AI Earnings Are Rising. The Market Is Paying Less for Them.
The Q3 earnings season has exposed a gap between reported profits and what investors are willing to pay.
As of October 30, 82% of S&P 500 companies reporting had beaten EPS estimates. But the aggregate surprise was just 1.9%, versus a five-year average of 7.5%, according to FactSet.
The distinction matters.
An EPS beat tells us a company exceeded the published estimate. It doesn't tell us whether the next two years of earnings justify its valuation.
For AI infrastructure and semiconductors, the market may be shifting from pricing revenue growth to pricing the durability and cost of that growth.
Higher capex, depreciation, financing costs and eventual returns on invested capital matter more as the installed asset base expands.
This is why falling multiples alongside rising earnings estimates are not automatically a buying opportunity.
The test is whether earnings and free cash flow keep rising after the next round of capex and depreciation.
The market may not be questioning AI demand. It may be questioning how much of that demand becomes durable shareholder earnings.
#stocks