Meta is undertaking a major reorganization of its AI division to focus all resources on achieving superintelligence. According to an internal memo that became public on August 20, 2025, the AI division, formerly known as FAIR, will be split into four key teams. The first team will focus on developing and scaling next-generation foundational models (Llama-4 and beyond). The second will concentrate on building computational infrastructure and R&D. The third will be entirely dedicated to AI Safety & Alignment, underscoring the companys serious approach to risks. The fourth team will be responsible for product integration and implementing advanced developments into the Meta ecosystem. This personnel and organizational strategy is a direct response to similar moves by competitors like OpenAI and Google and signals a shift from broad research to a targeted race to create AGI.
Meta Restructures AI Division to Accelerate Race to Superintelligence
