Meta’s Q3 2026 Earnings: AI Spend Meets Monetization Reality
By

Meta Platforms reported third-quarter 2026 results after the market close on October 21, just as NVIDIA CEO Jensen Huang’s GTC keynote put a spotlight on the massive infrastructure required to power the next wave of AI. The timing made Meta’s numbers a critical read on whether Big Tech’s AI spending is translating into measurable returns — or simply heavier bills.

What Investors Were Watching

The checklist was clear:

  • Ad revenue strength
  • Capital expenditures (the real cost of AI ambition)
  • Free cash flow
  • Reality Labs performance

Here’s how those key drivers landed.

The Headline Numbers

Meta delivered $64.6 billion in revenue, up 26% year-over-year and ahead of expectations. Advertising revenue accounted for the vast majority at roughly $63.1 billion, also rising 26%. Ad impressions grew 10% while the average price per ad climbed 14%, showing both volume and pricing power.

On the bottom line, the picture was more complicated. Meta recorded a $9.89 billion legal charge related to its multistate attorneys general settlement over youth safety issues. GAAP diluted EPS came in at $3.83. Excluding the charge, adjusted EPS was approximately $7.40. Operating margin compressed to 20% (or about 35% without the legal hit).

AI Infrastructure Costs Take Center Stage

Capex (including finance lease principal payments) jumped to roughly $34.9 billion in the quarter as Meta continued building out servers, data centers, and networking gear for AI training and inference. Free cash flow turned negative at –$2.12 billion, a sharp reversal from the strong positive figure a year earlier. Full-year 2026 capex guidance was narrowed to $135–145 billion.

This is the real-world counterpart to Huang’s GTC message: AI is no longer just software — it is an industrial-scale manufacturing process that demands enormous power, cooling, and compute infrastructure.

Reality Labs Still a Drag

Reality Labs generated $443 million in revenue (down 6% year-over-year) while posting an operating loss of about $4.13 billion. The company now expects Reality Labs losses of $18–19 billion for full-year 2026, with a reduction toward roughly $15 billion in 2027 as priorities shift more toward AI glasses and personal superintelligence initiatives.

Bull vs. Bear Takeaways

Bull case: Robust ad growth, healthy engagement (Family Daily Active People at 3.64 billion), early traction with generative AI tools for advertisers, and the launch of Muse (Meta’s first frontier model) suggest AI is beginning to feed the core advertising engine. Strong Q4 revenue guidance of $73–76 billion also reassured investors.

Bear case: Negative free cash flow, elevated and still-rising capex, a large legal charge, and continued heavy losses in Reality Labs highlight the cost of Meta’s dual bets on AI infrastructure and the metaverse. The company is spending aggressively while cash generation temporarily lags.

The Bigger Picture

Meta’s Q3 results arrived in the same week that Jensen Huang framed AI as a new manufacturing industry built on layered infrastructure. Meta is one of the largest customers of that infrastructure. The quarter showed that advertising remains a powerful cash engine capable of funding the AI buildout — for now. The longer-term test is whether the AI investments (models, agents, recommendation systems, and future glasses) can expand margins and create new high-margin revenue streams fast enough to justify the capital intensity.

For now, the market appears willing to fund the vision. The next few quarters will reveal whether the spend is compounding returns or simply compounding costs.