Meta is making another deep cut to its workforce, this time inside its most ambitious department. The company has laid off about 600 employees from its artificial intelligence division as it restructures to streamline operations and solidify Alexandr Wang’s control over the company’s AI strategy.
The layoffs, first reported by CNBC, affect workers across Meta’s AI infrastructure group, its Fundamental Artificial Intelligence Research unit (FAIR), and other product-focused teams. But one division was untouched: TBD Labs, which houses many of the high-profile AI researchers recruited by Meta this summer.
Those spared by the cuts report directly to Wang, who joined Meta in June after the company invested 14.3 billion dollars into his former company, Scale AI. The decision signals a clear bet by CEO Mark Zuckerberg on Wang’s vision and his costly new hires over long-serving staff.
Inside Meta, the AI division had been viewed as bloated. Legacy teams often competed for computing resources with newer research groups. When Wang’s team arrived to form Superintelligence Labs, it inherited what insiders described as an oversized and fragmented department. The layoffs, according to people familiar with the decision, are an effort to consolidate leadership, reduce redundancy, and align Meta’s AI work under one chain of command.
Following the cuts, Superintelligence Labs now has just under 3,000 employees. Workers impacted by the restructuring were told that November 21 will be their final day. Meta is offering 16 weeks of severance pay plus two additional weeks for every year of service.
Zuckerberg’s frustration with Meta’s AI performance has been growing for months. The release of the company’s Llama 4 models in April received a muted response from developers. Meta’s massive infrastructure spending has yet to deliver the breakthroughs its rivals, OpenAI and Google, have achieved.
Since Wang’s arrival, Meta has rebranded its flagship AI group as Superintelligence Labs, co-led by Wang and former GitHub CEO Nat Friedman. The pair have been tasked with unifying research and product development across the company’s platforms.
Meta’s financial outlook underscores the scale of that ambition. The company expects total expenses for 2025 to reach between 114 and 118 billion dollars, with higher costs projected in 2026 as its AI investments accelerate. The cuts in headcount are unlikely to slow spending but reflect a shift in who holds influence inside the company.
Earlier this week, Meta announced a 27 billion dollar partnership with Blue Owl Capital to fund its Hyperion data center in Louisiana, a project expected to occupy land comparable to much of Manhattan. The infrastructure will underpin Meta’s next phase of AI expansion.
Taken together, the layoffs and new construction signal a turning point for Meta. The company is trading its older research structure for a centralized, faster-moving operation built around Wang’s leadership. Efficiency is the public goal, but consolidation of control may be the true outcome.