Mark Zuckerberg told Meta staff at an internal town hall on Thursday that the company's work on AI agents has not moved as fast as he had hoped, according to a Reuters report on the meeting. He said that over the past four months, agent development had not accelerated in the way the company expected, and that the benefits Meta anticipated from a major AI focused reorganization had not come to fruition yet. For a company that has spent the year reshaping itself around AI, it is an unusually direct acknowledgment from the person who ordered the change that the early results have fallen short of the plan.
The reorganization he was talking about was not a small one. Earlier in the year Meta laid off about 10 percent of its global workforce, on the order of 8,000 people, and in May it reassigned roughly 7,000 employees into AI focused teams, including one named Agent Transformation. Zuckerberg told staff that the process had not been as clean as it should have been and that leadership had miscalculated the timing of the changes. In other words, the disruption was real and immediate, while the payoff the reshuffle was supposed to unlock has been slower to show up.
What makes the comments land is the sheer size of the bet sitting behind them. Meta is projected to spend as much as 145 billion dollars on AI infrastructure this year, one of the largest single commitments in an industry where big technology companies are together expected to pour more than 700 billion dollars into AI. When a company is spending at that level and the chief executive tells his own employees that the agents are not progressing as quickly as expected, it is a data point that cuts against the more confident public messaging the whole sector has been running on.
It is worth being precise about what Zuckerberg did and did not say, because this is not him walking away from the strategy. He told staff that he still expects Meta to see more significant benefits from its AI investments within the next three to six months, which reads less as a loss of faith in AI agents and more as a reset of the timeline. The honest version of the story is that the technology is taking longer to turn into working products and measurable results than the internal plan assumed, and that a rushed reorganization made the near term messier rather than faster.
The reason it matters beyond Meta is that it is a candid reality check from one of the people most invested, literally, in AI agents succeeding on schedule. It fits a pattern that has been building all year, in which the spending and the ambition keep climbing while the delivery of autonomous, agentic systems keeps slipping past the dates that were quietly attached to it. None of this means AI agents do not work or will not pay off, and three to six months is not a long horizon. But when the person writing some of the biggest checks in the industry tells his staff the agents have not accelerated the way he expected, it is a useful correction to the idea that the agent era has already arrived.