Zuckerberg Admits Meta’s AI Agent Progress Is Falling Behind Schedule

Meta CEO Mark Zuckerberg told employees this week that the company’s push into AI agents hasn’t moved as fast as leadership had hoped. Speaking at an internal town hall, he admitted that agentic development has barely picked up pace over the past four months, and that the major restructuring built around this bet hasn’t paid off yet.

Layoffs And Reorganization Under Scrutiny

Zuckerberg also conceded the recent shake-up, which included cutting about 10% of Meta’s workforce and shifting roughly 7,000 employees onto AI teams, wasn’t handled as smoothly as it should have been. He noted that back in January and February, top executives were anxious about falling behind competitors and were especially impressed by tools like Anthropic’s Claude Code at the time.

What Comes Next For Meta’s AI Push

Despite the setbacks, Zuckerberg expects real gains from Meta’s AI spending, projected near $145 billion this year, within the next three to six months. Separately, CTO Andrew Bosworth confirmed a review found no employee data was used in AI training after the company’s mouse-tracking tool was paused, and said any restart would be opt-in only.