Meta Platforms is reorganising its artificial intelligence efforts under a “superintelligence” group that includes a goal of developing smarter-than-human AI alongside the development of practical projects, chief executive Mark Zuckerberg said in a memo to staff on Monday.
The new unit, called Meta Superintelligence Labs, is the latest large-scale bet for Meta, which previously formed Reality Labs in 2020 to promote “metaverse” applications based on hardware such as Oculus virtual-reality headsets and more recently Ray-Ban-branded eyeglasses, while burning through more than $60 billion (£43.7bn) in the past five years.
The new AI unit has been staffed with a similar level of expenditure, with OpenAI chief executive Sam Altman saying last month that Meta had offered hiring bonuses of up to $100m to poach staff from his company.
Reorganisation
The unit is to be led by Alexandr Wang, former chief executive of AI data tagging start-up Scale AI, in which Meta bought a 49 percent stake last month for $14.3bn.
Former GitHub chief executive Nat Friedman is to co-lead the unit with Wang and will lead Meta’s work on AI products and applied research, Zuckerberg’s memo said.
The memo also introduced 11 new hires for the AI unit who include significant researchers poached from OpenAI, Anthropic and Google.
The appointments include former Google DeepMind researchers Jack Rae and Pei Sun; former OpenAI researchers Yu Jiahui, Bi Shuchao, Zhao Shengjia and Ren Hongyu; and Anthropic’s Joel Pobar, who previously spent more than a decade at Meta.
Earlier reports said OpenAI’s Zhao, Ren, Yu and Bi were hired on Friday before Meta announced their appointment on Monday.
All four studied at prestigious Chinese universities before continuing their studies in the US.
Key talent
Competition for talent has been intense, with Microsoft last year spending $650m to poach most of Inflection AI’s staff including co-founder Mustafa Suleyman and Amazon hiring key talent from Adept to form a lab focused on AI agents.
Tech companies including Meta are also spending huge amounts on data centre infrastructure to power compute- and energy-intensive generative AI services, with the largest companies saying they expect to spend a combined $320bn on AI this year.
The new unit follows the departure of senior staff at Meta and a poor reception from the AI community for its Llama 4 model, which it released with fanfare in April.
