Meta Chops AI Workforce with Focus on Major Talent

Meta has laid off approximately 600 employees across its artificial intelligence division in an internal restructuring, the most significant AI-focused cut in Silicon Valley to date. The layoffs, affecting its legacy FAIR research team and AI product/infrastructure units, were confirmed by multiple outlets and detailed in a memo from Chief AI Officer Alexandr Wang.

However, the company’s elite TBD Lab was spared and continues to hire selectively. Recently, Meta added Andrew Tulloch, co-founder of Thinking Machines Lab and a veteran of OpenAI and FAIR, to the lab.

While the broader AI division contracts, Meta appears to be consolidating talent into high-impact teams that align with CEO Mark Zuckerberg’s long-term vision for artificial general intelligence (AGI).

Why It Matters: Meta’s layoffs depart from big-tech’s traditional AI growth model, prioritizing small, elite teams over broad research structures. Simultaneously, Tulloch’s high-profile recruitment illustrates that Meta is still competing aggressively for top-tier AI talent.

• Meta Cuts 600 AI Roles and Restructures Core Divisions: The layoffs, confirmed by Meta, affect the company’s FAIR research group, product teams, and infrastructure divisions. The move is a departure from Meta’s past strategy of amassing large AI teams and follows years of heavy investment in AI-related infrastructure. Employees were notified before 7 a.m. on October 22 and placed on a “non-working notice period” through November 21, during which they can apply for other roles within the company or accept severance packages reportedly starting at 16 weeks’ pay.

• Memo Emphasizes Speed and Efficiency in Smaller Teams: In his internal memo, Alexandr Wang cited the benefits of streamlined operations, noting that “fewer conversations will be required to make a decision,” and that smaller groups would be more “load-bearing.” This echoes statements from CEO Mark Zuckerberg, who previously argued that a small group of highly capable individuals can make faster, more cohesive progress on complex AI systems. The message recalibrates Meta’s AI development to favor specialization over scale.

• TBD Lab Spared: Meta successfully recruited Andrew Tulloch to join its TBD Lab, the division tasked with building foundational and experimental AI models. Tulloch, who previously worked at OpenAI and FAIR, co-founded Thinking Machines Lab and led efforts focused on next-gen model infrastructure. His hiring came after Meta failed to acquire Thinking Machines outright, reflecting the company’s aggressive pursuit of AGI talent. Although Meta denies offering Tulloch a $1.5 billion package, insiders report that the deal was still exceptionally lucrative.

• From Volume to Precision: Rather than continuing to expand its AI workforce at scale, Meta is now centralizing around elite talent capable of driving results in high-stakes AGI development. This strategy is consistent with the recent internal reorganization of Meta’s Superintelligence Labs into four verticals under the direction of Alexandr Wang. Internal transfers now reportedly require top-level approval, reflecting a highly centralized approach to managing talent and resource allocation.

• A Recruitment Flashpoint in the Meta vs. OpenAI Rivalry: Tulloch’s move further fuels the escalating talent war between Meta and OpenAI. Both companies have repeatedly poached from one another in recent years, vying for the limited pool of researchers capable of building and scaling frontier AI models. Tulloch’s credibility and connections within the machine learning community offer Meta an increase in reputational capital that is becoming a critical asset in the AGI race. His onboarding suggests that Meta’s realignment is doubling down on depth over breadth.

Go Deeper -> Meta is axing 600 roles across its AI division – The Verge

In memo, Alexandr Wang explains 600-worker layoff at Meta – SFGATE

Meta Recruits Thinking Machines Co-founder to Join Their Roster – The National CIO Review