Meta's AI Pivot Is Cratering Employee Morale Even as Profits Soar

Meta is preparing to cut roughly 8,000 employees — about 10% of its workforce — on May 20, and the atmosphere inside the company has reached what workers describe as historically toxic lows. More than a dozen current and former employees paint a picture of a workforce paralyzed by anxiety, resentment, and a growing sense that the company views them as data sources rather than colleagues. The grim irony: Meta just posted nearly $27 billion in profit in the first quarter of 2025, yet median total compensation fell to $388,200 last year from $417,400 the year before, after two consecutive years of stock-grant cuts. Many employees are now actively hoping to be laid off so they can collect the 16-week minimum severance and 18 months of paid health coverage.

A central flashpoint is a mandatory tracking program called Model Capability Initiative (MCI), which Meta quietly installed on US employee laptops to record keystrokes and clicks — data used to train AI models that mimic human computer use. Opting out is not an option, according to three employees. When workers pushed back in internal channels, citing Meta's history of user data breaches, Chief Technology Officer Andrew Bosworth reportedly belittled and berated the dissenters. The tool has not been deployed outside the US, which employees attribute to stronger privacy protections in other jurisdictions. This week, a small protest group began posting flyers in US offices urging colleagues to sign a petition demanding the program's end.

The pay squeeze is compounded by an enormous and visible inequality inside the company. While rank-and-file workers absorbed compensation cuts, Zuckerberg last year offered select AI researchers as much as $100 million annually. Capital expenditures — mostly data center spending — are now projected at $125 billion to $145 billion for the year, up $10 billion from prior guidance. Meanwhile, at least 1,000 top engineers were forcibly transferred last month to a new Applied AI Engineering division, with refusal carrying the threat of layoff — an unusual hardball tactic in Silicon Valley, where engineers typically have relocation options during restructurings. Employees took to calling the move a 'draft.'

Legal and ethical turbulence has added to the gloom. In March, separate state court verdicts in California and New Mexico found Meta liable for product failures that harmed users, resulting in combined damages and civil penalties of nearly $380 million. The verdicts reopened internal wounds over the ethical costs of the company's social platforms. In the UK, frustrated employees are now organizing with United Tech & Allied Workers, the same labor group whose parent organization recently absorbed unionizing Google DeepMind workers. Organizers inside Meta wrote to colleagues that leadership was 'escalating their cruel and shortsighted behaviors.'

Not everyone is miserable: Meta's TBD Lab, which houses frontier AI research talent, appears largely shielded from the broader chaos, and some senior figures describe genuine excitement about what AI tools can enable. Zuckerberg himself told employees during an earnings call that projects once requiring months and dozens of people can now be completed by one or two people in a week — a statement that landed less as inspiration than as a warning to many on the receiving end. The only official HR guidance ahead of next Wednesday's cuts: make sure your personal email address is current in the system, and wait.