Meta Cuts 8,000 Jobs to Fund $125-145B AI Infrastructure Spending
Tags AI · Infrastructure · Enterprise
Meta will lay off approximately 8,000 employees (10% of its ~78,000 workforce) starting May 20, explicitly redirecting payroll savings toward its raised 2026 capex guidance of $125-145 billion for AI data centers and GPU compute, up from $72.2 billion in 2025. Despite the layoffs, Meta reported strong Q1 2026 results: $56.31 billion revenue (up 33% YoY) and $26.8 billion net income. CEO Mark Zuckerberg told employees at an April 30 town hall that Meta's two major cost centers are compute infrastructure and people spending, and investing more in one means less for the other. Employees report anxiety as Meta tracks computer work to train AI models and factors AI tool usage into performance reviews. Four mega-cap tech firms (Amazon, Microsoft, Alphabet, Meta) plan a combined ~$725 billion in 2026 capex, a 77% YoY increase going almost entirely into AI infrastructure.
Technical significance
Meta's layoffs represent the largest explicit reallocation of payroll to AI infrastructure in tech history, with capex nearly doubling year-over-year. The $125-145 billion commitment exceeds the GDP of most countries and signals that AI infrastructure spending has reached a scale where it directly competes with human capital budgets. The tracking of employee computer work for AI training data also raises questions about the boundary between workforce management and data collection.