Zuckerberg tells Meta staff AI agent rollout is behind schedule as $145B bet hangs in balance

Meta CEO Mark Zuckerberg told employees at an internal town hall on July 2 that the company's work on AI agents has not progressed as quickly as he had hoped — a rare concession from an executive who reorganized his entire company around that bet and committed up to $145 billion in capital spending for 2026 to back it. A recording of the meeting was reported by Reuters.

"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. He added that the company's bets on its new organizational structure "haven't come to fruition yet."

A Bet Built on Agentic Commerce

The admission is significant because Meta has staked an extraordinary amount on AI agents delivering returns. Earlier this year, Zuckerberg promised investors a wave of new AI models and products across Facebook, Instagram, and WhatsApp, with agentic shopping singled out as a high-priority initiative. An AI shopping agent for Instagram, internally codenamed "Hatch," was reportedly targeting a launch before Q4 2026. So far, no such agent is live on any Meta platform.

The company has spent aggressively to reach this point. Meta laid off approximately 10% of its global workforce in May and moved roughly 7,000 employees to newly formed AI units. Its capital expenditure forecast for 2026 was raised in April to between $125 billion and $145 billion — one of the largest single-year infrastructure bets in corporate history. Zuckerberg was candid about how the planning went wrong: early assumptions in January and February were "super optimistic" about tools like Anthropic's Claude Code, and those assumptions didn't hold when the work actually began.

Not a Pivot — A Timeline Shift

Despite the slower pace, Zuckerberg told employees that Meta should start seeing more substantial benefits from its AI investments within the next three to six months, pointing to the end of 2026 as a turning point. He also acknowledged that the company reorganization "wasn't as clean as it could have been" and that he will "almost certainly make more" mistakes as the transformation continues.

The agentic shopping thesis remains central to Meta's long-term plan. The company holds an unusual advantage in this space: it has deep behavioral data on billions of users across Facebook, Instagram, and WhatsApp — their social graphs, purchase histories, content interactions, and personal interests. If reliable AI agents can be built on top of that data, the commerce opportunity is enormous. The question is timing and execution, and on both counts, Zuckerberg is now signaling the original schedule was too aggressive.

A Data Security Incident, Too

At the same town hall, Meta CTO Andrew Bosworth addressed a separate matter: a data security review tied to Meta's mouse-tracking software, which had been monitoring employee activity for AI training purposes. The review found no employee data was included in AI training. Meta paused the program after it became public and is assessing whether to restore it.

The broader implication of Zuckerberg's admission runs beyond Meta. If the company that controls Facebook, Instagram, and WhatsApp — with more behavioral data about users than almost any other organization on earth — is struggling to deploy reliable AI agents, it raises a pointed question for the rest of the industry: how much harder is it for everyone else? As reported by SiliconAngle, based on a recording first obtained by Reuters, the honest answer from the most data-rich company in consumer tech is that agentic AI is taking longer than anyone publicly predicted.