Meta is once again reorganizing its AI division in a bid to accelerate its quest for “superintelligence.” In what marks the fourth major shakeup in just six months, the company is dissolving its recently created Meta Superintelligence Labs (MSL) and reallocating its teams into a new four-pillar structure. This is a high-stakes move in a race where rivals like Google and OpenAI are also rapidly evolving.
The internal memo outlining the changes reveals a more focused approach. The new structure includes a TBD Lab led by Chief AI Officer Alexandr Wang, which will focus on Meta’s large language models like the Llama series. In addition, the long-standing Fundamental AI Research (FAIR) lab will continue its work on long-term projects, while a new Products and Applied Research team, led by former GitHub CEO Nat Friedman, will be responsible for bringing AI innovations into consumer-facing products. Finally, the MSL Infra group will handle the crucial infrastructure needed to support these ambitious AI projects.
The constant reorganization is reportedly a response to both internal tensions and the intense pressure to compete at the highest level of AI development. Reports suggest the company’s latest Llama 4 model received a mixed reception, and this new structure is a way to address those performance issues directly.
