Meta Breaks Up AI Lab As Part Of Superintelligence Push

In brief

Meta will restructure the Superintelligence Labs into four new AI-focused divisions.
An internal memo reveals that AI chief Alexandr Wang will lead one of the new units.
Zuckerberg says Meta is committed to leading in the race toward AI superintelligence.

Meta is breaking up its AI Superintelligence Labs into four divisions focused on research, infrastructure, and product development, part of a broader effort to accelerate progress toward so-called superintelligence.

Meta’s chief AI officer, Alexandr Wang, said in the memo that the Superintelligence Labs will be divided into smaller units focused on AI research, infrastructure, hardware, product integration, and the company’s long-term superintelligence goals.

“Superintelligence is coming, and in order to take it seriously, we need to organize around the key areas that will be critical to reach it,” Wang wrote, according to an article on Bloomberg, which first reported the story.

Meta confirmed the reorganization in an email to Decrypt, but declined to provide further details.

The restructured Meta Superintelligence Labs (MSL) will include four groups:

TBD Lab, led by Wang
FAIR (Fundamental AI Research)
Products and Applied Research, led by former GitHub CEO Nat Friedman
MSL Infra, which will oversee Meta’s AI infrastructure

The shake-up follows an aggressive hiring spree in which Meta poached top talent from firms like OpenAI, Anthropic, GitHub, and Google DeepMind. In June, Meta invested $14 billion in Scale AI, naming Wang—Scale’s CEO—as Meta’s new chief AI officer. That same month, OpenAI CEO Sam Altman accused Meta of offering $100 million in job packages to lure his staff.

According to a separate New York Times report, which cited sources familiar with the matter, some executives are expected to leave following the restructuring. Meta is also reportedly considering integrating third-party AI models into its products, marking a shift from its past reliance on in-house AI development.

CEO Mark Zuckerberg has made AI and, more recently, achieving superintelligence central to Meta’s long-term vision. In the company’s second-quarter earnings call, CFO Susan Li said capital expenditures could hit $72 billion by year’s end, driven largely by AI-related infrastructure.

In a recent post, Zuckerberg doubled down on Meta’s push toward superintelligence.

“I am extremely optimistic that superintelligence will help humanity accelerate our pace of progress,” he wrote. “But perhaps even more important is that superintelligence has the potential to begin a new era of personal empowerment where people will have greater agency to improve the world in the directions they choose.”