Earnings Are Rising. AI Stock Valuations Are Falling. Which Signal Should Investors Trust?

AI companies are expected to make more money, but investors are paying less for each dollar of those expected profits. Here is what that gap tells us and what it doesn't.

There is a strange split in AI stocks right now: analysts have raised their earnings expectations, yet valuation multiples have fallen sharply.

From the end of June to mid-October, expected earnings for the next twelve months rose by 25% for memory stocks and 25% for AI infrastructure companies. Over the same period, their forward price-to-earnings multiples fell by 34% and 22%, respectively.

At first glance, that looks contradictory. If companies are expected to earn more, why are investors willing to pay less for them?

Because a share price reflects two things: how much a company is expected to earn, and how much investors are willing to pay for those earnings. Those two can move in opposite directions.

Figure 1. Changes from 30 June to 12 October 2026 in expected earnings for the next twelve months and forward P/E multiples.