Meta Beats Revenue Estimates but Misses on EPS as AI Spending Guts Profitability

Meta reported Q2 2026 revenue of $60.8 billion, up 28% year-over-year and ahead of the $59.5 billion Wall Street expected. Advertising revenue alone hit $59.3 billion, topping the $59.07 billion estimate. On the top line, Meta’s core business is still growing faster than analysts modeled.

The bottom line tells a different story. Earnings per share came in at $6.18, well below the $7.14 estimate — a 13% year-over-year decline. Net income fell 14% to $15.8 billion, dragged down by a $2.4 billion legal charge, severance costs, and rising research spending. Revenue growth didn’t translate into profit growth; it got absorbed almost entirely by the cost of building AI infrastructure.

Capex is the mechanism. Meta spent $31.1 billion on capital expenditures in Q2 alone, and raised its full-year 2026 guidance to a $135-145 billion range, up from the $125-145 billion it had previously set. Nearly all of that spending is going toward data centers and AI compute — the same GPU and memory supply chain currently facing the price hikes squeezing Nvidia and AMD’s consumer lineup. Some analysts now expect Meta’s capex to keep climbing toward $215 billion as the AI buildout continues, which would make this year’s guidance a floor, not a ceiling.

This is the tension every major AI lab and hyperscaler is navigating right now, and Meta’s quarter puts hard numbers on it: the AI infrastructure bet requires spending that outpaces the revenue it’s currently generating, and investors are starting to price that gap directly into earnings reactions rather than giving companies a pass on “growth investment.” A 28% revenue beat used to be an unambiguous win. Here it wasn’t enough to offset a capex number growing even faster.

For teams evaluating AI infrastructure costs of their own — whether that’s cloud GPU rental, on-prem hardware, or API spend — Meta’s numbers are a useful external benchmark: even at hyperscaler scale and hyperscaler negotiating leverage, capex growth is currently outpacing revenue growth in AI infrastructure specifically. That’s not a signal to avoid AI investment, but it is a signal that the unit economics haven’t settled yet, and betting on “it gets cheaper next quarter” isn’t currently supported by what the biggest spenders are seeing.

Sources

Related reading