Zuckerberg Says Meta 'Miscalculated' AI Overhaul After 8,000 Job Cuts

Meta Chief Executive Officer Mark Zuckerberg has acknowledged that the company's sweeping artificial intelligence (AI)-focused restructuring has not delivered the expected results, admitting senior executives "miscalculated" the timing of the overhaul despite laying off around 8,000 employees earlier this year.

According to a Reuters report, Zuckerberg made the remarks during an internal town hall this week, where he said Meta's AI agents had progressed more slowly than anticipated and the company's ambitious AI plans had yet to generate the expected returns.

The restructuring, announced earlier this year, involved cutting about 8,000 jobs, or roughly 10% of Meta's global workforce, while redeploying another 7,000 employees to AI-focused teams. The changes were aimed at supporting the company's AI infrastructure investments, which are projected to reach about $145 billion in 2026.

Zuckerberg Sees AI Progress Picking Up

During the town hall, Zuckerberg said the development of AI agents had not accelerated as quickly as executives had expected over recent months.

According to Reuters, he told employees that management had been "super optimistic" earlier this year about the rapid progress of AI coding tools such as Anthropic's Claude Code. However, Meta's own investments in AI agents have taken longer than anticipated to deliver meaningful results.

Despite the slower progress, Zuckerberg expressed confidence that the company would begin seeing more substantial benefits from its AI investments within the next three to six months.

The comments come as technology companies continue investing billions of dollars in AI infrastructure while facing growing pressure from investors to demonstrate commercial returns.

Meta Reviews Mouse-Tracking Tool

Meta Chief Technology Officer Andrew Bosworth also addressed concerns surrounding the company's mouse-tracking software, which had been paused following reports that it could expose sensitive information.

According to Reuters, Bosworth said an internal review found that no employee data had been used to train AI models.

He added that if the software is reinstated, participation will be voluntary, reversing the company's earlier stance that did not allow US employees to opt out.

"For people who are comfortable, that's great, they can contribute to this kind of great human survey. To people who are not, it is not an issue," Bosworth said.

AI Rollout Was Poorly Managed

Separately, Wired reported that Bosworth described Meta's rollout of its Applied AI division as "atrocious" in an internal memo to employees.

The division, created in March with around 6,500 engineers and product managers, was intended to accelerate Meta's generative AI efforts. However, according to the report, many employees criticised the restructuring as chaotic, with some describing the transition as deeply disruptive.

Bosworth acknowledged that the company had failed to clearly communicate its long-term vision and had weakened employee trust through rapid organisational changes.

According to Wired, Meta now plans to improve internal communication, strengthen career development opportunities and enhance workplace benefits to rebuild morale. The company will also cap managers at 20 direct reports, introduce AI coaching tools and invest in employee facilities, including office amenities, travel budgets and social events.

"I hope we can rekindle the best of the culture we joined," Bosworth wrote in the memo, according to Wired.