Meta Platforms came significantly closer to a far-reaching AI-led workforce restructuring than previously known, before CEO Mark Zuckerberg reportedly halted a planned second round of layoffs amid concerns over AI performance and growing employee resistance.
According to a Reuters report citing internal planning documents and company discussions, Meta had considered reducing headcount in some teams by as much as 60% under a confidential initiative known as Project OT, short for Organisation Transformation. A second wave of job cuts had been scheduled for November, but Zuckerberg reportedly cancelled the plan hours before the first phase of layoffs took effect in May.
The development offers a rare look into how one of the world's largest technology companies is attempting to balance aggressive AI ambitions with operational realities, workforce concerns and mounting infrastructure investments.
Inside Meta's AI-native vision
Reuters reported that Project OT emerged during a leadership retreat at Zuckerberg's Hawaii estate in January.
The initiative aimed to transform Meta into what executives described as an "AI native" organisation. Under the proposed model, AI agents would take over a substantial portion of routine work currently performed by employees, while smaller teams of generalist workers would oversee operations.
The restructuring blueprint reportedly included:
- Smaller "builder" teams replacing larger specialised groups
- Reduced layers of middle management
- Greater use of AI-assisted analysis in decision-making
- Expanded reliance on AI agents across product and engineering functions
- Workforce reductions that, in some scenarios, could have rivalled or exceeded Meta's 25% workforce contraction in 2023
Meta later acknowledged the existence of Project OT but said the more aggressive workforce reduction scenarios were exploratory options rather than confirmed plans. The company also clarified that the reported 60% reduction applied only to selected teams.
First phase moved ahead, second phase stopped
Meta proceeded with the first phase of Project OT on May 20, cutting roughly 10% of its workforce.
However, Reuters reported that Zuckerberg abandoned the planned November layoffs the night before the initial cuts took effect.
The decision followed growing evidence that Meta's AI systems were not yet delivering the operational gains anticipated by leadership.
Internal performance data reportedly revealed a widening gap between AI-assisted development activity and measurable business outcomes.
Key figures cited by Reuters showed:
- Code changes across Meta's AI platforms and infrastructure increased 220% year-on-year
- New or upgraded user-facing features rose only 36%
- Major technical and security incidents increased 40%
- Employee time spent resolving incidents grew 70%
At a July town hall, Zuckerberg reportedly acknowledged that AI development progress had not accelerated at the pace leadership expected and said the company's investments in AI agents had not yet delivered the intended results.
Employee backlash intensified pressure
Technology performance was not the only factor influencing the decision.
Reuters reported that employee opposition to the transformation strategy intensified as Meta expanded monitoring of workplace activity.
The company reportedly installed software on employees' computers in the United States to track keystrokes and cursor movements. The system was intended to help train AI tools, but many employees feared the technology would ultimately replace their roles.
Internal dissatisfaction grew rapidly.
According to Reuters:
- Employee approval scores in Meta's Pulse survey fell from 74% to 55%
- Labour organising efforts gained momentum
- Some engineering divisions experienced headcount reductions approaching 30% by late May
- Internal communication channels saw a surge in employee complaints
The developments highlighted growing tensions between AI-driven productivity initiatives and workforce confidence.
Billions at stake in Meta's AI strategy
The reported restructuring discussions come as Meta significantly increases spending on AI infrastructure.
The company plans to invest at least $130 billion in AI chips and infrastructure in 2026, according to the Reuters report.
Analysts at LSEG estimate that spending at that scale could consume the company's entire annual operating cash generation.
Meta's broader AI strategy partly depends on the possibility that AI-enabled teams can operate more efficiently with fewer employees. However, the internal productivity data cited by Reuters suggests the economic benefits remain uncertain.
While Wall Street continues to maintain a largely positive view of the company, investors are closely watching whether AI investments translate into measurable gains in productivity, product development and profitability.
Transformation remains on hold, not abandoned
Although Zuckerberg has reportedly told employees he does not expect further company-wide layoffs this year, uncertainty remains around the long-term direction of Meta's workforce strategy.
Reuters reported that smaller AI-assisted teams continue to operate across parts of the organisation, indicating the company has not abandoned its AI-first ambitions.
Instead, the decision to pause the second round of layoffs suggests Meta is reassessing how quickly AI can replace or augment human work at scale.
For now, Project OT appears to have shifted from rapid workforce transformation to a more cautious evaluation of whether AI can consistently deliver the productivity gains needed to justify such sweeping organisational change.