Meta Platforms reported second-quarter revenue of $60.8 billion on Wednesday, a 28 percent increase from a year earlier, but the headline number masked a dramatic deterioration in the company's financial flexibility as artificial intelligence spending consumed nearly all of its cash generation.
Free cash flow plummeted to roughly $784 million in the quarter, down from more than $12 billion in the prior three months, as Meta poured capital into data centers, networking equipment, and a new custom AI chip called Iris. The company raised its 2026 capital expenditure floor to $130 billion, with a projected range of $130 billion to $145 billion, a figure that would dwarf the annual budgets of most nations.
The spending surge reflects Meta's bet that owning the full AI stack, from silicon to software, will insulate it from supply constraints and reduce its dependence on Nvidia. The company confirmed plans to begin manufacturing the Iris chip in September, with the goal of doubling its computing capacity to 14 gigawatts next year. The announcement builds on earlier reports that Meta had accelerated its custom silicon program after struggling to secure enough high-performance GPUs.
Investors reacted harshly. Meta shares fell between 8 and 10 percent in after-hours trading, wiping out tens of billions in market value. Analysts said the sell-off reflected not just the magnitude of the spending increase, but also softer-than-expected third-quarter revenue guidance of $61 billion to $64 billion, which fell short of Wall Street consensus.
The quarter highlights a growing tension among technology giants. Companies like Meta, Google, and Microsoft are locked in an arms race to build the largest AI training clusters, even as the return on those investments remains uncertain. Meta's advertising business continues to generate strong growth, but the cash it produces is being reinvested faster than it can be counted.
For the broader technology sector, Meta's results serve as a warning. The AI infrastructure boom has driven record demand for semiconductors, data center construction, and power generation. But if the largest buyers begin to strain their balance sheets, the ripple effects could be felt across the supply chain, from memory manufacturers to real estate investment trusts that lease server space.
Image source: i.ibb.co