Meta spent billions on them. That's what they'll be doing.

As reported by Bloomberg, the new structure is intended to "accelerate" the company's pursuit of so-called superintelligence, according to an internal memo sent Tuesday by Alexander Wang, former CEO of Scale AI who recently joined Meta as chief AI officer.

A special group in Meta will deal with this.

“Superintelligence is coming, and to take it seriously, we need to organize ourselves around the key areas that will be crucial to achieving it – research, products, and infrastructure,” Wang wrote.

The group, known as Meta Superintelligence Labs, or MSL, will now consist of four parts:

• TBD Lab , led by Wang, who will oversee Meta’s extensive language models, including the Llama tools that underpin its AI assistant;
• FAIR , the company's in-house AI research lab, has been operating for over a decade. The team, named after its foundational AI research, focuses on long-term projects;
• Products and Applied Research, a team led by former GitHub CEO Nat Friedman, who will take these models and research and implement them in consumer products;
• MSL Infra , which will focus on the high-cost infrastructure needed to support Meta's AI ambitions.

Tuesday's reorganization did not include any layoffs, Bloomberg sources said.

Meta invests heavily in the development of artificial intelligence

Meta hopes to stabilize its AI operations after months spent poaching dozens of top AI researchers from high-salary competitors , many of whom earn hundreds of millions of dollars.

CEO Mark Zuckerberg has said the company's goal is to achieve superintelligence, or AI technology that can perform tasks better than humans, and expects to spend hundreds of billions of dollars on the talent and infrastructure needed to achieve this goal.

In July, Meta announced it was building several massive data centers to support its artificial intelligence projects. "The first one is called Prometheus and will launch in 2026," Meta's CEO said at the time.

Zuckerberg described the emerging data centers as multi-gigawatt clusters, which would place them among the largest in the world. While most data centers today are only hundreds of megawatts in capacity, several artificial intelligence and large technology companies, including OpenAI and Oracle, are planning to develop centers capable of handling several gigawatts.