Meta Layoffs Send More Than 7,000 Workers to AI Teams
Meta layoffs are moving more than 7,000 workers into new teams focused on AI agents and cloud infrastructure as the company reorganizes around artificial intelligence. Late last week, engineers received notice of the reassignment, and those selected were told they would report to the new teams by the end of this week.
Peter Hoose on Meta's AI shift
More than 7,000 workers are being redirected, while some engineers will join a cloud infrastructure team and others will move to Hatch, an internal AI agent team. Peter Hoose, Meta's vice-president of production engineering, said in an internal post that the company's direction has changed fast: "Our work, infrastructure and our products are fundamentally changing as a result of the continued acceleration of AI".
That move follows an earlier reshuffle of at least 1,000 engineers onto Applied AI, Meta's data labeling team. At first, those employees could volunteer, but workers were later told, "Transfers aren’t optional."
Cloud infrastructure and Hatch teams
About 25 people will sit on each of the two new teams, a scale that shows how tightly Meta is concentrating its AI buildout. The selected engineers are being funneled into roles tied to infrastructure and agents rather than staying in broader engineering groups.
Meta is also pulling some managers away from direct reports and into roles where they are expected to produce work more than oversee others. That flattening comes as the company prepares to cut approximately 10% of its workforce this week, even after reporting record earnings during the first three months of 2026.
MCI reaches Meta workers
Meta has begun rolling out Model Capability Initiative, or MCI, a monitoring tool that tracks mouse movements, keystrokes, laptop open and close activity, and copy-and-paste actions. The company says the data is fed into an AI model as training data, and a spokesperson said the tool captures inputs on certain applications to help train its models.
"If we’re building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them – things like mouse movements, clicking buttons and navigating dropdown menus," the spokesperson said. "To help, we’re launching an internal tool that will capture these kinds of inputs on certain applications to help us train our models." The same spokesperson added, "There are safeguards in place to protect sensitive content, and the data is not used for any other purpose."
For employees, the immediate issue is not just which team they join, but how much of their work will now be measured, reassigned and restructured around AI priorities. With layoffs, transfers and surveillance tools all moving at once, Meta’s workforce is being pulled into the same shift from three directions.
