Particle.news

Meta Scales Back AI-Native Overhaul After Layoffs and Staff Revolt

Employee protests, weak agent reliability and a spike in technical incidents have forced leadership to temper plans to replace large swaths of staff with AI.

Overview

  • Project OT was an internal January planning effort that modeled making Meta “AI-native” by shifting work to AI agents and shrinking some teams by as much as 60 percent in scenario exercises.
  • Meta cut about 10 percent of its workforce on May 20, 2026, closed thousands of open roles and reassigned roughly 6,000–7,000 employees into a new Applied AI unit instead of carrying out the larger reductions originally envisioned.
  • A planned second wave of layoffs scheduled for November was halted after sustained employee pushback and internal leaders signaled they would not pursue further company-wide cuts in 2026.
  • Staff revolts focused on a keystroke and screen-tracking program used to train models, and Meta paused that tracking as internal morale fell from 74 percent favorable to 55 percent on Pulse surveys.
  • Internal data showed AI work produced a 220 percent rise in code changes but only a 36 percent rise in user-facing features and a roughly 40 percent increase in major technical and security incidents, while Meta continues very large AI infrastructure spending that has drawn investor scrutiny.