Meta CEO Mark Zuckerberg told employees Thursday that the company's AI agent technology is not advancing as quickly as executives projected, a disclosure that comes as the social media giant pours up to $145 billion into AI infrastructure this year alone. The slower-than-expected trajectory raises questions about when Big Tech's massive AI bets-totaling more than $700 billion across the industry-will translate into functional workplace tools.
Speaking at an internal town hall, Zuckerberg said the "trajectory of the agentic development over at least the last four months hasn't really accelerated in the way that we expected," according to a recording heard by Reuters. He added that the company's restructuring bets "haven't come to fruition yet."
The acknowledgment follows a sweeping reorganization in May that cut roughly 10% of Meta's global workforce and reassigned about 7,000 employees to AI-focused teams-moves that prompted employee pushback and raised morale concerns. Zuckerberg described the restructuring as not as "clean" as it could have been, with executives miscalculating the timing of the changes.
Executive optimism meets reality
When planning began in January and February, conversations with "our top people" revealed a different sentiment. Executives were "super optimistic" about tools like Claude Code from AI startup Anthropic, Zuckerberg said. He recalled that leadership was "worried that we weren't going to move fast enough to adapt." Four months later, the acceleration they expected has not materialized.
Despite the slow progress on AI agents and automation, Zuckerberg said he expects the company to begin seeing more significant benefits from its AI investments within the next three to six months. He told employees in May he did not anticipate further companywide layoffs this year, though some workers remained skeptical. A Meta spokesperson declined to comment.
Mouse-tracking software may return on opt-in basis
Chief Technology Officer Andrew Bosworth addressed the company's controversial mouse-tracking program at the same town hall. Meta paused the software last month-which tracks employee mouse movements and digital activity for AI training-while investigating a data security incident. Bosworth said the review found no employee data was included in AI training.
If the program resumes, it will operate on an opt-in basis. "For people who are comfortable, that's great, they can contribute to this kind of great human survey. To people who are not, it is not an issue," Bosworth said. When Meta first installed the program on U.S. employees' computers in April, Bosworth had told them there was no way to opt out.
Why this matters for IT and development professionals
Zuckerberg's candor about AI agent timelines is a rare public signal from a Big Tech leader that the technology is not maturing as quickly as internal roadmaps assumed. For developers and IT teams evaluating agentic tools or building automation pipelines, the gap between executive funding decisions and engineering reality is worth tracking. The fact that Meta's leadership greenlit a workforce restructuring based on optimistic January projections-only to acknowledge in July that those projections haven't held-suggests that even well-resourced organizations are struggling to forecast AI capabilities accurately. When vendors or leadership promise agent-driven efficiency gains on tight timelines, skepticism backed by concrete technical benchmarks remains a prudent posture.
Your membership also unlocks:
