Meta's AI Bill Hits $145B as Stock Tumbles on Earnings Beat

Meta delivered a blockbuster first quarter on April 29, posting $56.3 billion in revenue, up 33% year over year, with operating income of $22.9 billion and net profit of $26.8 billion. Advertising revenue alone climbed 33% to $55 billion. By almost any normal metric, it was a triumph. But investors did not see it that way.

The reason is the cost of building AI. Meta raised its 2026 capital expenditure guidance to a range of $125 billion to $145 billion, up from the $115 billion to $135 billion it forecast in January and nearly double the $72.2 billion the company spent in 2025. The new ceiling exceeds Meta's combined capex for 2024 and 2025. The company cited higher component pricing — particularly memory — and additional data center costs to support future capacity. Q1 capex alone hit $18.99 billion, a 47% jump from a year earlier.

Wall Street flinched. Meta's stock fell more than 6% in after-hours trading, in stark contrast to Alphabet and Amazon, which both rose on their earnings reports the same week. Analyst Melissa Otto captured the mood, telling reporters investors were left asking "what is the real ROI on all this capex that they're spending." On the call, CEO Mark Zuckerberg deflected a direct question on return metrics as "a very technical question," but said he was "quite comfortable" the AI lab Meta is building "is on track to be a leading lab in the world."

The strategy behind the spending is starting to come into sharper focus. Meta plans to deploy more than one gigawatt of custom silicon co-developed with Broadcom, alongside large purchases of AMD accelerators and new Nvidia systems, in an attempt to wring more efficiency out of each dollar of compute. Zuckerberg has positioned that infrastructure as the foundation for a future generation of personalized AI that can develop, in his words, a "first-principles understanding of what you care about."

Even with the capex shock, Meta is also tightening its belt elsewhere. The company confirmed plans to cut hundreds of jobs in May across sales, recruiting and hardware divisions — the second consecutive quarter in which it has paired record-breaking AI investment with targeted layoffs. Combined with similar capex revisions from Microsoft, Google and Amazon, the Q1 2026 earnings cycle has now committed roughly $650 billion in big-tech AI infrastructure spending for the year, with no end in sight.