Meta is laying off 8,000 people. Business has never been healthier.

Meta made $201 billion in revenue last year. $60 billion in net income. On May 20, 2026, it began laying off 8,000 people — 10% of its workforce — not because the business is failing, but to fund $125 to $145 billion in AI infrastructure spending this year. The memo to employees was precise: the cuts are "part of our continued effort to run the company more efficiently and to offset the other investments we're making."

Read that carefully. The humans are the offset. The AI is the investment.

Meta is not alone. So far in 2026, almost 110,000 workers have been laid off across 137 tech companies — after 125,000 cuts in 2025. Amazon is cutting 16,000. Microsoft offered buyouts to 7% of staff. Salesforce cut 1,000 roles linked to AI automation. The pattern is consistent: profitable companies eliminating people not because the business demands it, but because the balance sheet can be redeployed toward infrastructure.

The financial logic is clean. But inside Meta, the reality looks different. The company deployed surveillance software on US employees' laptops in April — tracking activity to train its AI models. Performance reviews now evaluate whether employees use AI. Median compensation fell from $417,400 in 2024 to $388,200 in 2025. Raises were cut 10% in 2025 and a further 5% in February 2026. Then the layoff announcement arrived. Employees built countdown websites to May 20. One carried the header: "Big Beautiful Layoff." That is not the behaviour of a motivated workforce. And motivation is precisely what Meta needs from the 70,000 people still at their desks.

This is not unique to Meta. Research from Leadership IQ found that 74% of employees who kept their jobs after layoffs reported declining personal productivity. 71% reported declining motivation. 77% observed more errors around them. The Lattice 2024 Strategy Report found morale takes four months to a year to recover. A peer-reviewed NIH study found job insecurity specifically undermines change-oriented behaviour — the willingness to take risks and challenge assumptions. Those are the capacities that distinguish humans from the work AI handles. A workforce in fear complies. It does not innovate. Compliance is what AI already does, faster and cheaper.

One variable nobody is modelling. Meta's CFO said executives "don't really know what the optimal size of the company will be in the future" — and that they have "continued to underestimate compute needs." The AI bill keeps growing. If frontier AI costs continue to exceed projections — and they have, consistently — the efficiency calculation changes. The humans are gone. The bill got bigger.

Meta is betting $145 billion that AI delivers the output those 8,000 people would have produced. The bet may pay off. But it is being placed on top of a workforce that knows their employer views them as a budget line — now expected to produce the creative, risk-taking work no spreadsheet can mandate.

The efficiency is real. The cost is being paid by the people the model did not account for.