Meta's AI investments impact profit, with a graph showing a 91% drop, alongside a data center deal.

Editorial illustration for Meta's AI Investments Cut Profit 91% Amid New Data Center Deal

Meta's AI Investments Cut Profit 91% Amid New Data...

By AI Daily Post Edited by Brian Petersen, Editor-in-Chief

July 30, 2026 • 3 min read

Meta's net income fell 83% in the third quarter, dragged down by a one-time $16 billion tax charge tied to the Trump administration's spending bill, plus the company's continued spending spree on AI infrastructure. Capital expenditures for 2025 are now projected to land between $70 billion and $72 billion, and Zuckerberg has told investors to expect even higher spending next year. The company also signed a new arrangement with Blue Owl Capital to help finance a data center buildout in Louisiana, a sign of how much off-balance-sheet financing Meta now needs to keep pace with its AI ambitions.

None of that spending is aimed at a single product launch. Zuckerberg used Wednesday's earnings call to lay out a much bigger bet: that AI will stop being something people occasionally query and start being something that works for them constantly, across finance, health, relationships and daily logistics. He framed WhatsApp and Meta's other messaging apps as the delivery point for that shift, betting that billions of users already checking in with Meta AI there will eventually hand off real tasks to it.

Meta founder and CEO Mark Zuckerberg is trying to sell investors on his prediction for the future — one where billions of people will have their own personal AI agents in the next five years.

Mark Zuckerberg predicts that billions of people will have personal AI agents in five years - TechCrunch AI

Why this matters

A 91% profit drop is a real number, not a rounding error, and Zuckerberg is asking shareholders to eat it on the promise that billions of people will want a personal AI agent within five years. For founders and developers watching Meta's spending, the $14 billion BlackRock deal in El Paso is the tell: infrastructure at that scale only pays off if "selling intelligence" becomes a margin business, not a cost center. That's the bet Meta is making public, and it's worth taking seriously rather than dismissing as hype, because Meta has the user base to actually test the "billions of agents" claim in a way smaller labs can't.

But the math only works if adoption matches the spend, and right now we have a concrete infrastructure commitment paired with a vague consumer-behavior prediction. Anyone building on top of Meta's AI stack, or competing against it, should watch whether usage numbers start closing that gap in the next few quarters. If they don't, this quarter's profit hit won't look like a bold bet.

It'll look like the first of several.