Zuckerberg: AI Agents Are Advancing Slower Than Expected

Zuckerberg made the remarks at an internal town hall, according to a recording heard by Reuters. He acknowledged that a company reorganization that included major job cuts was not as “clean” as it could have been and that executives had miscalculated on the timing. Meta laid off about 10% of its global workforce in May and moved roughly 7,000 employees to AI-focused teams, Reuters reported. The restructuring was designed to fund heavy AI infrastructure investments and position the company to capture efficiency gains from AI-assisted work.

“The trajectory of the agentic development over at least the last four months hasn’t really accelerated in the way that we expected,” Zuckerberg said, per the Reuters report, adding that the company’s bets on the new structure “haven’t come to fruition yet.” He said executives had been “super optimistic” about tools like Anthropic’s Claude Code when planning began in January and February. He still expects Meta to see more significant benefits from its AI investments within the next three to six months.

At the same town hall, Meta CTO Andrew Bosworth addressed a review of a data security incident tied to the company’s mouse-tracking software, which monitors employee activity for AI training. The review found no employee data was included in AI training, Reuters reported. Meta paused the program last month and may restore it on an opt-in basis, a reversal from April, when employees were told they could not opt out.

Zuckerberg’s caution contrasts with momentum elsewhere. PYMNTS reported that Visa, Mastercard and American Express are building agentic commerce into their core networks, that Goldman Sachs projects AI agents will drive a 24-fold increase in token consumption by 2030, and that Adyen’s agentic commerce lead rates the market at just 0.5 on a five-point scale, with the hard work sitting in payments plumbing rather than the AI itself.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.