Meta’s AI Push Sparks Layoffs, Internal Turmoil, and Questions Over the Payoff

Meta CEO Mark Zuckerberg has admitted the company made mistakes as it tries to reposition itself as a major force in artificial intelligence. In an internal memo sent to employees, he wrote, “Given the complexity of the changes, we made mistakes and almost certainly will make more,” signaling both acknowledgement of problems and an effort to project stability.

The remarks come after a turbulent period for the tech giant. In May, Meta cut about 8,000 jobs, roughly 10% of its workforce, and reassigned another 7,000 employees to AI development and training roles. Altogether, nearly one-fifth of the company’s employees were affected. The restructuring reflects the high cost of building AI infrastructure and Meta’s attempt to catch up with or surpass rivals such as Google, OpenAI and Anthropic.

The company has said its AI investment could reach as much as $145 billion by 2028. But the internal mood has deteriorated, according to U.S. media reports, with some employees describing the atmosphere as among the worst in the company’s history. Workers have protested the use of AI-based monitoring tools, while others say they remain uncertain about their new roles after the mass transfers.

Meta is also facing technical setbacks. Its flagship advanced-model project, code-named “Avocado,” has been delayed after failing to meet performance expectations against competitors. That has raised fresh doubts about whether Meta can keep pace in frontier AI development. Company officials say Zuckerberg’s softer tone is meant to reassure remaining staff and reinforce the message that no broad layoffs are expected this year, but the central question remains whether the massive AI bet will pay off or prove too costly.