Meta CEO Mark Zuckerberg has abandoned plans for a second wave of layoffs after employees rebelled and autonomous AI agents failed to deliver the expected productivity gains, according to a Reuters investigation based on internal documents and interviews with more than 20 people familiar with the company’s operations .
The plan, codenamed Project OT (Organization Transformation), was developed during an executive retreat at Zuckerberg’s Hawaii compound in January 2026. It envisioned a future where AI agents would take over much of the routine work performed by thousands of employees, with smaller teams of highly skilled workers supervising AI-powered operations .
Internal scenario-planning exercises considered reducing the size of some Meta teams by as much as 60 per cent. The restructuring was planned in two waves, with layoffs beginning in May and another round scheduled for November .
However, hours before the first round of cuts on May 19, Zuckerberg abandoned plans for the second wave, according to an internal document reviewed by Reuters. Meta ultimately laid off about 10 per cent of its workforce the following day .
The reversal came as employees increasingly believed that the company’s AI transformation was partly intended to replace them. At the same time, internal data showed that autonomous AI “agents” at the centre of the strategy were not producing the expected productivity gains .
Internal data revealed that while code modifications had surged by 220 per cent, features actually delivered to users had increased by only 36 per cent. Major technical and security incidents had risen by 40 per cent in a year, and the time spent resolving these issues had jumped by 70 per cent .
Staff morale took a significant hit, with Meta’s internal satisfaction score dropping from 74 per cent to 55 per cent in six months. Many employees flooded Meta’s internal communications network with angry messages and memes, including images of elephants to symbolise the layoffs that no one wanted to address .
In July, during an internal meeting, Zuckerberg acknowledged timing errors and admitted that the technology for AI agents had not accelerated as expected. In August, he repositioned his public messaging, predicting an “abundance of jobs” while conceding that each company might eventually employ fewer people .
The company also came under scrutiny for installing tracking software on US employees’ devices to capture their keystrokes and mouse clicks to train AI agents. This was later suspended after concerns were raised .
Meta confirmed the existence of Project OT in response to the Reuters investigation, describing it as a year-long initiative focused on cost reduction, restructuring teams and moving employees into priority areas, including the production of training data for AI models .
However, the company emphasised that performance rating and promotion decisions were made by people, not AI. “Performance rating and promotion decisions were and are made by people, not AI,” Meta said in a statement .
