Zuckerberg tells staff Meta’s AI restructuring and “agent” timeline are behind expectations

Meta Platforms has told employees that parts of its push to reorganize around artificial intelligence and to accelerate the development of “AI agents” have not yet delivered on the timeline leadership expected. In remarks discussed by multiple market outlets, CEO Mark Zuckerberg said the company’s expectations for agentic development have lagged over the prior several months and that the broader organizational changes were not “as clean” as planned.

According to a report summarizing a recording reviewed by Reuters and reproduced by Benzinga, Zuckerberg told staff that “the trajectory of the agentic development over at least the last four months hasn’t really accelerated in the way that we expected.” He added that the company’s bets on the new organizational structure “haven’t come to fruition yet.” The comments suggest Meta’s internal goals for faster progress in building AI systems that can take actions, not just generate text, have been harder to meet than anticipated.

The same report also tied Zuckerberg’s admissions to Meta’s May restructuring. In May, Meta cut about 10% of its global workforce and reassigned roughly 7,000 employees to AI-focused teams, as part of a broader effort intended to support growing investment in artificial intelligence. Zuckerberg reportedly acknowledged that the reorganization did not unfold as smoothly or on the timing executives had projected, pointing to internal execution risk rather than a simple slowdown in hiring or funding.

Meta’s stock performance in 2026 has reflected investor skepticism around how quickly its AI strategy would translate into results. Through July 2, Meta shares were described as down about 11.7% year to date, trailing an S&P 500 advance of roughly 9% over the same period, according to Yahoo Finance’s market recap. While stock moves do not prove what is happening inside the company, they set the backdrop for why investors have been watching whether Meta’s operational shift toward AI is translating into measurable momentum.

Zuckerberg’s comments also arrive amid a broader public narrative that Meta is betting heavily on AI infrastructure. Some coverage has pointed to projections that Meta could spend as much as $145 billion on AI infrastructure in 2026, though those figures were described as “projected” or “as much as” ranges in secondary reports rather than confirmed by a specific company disclosure in the materials reviewed for this story. The company’s intent, in general terms, is to build and deploy AI systems at scale, which typically requires substantial compute and engineering resources.

AI agents, the centerpiece of Meta’s messaging in these remarks, are typically understood as systems that can perform tasks toward a goal using tools and intermediate steps, rather than simply producing a response. For Meta, the strategic relevance is straightforward: if agents work reliably, they can be used across products such as recommendation, advertising workflows, customer support automation, and developer tooling. But turning agent concepts into dependable products often involves long development cycles, safety and evaluation work, and integration with existing services.

What Meta did not disclose in these reported remarks is as important as what it did say. The town hall summary, as described by outlets, focused on pace and internal execution rather than providing quantified milestones, product timelines, or detailed performance benchmarks for agents. There was also no disclosure of whether the company expects a near-term course correction, which specific agent capabilities are behind schedule, or how management is reallocating resources in response.

Going forward, the key question is whether Meta’s next external updates will show signs of acceleration after internal reassignment and cost actions. Investors and customers will likely look for clearer indicators such as new AI features, measurable improvements in agent reliability, or evidence that restructuring has reduced friction in product delivery. Until then, Zuckerberg’s comments effectively confirm that at least some parts of Meta’s AI reorganization and agent timeline have been slower than planned, and that operational execution remains a moving target.