Mark Zuckerberg: Meta’s AI agents lag, gains expected in 3-6

Mark Zuckerberg told staff at an internal town hall that the pace of AI agent development had not "accelerated in the way" executives had previously expected, and that the perceived upside of the company’s AI restructuring "had not come to fruition yet," though he said he expected improvements within the next three to six months.

What did Zuckerberg actually say about AI agents?

Zuckerberg said the pace of AI agent development had not "accelerated in the way" company leaders had expected, and he added the upside of the AI-focused reorganization "had not come to fruition yet," while forecasting visible improvements in three to six months. Reuters reported the town hall comments and TechCrunch reached out to Meta for comment.

The town hall remarks framed recent personnel moves: earlier this year Meta laid off some 8,000 employees, described in the article as approximately 10% of its corporate workforce, and reassigned another 7,000 employees into various AI groups, including one called Agent Transformation. Bloomberg has described the months-old AI unit as a "soul-crushing gulag," according to some engineers assigned to it.

How has Meta reorganized and funded its AI push?

Meta cut roughly 8,000 corporate jobs and reassigned about 7,000 people into AI roles as part of an internal refocus on artificial intelligence, and the company is expected to spend as much as $145 billion on AI infrastructure this year. Those moves included the creation of teams such as Agent Transformation to concentrate work on agentic systems.

Executives made the headcount changes, Zuckerberg said, because they "were worried that we weren’t going to move fast enough to adapt" to the changing tech landscape. The town hall comments tied those workforce decisions directly to the expectation that an AI-first structure would accelerate product and engineering outcomes.

Why it matters

Large-scale layoffs and the reassignment of thousands of employees create immediate pressure to convert investment into results. Meta’s expected AI infrastructure spend of as much as $145 billion this year raises the stakes: the company has shifted people and capital toward agent work, and Zuckerberg’s public acknowledgement that progress has lagged tightens the timeline for showing returns. The combination of heavy spending, major staff moves, and internal morale concerns amplifies the cost of further delays.

What to watch

Zuckerberg’s three- to six-month window is the clearest near-term milestone: look for concrete signs of improved agent performance, internal metrics or product rollouts tied to the Agent Transformation teams during that timeframe. Also watch whether leadership revises staffing or structure again if the expected gains do not materialize.