Meta Plans 60% Team Cuts, Replacing Staff With AI

Meta’s ambitious attempt to overhaul its workforce by replacing human labor with AI agents has been scaled back after the technology failed to meet performance expectations. The initiative, internally known as Project OT, was designed to reduce team sizes by up to 60% in favor of smaller groups of highly skilled employees managing virtual workers. However, the company abandoned the most aggressive phase of this restructuring after internal data revealed that the AI systems were generating significant operational risks without delivering proportional value.

The strategy was developed at Mark Zuckerberg’s Hawaii residence in January. It envisioned a two-stage reduction process: an initial wave scheduled for May and a second follow-up in November. This plan included not only layoffs but also the closure of open roles and the removal of underperforming staff. Internal projections suggested the cuts could match or exceed the roughly 25% reduction Meta executed three years prior. On May 19, hours before the first wave of terminations, Zuckerberg halted the November planning. The following day, the company cut approximately 8,000 jobs, representing about 10% of its workforce, while simultaneously reassigning 7,000 employees to AI-focused roles.

The Origin of the Restructuring

The impetus for Project OT came from executive observations of how artificial intelligence startups operate. Senior leaders, including Chief Data Officer Alex Schultz and Head of Product Naomi Gleit, traveled to Asia to study these models. They admired how certain organizations had built their organizational charts around AI integration rather than traditional hierarchies. Gleit noted during a June interview that her time in Meta’s Singapore office inspired changes in California and New York teams. She described many of these ideas as emerging from the bottom up within those regional offices.

Despite the high-level enthusiasm, the execution faced immediate hurdles. Meta confirmed the existence of the year-long effort to cut costs and redesign structures but clarified that the 60% figure applied only to specific scenarios within certain teams, not the entire company. Several major units were excluded from the exercise. In a statement regarding the cancellation of the second wave, the company explained: “This ultimately resulted in moving thousands of employees to do priority work on several newly-established teams, as has been publicly reported. Ultimately, we didn’t move forward with every scenario from the exercise, and it was never assumed we would.”

Performance Gaps and Security Risks

Internal metrics quickly undermined the justification for massive workforce reductions. Data indicated that AI agents were producing volume rather than meaningful output. According to a June post by Chief Technology Officer Andrew Bosworth, code changes to internal platforms rose 220% year-over-year, yet user-facing features improved by only 36%. This disparity led to increased instability. An April report warned that unchecked AI agents were performing large-scale actions that humans would typically avoid. Consequently, major technical and security incidents, including service disruptions and potential data leaks, increased by 40%, while the time spent resolving them grew by 70%.

Reliability concerns escalated throughout the spring. By June, hackers exploited Meta’s new AI customer support bot to gain access to high-profile Instagram accounts, including the dormant Obama White House page. Infrastructure teams had raised warning signs as early as March, but the pace of implementation outstripped the system’s ability to maintain stability. Meta declined to comment further on the internal data surrounding these incidents.

Employee Sentiment and Structural Changes

As the project progressed, employee morale deteriorated significantly. Many engineers felt they were effectively training their replacements by writing software puzzles used as training data. Others criticized the mandatory tracking software installed on US devices, which captured keystrokes and mouse clicks to teach AI systems how to use computers. Meta paused this monitoring program in June, though twenty-six employees have since filed lawsuits alleging that AI systems targeted workers on medical leave for termination.

Internal sentiment scores dropped from 74% favorable to 55% in Meta’s half-year Pulse survey. Staff expressed dissent through symbolic gestures, such as posting images of elephants in response to executive communications. Meanwhile, labor organizing efforts gained momentum within the company.

To replace the traditional org chart, Meta introduced the AI-Native Playbook. This framework proposed shrinking product teams of 10 to 20 specialists into pods of three to five members. Engineers, designers, and product managers would share a unified title of “builder,” eliminating middle management layers. Daily priorities would be set through agent-assisted analysis. By June, at least eleven units had adopted this pod structure, with unit heads overseeing 30 to 50 people each, supported by HR staff and unspecified AI systems. Meta maintained that performance ratings and promotions remain decisions made by people, not algorithms.

Future Outlook and Investment

At a July town hall, Zuckerberg acknowledged that he had misjudged the timeline for AI development. He admitted that the trajectory of agentic capabilities had not accelerated as anticipated over the preceding four months. Despite this setback, he predicted benefits would emerge within three to six months. However, Bosworth later instructed employees in August to stop requesting time off based on expected AI efficiencies.

Public messaging has shifted to emphasize investment in human talent. Recent advertising campaigns highlight Meta’s commitment to people, and Zuckerberg published an essay predicting an abundance of jobs globally, even as individual companies may shrink. He drew parallels to the transition from industrial giants to tech firms, noting that company sizes naturally fluctuate during technological shifts.

Financially, the commitment to artificial intelligence remains robust. Meta plans to invest at least $130 billion in AI chips and infrastructure this year. Its second-quarter filing listed 75,472 employees at the end of June, a figure that still accounted for the approximately 8,000 departures under the May cut. While Zuckerberg has promised no further company-wide layoffs this year, the door remains open for team-level reductions and moves in 2027. The ultimate measure of success will likely be whether feature output ever catches up with the volume of code generated by its AI agents.