Big Tech just put the AI bubble talk on ice

The “prove it” quarter

For months, the skeptics had a pretty simple script: Big Tech was dumping cash into AI infrastructure, and someday—maybe—someone would explain how that turns into actual profit. This week, Meta, Amazon, Microsoft, and Alphabet basically kicked down the door and answered, “Like this.”

The receipts are getting chunky

The headline here isn’t just that capex is huge. It’s that the businesses are already producing enough AI-related revenue to justify the build-out. Microsoft pointed to a $37 billion annualized AI revenue run rate, Amazon said AWS AI services are running above $15 billion, Meta’s AI-powered ad products are scaling fast, and Alphabet said its cloud growth is being pulled higher by generative AI demand.

That matters because it shifts the story from “can they monetize AI?” to “how much can they scale before everyone else catches up?” And that’s a very different ballgame for your portfolio.

Capex, but make it organized

The other big takeaway: this isn’t a blind leap. Management teams said a big chunk of the spending is already underwritten by customer commitments and backlog. In plain English, they’re not just building an expensive AI playground and hoping the neighbors show up.

• Amazon pointed to massive backlog and chip commitments
• Alphabet highlighted a cloud backlog that ballooned
• Microsoft and Meta both showed AI products moving from demo mode to actual revenue mode

Big picture

The “AI is just 1999 with a better logo” crowd got a rough afternoon. These companies aren’t selling a dream anymore—they’re selling compute, software, ads, and cloud capacity at industrial scale. If the numbers keep holding up, the AI capex boom may be less bubble and more the opening chapter of a very expensive, very profitable arms race.