Technology

Mark Zuckerberg says that development of AI agent technology is progressing more slowly than was previously expected.

Published On Fri, 03 Jul 2026
Ananya Bhattacharjee
7 Views
news-image
Share
thumbnail

Mark Zuckerberg, speaking at an internal town hall, reportedly acknowledged that Meta’s recent large-scale restructuring has not unfolded as smoothly as planned, according to a recording reviewed by Reuters. He noted that the company’s development of AI agent systems has been slower than anticipated, and that several of the expected outcomes from the reorganization have yet to be realized.

Zuckerberg said the restructuring—carried out earlier this year to redirect resources toward artificial intelligence—was not executed as cleanly as it should have been. The changes included major workforce reductions of around 10% globally and the reassignment of approximately 7,000 employees into AI-focused teams in May. While the move was intended to accelerate AI progress and improve long-term efficiency, it also led to internal pushback and concerns about employee morale.

He further explained that, in hindsight, the timing and execution of the transition were miscalculated. Conversations among senior leaders earlier in the year reflected concerns that Meta needed to move faster in adapting to AI-driven competition. At the same time, executives were highly optimistic about emerging tools such as Anthropic’s Claude Code, which influenced expectations for rapid progress.

Zuckerberg said the trajectory of “agentic” AI development over the past several months has not accelerated as expected, and the benefits of the new organizational structure have not yet materialized. Despite this, he projected that Meta could begin seeing more tangible gains from its AI investments within the next three to six months. The company is also investing heavily in AI infrastructure, with spending estimated as high as $145 billion, part of a broader tech industry investment exceeding $700 billion.

Meta’s CTO Andrew Bosworth said an internal review of its mouse-tracking software found no evidence that employee data was used for AI training. The tool, previously paused after concerns over privacy and data exposure, may be reinstated on an opt-in basis, after initially being mandatory for employees.

Disclaimer: This image is taken from Business Standard.