$56 billion. That's what Meta recorded in Q1 2026 revenue - a strong quarter by almost any measure. The company then announced it would cut another 8,000 employees.
The apparent contradiction resolves when you look at where the money is going. Meta has committed to spending $60-65 billion on capital expenditure in 2026, the majority of it on AI infrastructure: data centers, custom chips, and the research teams behind the Llama family of open-weight models (AI systems that anyone can download and run, unlike closed API-only models). Maintaining that spending pace while keeping headcount flat doesn't work on a spreadsheet, so headcount is being cut instead.
The Same Playbook, Larger Scale
Microsoft, Google, and Amazon have each run versions of this pattern over the past 18 months: record profits, announced layoffs, AI infrastructure commitment, repeat. The logic is consistent across all four companies: AI infrastructure is a capital investment that compounds returns over time, while human labor costs scale linearly. The math looks better when you can point to AI absorbing work that previously required people.
Meta's AI revenue story is already concrete, not theoretical. Its advertising business grew significantly in Q1 partly because AI-optimized ad targeting is outperforming manually managed campaigns. The better Meta's targeting models get, the more advertisers pay for access to its 3.3 billion daily users. That creates a direct feedback loop: AI infrastructure spend increases ad revenue, which funds more AI infrastructure.
The 8,000 cuts follow Meta's 2023 "Year of Efficiency," which eliminated around 21,000 positions. This round appears concentrated in middle management and operations rather than engineering - roles that became redundant after previous reorganizations consolidated teams.
For people who use Meta's apps daily, the practical outcome is more AI in everything: AI-curated feeds, AI-generated content recommendations, AI-optimized ad delivery. That trajectory was already set before Q1. The profits just mean Meta can execute it faster.