Meta Fires 8,000 Employees to Fund $145B AI Surge After Spying on Their Workday Data to Train Models

Meta is set to begin laying off about 8,000 workers this week as it pours more money into artificial intelligence, with the latest cuts expected to hit roughly 10 per cent of the company's workforce and leave a further 6,000 roles unfilled. The move lands as the company pushes its 2026 capital spending plan as high as $145 billion, and after reports that it is also using employee workplace activity data to help train internal AI models.

Meta was preparing to tighten headcount while widening its AI bet. Reuters reported in March that the company was planning sweeping layoffs as AI costs mounted, and later coverage said the first wave of cuts was due to begin on 20 May, with more reductions possible later in the year. Nothing is confirmed yet on the full scale or timing of every subsequent round, so the safest reading is that this is the beginning of a longer restructuring rather than a one-off purge.

Meta raised its 2026 capital expenditure forecast by as much as $10 billion last month, lifting projected spending to as high as $145 billion. That kind of money does not disappear into the ether.

It is being channelled into data centres, chips, infrastructure and the still-uncertain machinery of AI. The company, quite plainly, has decided that the next phase of growth will be bought with capital rather than payroll.

Meta And The Workplace Data Debate

An internal AI initiative at Meta has become a flashpoint for concern among employees and outside observers. The project envisages a tool that would log workers' mouse movements, clicks, keystrokes and periodic screen images, with the stated aim of using that data to train models to carry out routine digital tasks more efficiently.

Meta has said the information would be used solely for model training rather than for monitoring individual performance, and insists that safeguards have been built in to shield sensitive content.

Meta wants its AI systems to learn how people actually use computers, because that is the gap between a flashy demo and a useful product. What is less tidy is the fact that the company is apparently asking employees to generate the training data while also trimming thousands of their jobs.

The layoffs themselves are expected to affect about 8,000 employees, or roughly 10 per cent of the company's global workforce. Meta has also decided not to fill around 6,000 open positions, a reminder that job losses do not always arrive with a single announcement.

Meta And A Brutal Market

Layoffs.fyi says nearly 110,000 job cuts have been announced across 137 tech companies so far this year. That figure gives the wider backdrop a sharper edge. The tech sector is still trying to convince investors that AI can justify colossal spending, but the route to that future is looking messy and, for many employees, expensive in the most literal sense.

For investors, the question is whether Meta's heavy AI spending will pay off quickly enough to protect margins and keep growth on track. For employees, the question is simpler and more immediate.

If a company is buying the future at $145 billion a year, who exactly gets to stay in the room while the bill is settled? Meta has not said that every cut is final. The company wants leaner operations, more AI, and fewer people doing the work between the two.

Meta did not offer a public comment beyond saying the reductions are part of a push to operate more efficiently while redirecting resources towards AI-related initiatives. The markets may applaud the discipline.

Meta's strategy is becoming clearer by the week, and it is not especially subtle. Spend more, automate more, hire less, and make the workforce help train the machines that may one day replace parts of it.

Meta's latest gamble is that Wall Street will see efficiency and progress where thousands of workers will see something much less flattering.