After 8,000 layoffs, Mark Zuckerberg admits Meta AI agent development is slower than expected

Meta is investing billions of dollars into AI with a key focus on AI agents. However, Meta CEO Mark Zuckerberg reportedly accepted that AI agents are progressing slower than the company expected. His comments come just weeks after Meta laid off roughly 8,000 workers, and reassigned 7,000 employees.

In Short

• Mark Zuckerberg says agentic AI development not as fast as expected
• He accepts Meta’s 8,000 job cuts and restructuring could have been more clean
• Meta is said to invest $145 billion on AI infrastructure this year alone

Meta has been pushing aggressively when it comes to AI. The company is increasing its investment when it comes to AI infrastructure, while reducing costs elsewhere. This includes laying off roughly 8,000 workers in May, and reassigning 7,000 employees into new AI-native teams. But now, Meta CEO Mark Zuckerberg has acknowledged that one of the company’s big bets, development of AI agents, is progressing slower than expected. Zuckerberg also admitted that the company’s restructuring may not have been as clean as it had hoped to be.

As per a report from Reuters, 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.” AI agents are essentially tools that run on AI models, and can do tasks for you.

The remarks came as Meta continues a costly AI push, with the company projected to spend as much as $145 billion on AI infrastructure this year alone. Zuckerberg added that Meta was still on a “journey to superintelligence.” He claimed that the company expected to begin seeing more meaningful benefits from its AI investments within the next three to six months.

Meta had previously acquired Moltbook, the viral social media platform for AI agents, and brought its team to work on AI agents within the company.

Meta executives were worried over not being fast enough

Meta has made some big changes internally in recent months. However, Mark Zuckerberg says that these changes “haven't come to fruition yet.” According to the Meta CEO, in January and February this year, “top people” at the company were worried that it wasn’t moving fast enough when it came to restructuring. At the time, he said, executives were 'super optimistic' about tools such as Claude Code from Anthropic.

In March, the company set up a new Applied AI division and shifted around 6,500 engineers and product managers from other teams into AI work. However, as per reports, some employees were not happy with this change. In addition, the company laid off about 10 per cent of its global workforce, or roughly 8,000 employees, reassigned around 7,000 workers to AI-focused or AI-native teams, and closed 6,000 openings it had previously planned to fill.

Previously, Meta’s chief technology officer, Andrew Bosworth, also acknowledged problems with how the restructuring had been handled. As per reports, he said in an internal memo that Meta had done an 'atrocious' job of explaining its vision behind the changes.

Meta says no mandatory mouse tracking

In April, Meta started using software to track mouse movements and keystrokes of employees to train AI agents. This was met with widespread backlash and concerns over user privacy, forcing the company to pause the programme. During the town hall, Bosworth said a review of Meta's mouse-tracking software had found that no employee data was included in AI training.

The Meta executive informed staff that even if the programme was to be brought back, no employee would be forced to join it. He said, “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.”

- Ends