Meta's AI Shake-Up Ignites Talent War in Quest for Superintelligence

Meta is reorganizing its artificial intelligence division for the fourth time in six months, aiming to accelerate its pursuit of superintelligence, according to reports from Reuters and Bloomberg. The restructured group, Meta Superintelligence Labs (MSL), has been split into four key teams: the TBD Lab, led by newly appointed chief AI officer Alexandr Wang; the long-standing Fundamental AI Research (FAIR) lab; a products and applied research team; and an infrastructure-focused division. The reorganization, detailed in an internal memo, is intended to streamline efforts across research, product development, and infrastructure, reflecting the company’s aggressive bet on achieving artificial general intelligence (AGI)—systems that can perform tasks better than humans.

The new structure represents a shift in Meta’s AI strategy. Previously, the company had three primary AI teams—FAIR, an AI products group, and the AGI foundations team. The consolidation into MSL aims to unify these efforts under a single umbrella, emphasizing long-term research and product development. The TBD Lab, for instance, will oversee Meta’s large language models, including the Llama tools used in its AI assistant. The FAIR team will continue its foundational research while operating with relative independence, according to insiders.

Meta’s AI reorganization comes amid rising internal tensions and workforce instability. Current and former employees have reported a growing sense of discontent among existing AI researchers, particularly those at the GenAI team, who worked on the recently released Llama 4 model. The new superintelligence team has been criticized for offering significantly higher compensation to newly hired researchers, creating a perception of favoritism and diminishing morale. Some employees have even considered leaving the company or threatening to resign to gain a spot in MSL, according to sources familiar with the situation.

Despite the internal friction, Meta is doubling down on its superintelligence ambitions, with CEO Mark Zuckerberg advocating for the creation of AI systems that can surpass human capabilities. The company has already begun aggressive recruitment, poaching dozens of top AI researchers from competitors with high compensation packages. Zuckerberg has pledged to invest hundreds of billions of dollars into talent and infrastructure to support these goals. Additionally, Meta recently raised its annual capital expenditures forecast by $2 billion, projecting a range of $66 billion to $72 billion for the year.

The restructuring is also expected to impact Meta’s expense growth. Rising costs associated with data center expansion and competitive salary offers could push 2026 expense growth above the pace of 2025. The company has already tapped PIMCO and Blue Owl Capital for support in building AI data centers in rural Louisiana. Meanwhile, rival firms like Elon Musk’s xAI and Microsoft are reportedly also trying to attract Meta’s AI talent, further intensifying the competition in the AI talent market.

Source: [1] Meta plans fourth restructuring of AI efforts in six months (https://www.reuters.com/business/meta-plans-fourth-restructuring-ai-efforts-six-months-information-reports-2025-08-15/) [2] Meta Restructures AI Group Again in Pursuit ... (https://finance.yahoo.com/news/meta-restructures-ai-group-again-163342766.html) [3] Meta's Superintelligence Push Sparks Tension, Threats of ... (https://www.businessinsider.com/meta-ai-talent-war-superintelligence-push-tension-desertion-2025-8)