Meta cuts 8,000 jobs for AI

Meta is cutting jobs because AI has become an infrastructure race, not just a product race. The company told employees on April 23 that it plans to eliminate 10% of its workforce — about 8,000 people — starting May 20, and it also scrapped hiring for 6,000 open roles. That came after smaller cuts earlier this year and a broader push to replace some contractor-heavy work with AI systems. Why does that matter? Because this is less about one bad quarter and more about Meta changing what it spends money on. On April 29, just six days after the layoff memo, Meta raised its 2026 capital-expenditure forecast to $125 billion to $145 billion, up from $115 billion to $135 billion. The company said the increase reflects higher component prices and added data-center costs to build future AI capacity. In plain English — fewer people, more chips, more servers, more buildings full of compute. The scale is what makes the story land. Meta had 77,986 employees at the end of March 2026. So cutting about 8,000 jobs means removing roughly one in 10 workers from a company that was still growing headcount year over year. At the same time, first-quarter capex alone hit $19.84 billion. That’s the tradeoff investors and employees are staring at: labor is becoming the flexible cost, while AI infrastructure is turning into the fixed priority. But there’s a second layer here. Meta isn’t only trimming employees on payroll. It has also been moving away from some third-party vendors and contractors in content enforcement, with AI taking over more of that work over time. Earlier reports in March said internal planning considered cuts that could affect 20% or more of the company, though Meta publicly pushed back on that framing at the time. What actually materialized, at least for now, is the 10% workforce reduction plus the canceled hiring pipeline. So why now? Basically, Meta is trying to win two races at once. One is the public AI race against OpenAI, Google, and Anthropic. The other is the quieter infrastructure race — securing enough GPUs, custom silicon, power, and data-center capacity before those inputs get even more expensive or harder to find. When Zuckerberg talks about “personal superintelligence,” that vision doesn’t just require better models. It requires an absurd amount of compute behind the scenes. The market reaction shows the tension. Meta’s business is still throwing off huge revenue and profit — $56.31 billion in Q1 revenue, up 33% year over year — but investors still flinched at the spending outlook. That’s because AI capex this large can squeeze free cash flow for a long time before the payoff is obvious. The company is telling Wall Street to trust that the infrastructure binge will produce future dominance. Wall Street is asking how many people and how much margin get sacrificed on the way there. The bottom line is simple. Meta is reallocating from humans to compute. Not because the business is collapsing — turns out it’s doing very well — but because management thinks the next moat is owned by whoever can afford the biggest AI machine.