Author(s): Kashif Mehmood

Originally published on Towards AI.

Zuckerberg Admits AI Agents Are Behind Schedule. Meta’s Bill So Far: $145B and 8,000 Jobs

Mark Zuckerberg spent the first half of 2026 laying off about 8,000 people, reassigning roughly 7,000 more into AI teams, and raising Meta’s capital expenditure guidance to as much as $145 billion. On Thursday, 2 July, he stood in front of employees at an internal town hall. He said, in a recording heard by Reuters, that “the trajectory of the agentic development over at least the last four months hasn’t really accelerated in the way that we expected.”

After the lead, the article argues that the “not accelerated” admission pins down a specific four-month window (March through June 2026) in which Meta’s restructuring plan—8,000 layoffs plus moving thousands into AI teams—failed to deliver the agentic momentum it was used to justify. It connects Meta’s delayed agent progress to larger spending (capex guidance raised toward $125–145B), compares the situation to repeated organizational churn in Meta’s AI leadership and structures, and highlights related reporting that Meta restricted engineers’ use of external agent-coding tools over data-copying/distillation concerns. The piece also discusses how this pattern fits broader industry data showing many GenAI pilots fail to produce measurable financial returns, while offering a “steelman” counterpoint: Meta may still believe agentic progress is happening (just not on the expected acceleration curve), markets largely didn’t punish the capex plan, and code-writing agents can work elsewhere—suggesting Meta’s challenge may be adoption, integration, and reliability rather than the underlying tech.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI