AI Bubble Fears Collide With Wall Street’s Rate Reckoning

A single trading session recently erased hundreds of billions of dollars from some of the world’s largest companies, then shifted the money elsewhere. That violent rotation is becoming the clearest warning yet that investors are no longer treating the artificial-intelligence trade as one unified bet.

Microsoft’s market value jumped roughly $450 billion on July 30. The following day, Apple lost about $360 billion, while Amazon added approximately $388 billion and Meta shed $102 billion. Owen Lamont, a behavioral economist and portfolio manager at Acadian Asset Management, said the market’s dispersion has reached one of its most extreme readings in nearly 2,850 trading sessions.

That does not prove a bubble is bursting. It does show that confidence is thinning.

Capital Economics economist James Reilly has assembled a more direct bearish case. In a September 10 report, he said valuations, expected earnings growth, market concentration and the volume of new equity issuance were all approaching levels associated with previous market peaks. His firm expects the S&P 500 to rise further this year before falling at least 30% from its high.

The pressure point is not only valuation. It is the cost of financing the AI buildout.

The 10-year Treasury yield briefly crossed 5% on September 14, a level not consistently sustained since the dot-com era. Higher yields raise the hurdle rate for long-duration growth stocks and make the debt-heavy construction of data centers more expensive. Rockefeller International Chairman Ruchir Sharma has argued that a decisive move above 5% would mark a new phase of tighter money for AI infrastructure.

Yet the market’s reaction to fresh warnings about AI development was selective. Nvidia fell 3.4%, while Intel, AMD and Marvell dropped between 5% and 6%. The Philadelphia Semiconductor Index lost nearly 6%. Alphabet gained almost 2%, Microsoft rose 1.6% and Meta added about 1.4%.

That split reveals what investors are pricing. If frontier-model development slows, chip suppliers and data-center operators take the first hit. Hyperscalers, by contrast, can reduce capital spending, use existing capacity and convert more revenue into cash flow.

Morgan Stanley’s Lisa Shalett called the September-October period Wall Street’s “silly season,” but maintained an S&P 500 target of 8,000 for year-end. The bulls still have a case. Earnings are growing rapidly.

The question is how much of that growth is already in the price.