Mark Zuckerberg told employees at a company town hall on April 30 that Meta has two major cost centers: compute infrastructure and people. He is cutting the people to fund the compute. On May 20, Meta begins laying off 8,000 employees, roughly 10% of its workforce, with a second wave planned for the second half of 2026. The company raised its full-year capital expenditure guidance to $125 billion to $145 billion, up from an already staggering $115 to $135 billion range, nearly double the $72 billion it spent in 2025. Zuckerberg told a company town hall that the layoffs are a direct consequence of the AI infrastructure budget. He was not vague about the trade-off. He named the two line items competing for the same pool of capital: machines and humans. He chose machines. Meta’s 2025 revenue was $201 billion. Its Q4 net income set a record. Free cash flow hit $43.6 billion for the year. The company could fund every dollar of the AI buildout without firing a single person. It is choosing to fire anyway because the roles being eliminated are not the roles that build what comes next. Surviving employees are being reorganized into AI-focused units under the Superintelligence Labs division, led by 28-year-old Chief AI Officer Alexandr Wang. Zuckerberg has said publicly that a small number of talented people working alongside powerful AI systems can accomplish what previously required entire departments. In February, Microsoft AI CEO Mustafa Suleyman told the Financial Times that AI would achieve “human-level performance on most, if not all professional tasks” within 12 to 18 months. Accounting, legal, marketing, project management, anything involving sitting at a computer. Three months have passed. Fortune reported that mounting evidence shows AI “is kind of a bust” for most enterprise deployments. Gartner forecasts that 40% of agentic AI projects will be abandoned by 2027, the same year Suleyman’s timeline predicts human-level performance arrives. The builder and the analyst disagree on the same deadline. The money is following the prediction, not the evidence. Five hyperscalers have committed $725 billion in AI capex for 2026. The tech industry laid off nearly 80,000 workers in Q1 alone. Meta’s layoffs begin three days before its annual shareholder meeting. The restructuring is not a response to financial pressure. It is a reallocation inside the most profitable social media company in history, funded by record earnings, directed toward infrastructure whose returns have not yet materialized at the scale the spending implies. The physical world is straining to absorb the bet. Over 100 UK data centers have requested gas-generated electricity totaling more than 15 terawatt-hours annually, threatening national climate targets. xAI is installing gas turbines in Mississippi to bypass utility constraints. The AI buildout that was supposed to run on renewables is being powered by fossil fuels because the grid cannot deliver clean electricity fast enough. Zuckerberg’s formulation is the clearest statement any CEO has made about what is actually happening. Two cost centers. Compute and people. The company chose compute. It could afford both. It decided it did not need both. The 8,000 people being let go are not casualties of a downturn. They are the output of an optimization function that concluded machines are the better allocation of the next marginal dollar. That function is now running inside every company spending on AI infrastructure. Meta just said it out loud.