Meta plans major layoffs as AI spending surges: Report

Meta plans to invest $600 billion to build data centres ‌by 2028, expanding its footprint globally
The logo of Meta is seen at Porte de Versailles exhibition center in Paris, France, June 11, 2025. — Reuters
The logo of Meta is seen at Porte de Versailles exhibition center in Paris, France, June 11, 2025. — Reuters

Meta is reportedly planning sweeping layoffs that could impact 20% or more of the company, as it seeks to offset costly artificial intelligence infrastructure bets and prepare for greater efficiency brought about by AI-assisted workers.

According to Reuters, no data has been set for the cuts and the magnitude has not been finalized. Top executives have recently signaled the plans to other senior leaders at Meta and told ​them to begin planning how to pare back, two of the people said. The sources spoke anonymously because they ​were not authorised to disclose the cuts.

Meta spokesperson Andy Stone ⁠said: "This is speculative reporting about theoretical approaches."

As per Reuters, If Meta settles on the 20% figure, the layoffs will be the company's most ​significant since a restructuring in late 2022 and early 2023 that it dubbed the 'year of efficiency'. It employed nearly 79,000 people as ​of December 31, according to its latest filing.

The company laid off 11,000 staffers in November 2022, or around 13% of its workforce at the time. Around four months later, it announced it was cutting another 10,000 jobs.

Last year, Meta CEO Mark Zuckerberg has been pushing the company to ​compete more forcefully in generative AI. The company has said it plans to invest $600 billion to build data centers ‌by 2028. ⁠Earlier this week, it acquired Moltbook, a social networking platform built for AI agents. Meta is also spending at least $2 billion to buy Chinese AI startup Manus.

It is important to note that Meta's planned AI investments follow a series of setbacks with its Llama 4 models last year, including criticism that it provided misleading results on the benchmarks it used for early versions.