Why did chip stocks keep falling even after a string of earnings beats?

Google, Intel, and TSMC all beat estimates this earnings season. Yet chip stocks kept selling off the following week. Semiconductor analyst Jukan compared the pattern, on July 28, to last autumn and winter: good earnings, but no rally.

Why did the stocks fall despite good numbers

Google's Cloud revenue grew 82% year over year in Q2. The problem wasn't the revenue. It was the spending. Capex for the quarter came to $44.9 billion, up 26% from $35.7 billion in Q1.

Intel also beat with $16.1 billion in revenue, and AI-driven server revenue jumped 59%. TSMC and ASML both raised their 2026 spending and order outlooks. On the surface, there's no reason for this cycle to be rolling over.

Google stock falls on capex worries despite 82% Cloud growth (TradingKey, 2026-07-22)

출처: tradingkey.com

Google's own earnings day proved the point. Cloud growth beat expectations, but the stock wobbled anyway, dragged down by the bigger spending number. Good news and a bad reaction landed in the same report.

This isn't the market's first round of doubt

This skepticism didn't appear out of nowhere this week. A similar selloff hit on July 8, before earnings season had even started. The Philadelphia Semiconductor Index fell 10.8% in ten days, and about $1.3 trillion in market value disappeared, even though no one had reported bad numbers yet.

Chip index falls 10.8%, Intel down 21%, before earnings even arrive (Forbes, 2026-07-08)

출처: forbes.com

The blamed cause back then was the same one in play now. Big Tech's AI spending had jumped 67% to $650 billion in a year, and the doubt was whether that spending would ever pay for itself. In other words, this selloff isn't happening because results came in weak. It's happening because the doubt kept building for months, and good results alone can't dissolve it. In Jukan's own framing, the market has moved out of the "AI Summer" of May and June, and into a phase where it isn't debating how much capex is growing, but whether it will pay back and hold up. Google's 82% growth might have pushed the stock higher back in May. Right now it gets met with "so what happens in 2028" instead.

So what settles this

It comes down to one question. Is this doubt a fair reassessment of the fundamentals, or a temporary de-rating that lasts until a new product proves demand is still there? Jukan's own answer points to something like "Coding 2.0," a new blockbuster AI use case. If the market sees AI agents genuinely replacing part of software development work, capex gets read again as "productivity investment" instead of "cost." If that proof doesn't arrive by the next earnings season, this round of doubt gains one more data point. What's certain right now is that good earnings alone no longer move these stocks higher. The market is simply asking for more proof than it used to.