Meta reorganizes its AI division once more, raising questions about its future direction.
Meta, the tech giant behind Facebook, has announced a significant restructuring of its AI division, carving it into four separate teams under Meta Superintelligence Labs (MSL). This is the fourth overhaul of Meta's AI operations in less than six months.
The restructuring led to a halt on all hiring, according to a report by the Wall Street Journal. Meta imposed a hiring freeze on MSL, halting both external hires and internal transfers unless personally approved by Alexandr Wang, the new Chief AI Officer.
Wang joined Meta after a $14 billion deal to invest in his company Scale AI. He is now heading one of the teams, which is responsible for weaving Meta's AI research into consumer products.
The new division is focused on achieving personal superintelligence that surpasses human intelligence in every way. TBD Lab, another team, is exploring an "omni model," potentially a multimodal system that would handle text, visual, audio, and other data types. Their ultimate goal is to train and scale Meta's largest models, with the aim of achieving superintelligence.
Tensions have started mounting between the new hires and Meta's veteran AI researchers, some of whom have threatened to quit. This shift has reportedly intensified tensions between these newcomers and the old guard.
FAIR, Meta's long-standing AI research arm, will work together with TBD Lab on research, with FAIR serving as an "innovation engine" for MSL. Products and Applied Research is in charge of weaving Meta's Llama models and other AI research into consumer products. MSL Infrastructure is responsible for the infrastructure needed to power Meta's AI research and development.
The hiring freeze comes after a months-long hiring spree led by Meta's CEO Mark Zuckerberg, during which Meta poached over 50 AI researchers and engineers. Meta offered to buy a stake in Nat Friedman's venture firm to woo him and co-founder Daniel Gross.
Meta's capital expenditures could be as much as $72 billion in 2025, with the bulk of it going toward building new data centers and hiring researchers. However, the company is considering making its new AI model "closed," a departure from its long-standing practice of open-sourcing its model weights.
The new division is not without controversy. The lukewarm reception of Meta's new Llama 4 language models reportedly prompted CEO Mark Zuckerberg to "handpick" MSL's team himself. Cat Valverde, founder of Enterprise AI Solutions, suggests that Meta wants to stabilize reporting lines and prove the new org design works before layering on fresh talent.
The restructuring of Meta's AI division marks a significant step in the tech giant's quest for AI dominance. As the competition heats up, it remains to be seen how these internal tensions will play out and what impact they will have on Meta's AI development.
Read also:
- List of 2025's Billionaire Video Game Moguls Ranked by Fortune
- Transformation of Decarbonization Objectives in the Iron Ore Pellets Sector
- Condolences offered by Cuba for earthquake tragedy in Turkey
- Affordable, Multifunctional Storage Solution for Small-Scale Power Plants: Marstek Jupiter C Plus, Offering Energy Storage below 220 € per Kilowatt-hour, Now Available with a 100 € Discount for Each Set.
