Meta is cutting around 600 roles in AI unit (finance.yahoo.com)

Meta is cutting roughly 600 roles from its Superintelligence Labs AI organization — a reduction spanning FAIR, product AI and AI infrastructure teams while reportedly sparing the newly formed TBD Lab. The move, disclosed in an internal memo and reported by Axios (unverified by Reuters), is framed by Chief AI Officer Alexandr Wang as a way to streamline decision‑making and broaden the scope and impact of remaining roles. Meta is encouraging affected staff to apply internally and expects most to be redeployed within the company. For the AI community this signals both consolidation and a strategic pivot: after reorganizing Superintelligence Labs in June and a mixed reception to open‑source Llama 4, Meta appears to be tightening team structures while continuing heavy capital investment in scale — notably a $27 billion financing deal with Blue Owl to fund massive data‑center projects. Technically, cuts across FAIR and infra could slow exploratory research or shift work toward engineering and production priorities, while sparing TBD Lab suggests focus on prioritized projects or architectures. The staffing change, paired with external financing to offload upfront data‑center risk, points to Meta prioritizing large‑scale model training capacity and execution efficiency over broader research headcount.