Meta Cuts 8,000 Jobs — 10% of Its Workforce — as 2026 AI Spending Hits $135B

Meta told its employees on Thursday that approximately 8,000 of them — 10% of the global workforce — will be let go starting May 20. The company is also canceling 6,000 open job requisitions. Combined, this is one of Meta’s largest headcount reductions since the 2022 “Year of Efficiency” layoffs that cut 21,000 jobs across two rounds.

The stated reason is efficiency. The unstated reason is arithmetic. Meta’s 2026 capital expenditure guidance is now $115 billion to $135 billion — almost entirely AI infrastructure. Data centers, custom silicon, training compute, and the engineering talent to run it. The company cannot grow headcount in product and operations while simultaneously funding that kind of capital program without pressure on margins. Something had to give.

Mark Zuckerberg laid the groundwork for this at the start of the year. He told investors that AI was beginning to replace the need for large teams on projects that previously required many engineers. That wasn’t a vague forecast — it was a warning that the productivity math would eventually justify cuts.

US employees will receive 16 weeks of base pay severance plus two additional weeks per year of service, with 18 months of continued health coverage. That’s a generous package by industry standards, which typically signals the company expects the PR hit and is trying to limit it.

The broader context: Meta isn’t alone. Oracle announced 30,000 cuts earlier this year to fund data center investment. Snap cut 1,000 jobs and credited AI for writing 65% of its code. The pattern is consistent — headcount that used to produce output is being replaced by compute that does the same work faster and cheaper.

What makes Meta’s case different is scale. 8,000 people is larger than the entire employee count of most tech companies. And Meta’s AI bets are not speculative — Llama 4, Muse Spark, and the newly formed Meta Superintelligence Labs are already shipping. The company is cutting human headcount while simultaneously accelerating AI model output. That is the efficiency trade in practice, not in theory.

The stock rose after the announcement. Markets have seen this movie before: announce layoffs, beat on AI spend, stock goes up. Investors have been trained to interpret headcount cuts as discipline rather than distress — at least when they come from profitable companies with growing revenue.

Affected employees will be notified individually starting May 20. The 6,000 canceled open positions represent roles that will never be filled, many of them in mid-level engineering, sales, and operations. The company is not pausing hiring in AI research or infrastructure.

The question that matters for the next two years: does AI-driven productivity actually allow Meta to do more with fewer people, or does it just do the same things for less cost while the growth ceiling stays the same?