in 2023 — the scenario the firm's smaller-bubble forecast did not assume.