Zuckerberg Admits Meta’s AI Restructuring Did Not Go as Planned

Meta restructuring problems have forced Mark Zuckerberg to admit that the company’s recent internal overhaul did not go as smoothly as expected.

According to Reuters, Zuckerberg acknowledged that Meta executives misjudged the timing of the changes. The company now wants to soften some parts of the restructuring while keeping its broader AI-first strategy in place.

The admission comes after a difficult period for Meta employees, who faced layoffs, team changes, and growing pressure around the company’s AI ambitions.

Meta’s AI Restructuring Caused Employee Pushback

Meta laid off about 10% of its global workforce as part of its latest restructuring effort.

The company also reassigned around 7,000 employees to AI-focused teams in May. These moves aimed to push more resources into artificial intelligence, which Zuckerberg sees as one of Meta’s most important long-term priorities.

However, the changes caused frustration inside the company. Employees raised concerns about morale, job security, and the speed of the transition.

Zuckerberg told employees in May that he did not expect more companywide layoffs this year. Still, some workers reportedly remained skeptical after the scale of the earlier cuts.

Zuckerberg Says Some AI Bets Have Not Paid Off Yet

Zuckerberg also admitted that Meta’s AI agent development has not moved as quickly as expected over the past four months.

The company had hoped that the new structure would help teams build and ship AI products faster. Instead, Zuckerberg said some of Meta’s bets on the reorganization have not produced results yet.

He also said executives worried earlier this year that Meta would not adapt quickly enough if it kept its old structure.

That concern helped drive the major reshuffle, even though the execution created new internal problems.

Meta Still Plans To Spend Heavily On AI

The restructuring partly helped Meta fund its massive AI infrastructure plans. The company expects to spend up to $145 billion on AI infrastructure this year.

That spending includes the computing power, data centers, and technical systems needed to support Meta’s AI tools and future agent-based products.

Zuckerberg said he expects the company to see more meaningful benefits from its AI investments within the next three to six months.

That timeline suggests Meta wants quick proof that the restructuring and spending push can deliver real product improvements.

Meta Reviews Controversial Mouse-Tracking Program

Meta CTO Andrew Bosworth also addressed the company’s controversial mouse-tracking software.

The software tracked employee mouse movements and digital activity for AI training purposes. Meta paused the program last month after a data security incident exposed sensitive data.

Bosworth said Meta’s review found that no employee data had been included in AI training.

The program still raised concerns among workers because of how closely it monitored employee activity.

Mouse-Tracking Program Could Return As Opt-In

Bosworth said Meta may bring the mouse-tracking program back, but only as an opt-in system.

That means employees who feel comfortable with the program could choose to contribute. Employees who do not want to participate would not have to take part.

This would mark a major change from April, when Bosworth reportedly told U.S. employees that they could not opt out.

For now, Meta appears focused on moderation rather than reversal. The company wants to adjust how it manages teams, employee data, and AI development without stepping back from its larger AI investment plan.

In other news, Google has reportedly limited Meta’s Gemini AI access after the company exceeded capacity.