Meta has laid off about 600 employees from its artificial intelligence division… one of the company’s biggest shake-ups since it began its multibillion-dollar push into advanced computing last year.

The cuts, confirmed by Meta today, affect teams across its AI infrastructure, Fundamental Artificial Intelligence Research (FAIR), and product-related units. The move comes as CEO Mark Zuckerberg looks to simplify operations and speed up decision-making while ramping up what he’s calling Meta’s “superintelligence race.”

In a memo to staff, chief AI officer Alexandr Wang said the goal was to “reduce layers and operate more nimbly.” Employees were told their internal access would be removed immediately and that their official termination date is November 21. Until then, they’ll remain on paid “non-working notice.” Meta said it will offer 16 weeks of severance pay, plus two additional weeks for every year of service.

The restructuring marks a turning point in Meta’s AI strategy. Over the past two years, the company has poured billions into data centers, model development, and hiring to keep pace with OpenAI and Google. Meta’s AI spending is expected to exceed $114 billion in 2025 and climb higher in 2026 as its next wave of data centers and models come online.

Wang (who joined Meta in June after its $14.3 billion investment in Scale AI) framed the move as a “recalibration” rather than a step back. “By reducing the size of our team, fewer conversations will be required to make a decision, and each person will be more load-bearing,” he wrote, encouraging affected employees to apply for other roles within the company.

People familiar with the matter said the cuts mainly hit older divisions whose research output had slowed, while newer teams (like the TBD Lab and Meta Superintelligence Labs) continue to expand. Those groups, led by Wang and former GitHub CEO Nat Friedman, are focused on building Meta’s next generation of AI systems.

The shake-up follows months of frustration from Zuckerberg, who reportedly grew impatient after Meta’s Llama 4 model received a lukewarm reception in April. Since then, the company has been shifting from open-ended research toward more practical AI tools, such as recommendation engines, generative features, and applied “machine reasoning.”

Analysts say the changes reflect Meta’s attempt to strip out bureaucracy in a company that swelled to more than 69,000 employees during its pandemic-era hiring boom. “Meta is trying to become the AI equivalent of SpaceX… fast, experimental, and capital-intensive,” said Daniel Ives of Wedbush Securities. “To do that, it needs fewer layers and more decisive leadership.”

The restructuring comes just one day after Meta announced a $27 billion partnership with Blue Owl Capital to fund its Hyperion data center in Louisiana… an installation Zuckerberg described as “large enough to cover a significant portion of Manhattan.” The facility will anchor Meta’s long-term AI roadmap, powering future Llama and multimodal systems.

Even with the layoffs, Wang said Meta’s focus on building smarter, more capable AI models hasn’t changed. “I’m really excited about the models we’re training, our compute plans, and the products we’re building,” he wrote. “I’m confident in our path to build towards superintelligence.”

It’s too early to tell if this leaner structure will deliver the results Zuckerberg expects… but at some point, the spending spree has to show real-world returns.

At the time of publishing this article, Stocks.News holds positions in Meta as mentioned in the article.