Meta CEO Mark Zuckerberg acknowledged during a recent internal town hall that the company’s push into artificial intelligence agents has not met internal performance expectations. According to reports, Zuckerberg admitted that top executives miscalculated the timeline for AI development, noting that progress has not accelerated as anticipated over the past four months.
This admission follows a period of significant volatility at the company. Earlier this year, Meta initiated a restructuring plan that included laying off 8,000 employees—approximately 10% of its global workforce—and reassigning 7,000 staff members to AI-focused teams. These efforts are part of a massive infrastructure investment strategy projected to reach $145 billion in 2026.
Internal morale has become a focal point of the company’s challenges. Chief Technology Officer Andrew Bosworth recently described the rollout of the company’s ‘Applied AI’ division as ‘atrocious,’ citing chaotic transitions and poor communication that undermined employee trust. In response, leadership has pledged to improve internal culture by limiting manager direct reports, offering AI coaching tools, and reinstating office perks.
Additionally, Bosworth addressed concerns regarding the company’s mouse-tracking software, which was previously paused due to data exposure fears. He confirmed that no employee data was used for AI training and announced that if the program is reinstated, it will operate on a strictly opt-in basis.