Meta Targets Up to $145 Billion in AI Spending Despite Shedding 20 Million Daily Users

Meta’s Q1 2026 earnings delivered a contradictory message to investors: revenue surged 33% year-over-year to $56.3 billion while the company simultaneously disclosed a loss of 20 million daily active users and plans to spend up to $145 billion on AI infrastructure in 2026. Markets responded by sending Meta’s stock down more than 7%.

The Capex Revision That Spooked Investors

The most consequential disclosure from Meta’s earnings call wasn’t the revenue beat — it was the spending escalation. According to The Verge AI, Meta raised its 2026 capital expenditure guidance to a range of $125–145 billion, a $10 billion increase over prior estimates. Chief Financial Officer Susan Li explained that Meta had “underestimated our compute demand in the past,” a candid admission that obligates the company to dramatically higher infrastructure outlays driven by rising component costs and future data center expansion.

Revenue Growth Can’t Mask User Losses

Meta’s advertising business remains a powerful engine: $56.3 billion in quarterly revenue, up from $42.3 billion in the same period last year, represents the company’s fastest growth rate since 2021. Yet the user metrics tell a more complicated story. Meta’s bundled “Family daily active people” metric — which aggregates daily users across Facebook, Instagram, WhatsApp, and Messenger — dropped by 20 million quarter-over-quarter. The Verge AI reports that Meta attributes the decline to network access failures in Iran and a WhatsApp block imposed in Russia. Because Meta reports only this blended figure, it is structurally impossible for outside observers to determine which individual platform bore the steepest losses — an approach that offers the company considerable interpretive latitude.

Reality Labs Adds to the Pressure

Meta’s hardware division reported a $4.03 billion operating loss for the quarter, a figure made more striking by the two rounds of layoffs that have swept Reality Labs since January. The unit’s sustained losses stand in sharp contrast to the accelerating AI infrastructure side of the business, suggesting an internal reprioritization is already underway.

Why This Matters

Meta’s Q1 results crystallize a tension defining the large-platform AI era: enormous infrastructure commitments are being made before the revenue streams that justify them fully materialize. The fact that Meta’s CFO openly acknowledged underestimating compute demand raises an unsettling follow-on question — if the revised $125–145 billion range proves similarly insufficient, subsequent corrections could be even larger. For investors and industry observers alike, Meta’s upward capex revision reinforces that AI infrastructure demand continues to surprise even its largest and best-resourced buyers.