Meta Launches “Meta Compute” to Scale AI Infrastructure and Power Superintelligence Push

Meta has unveiled a new initiative called Meta Compute, aimed at building large-scale artificial intelligence infrastructure and managing the company’s global network of data centres and supplier partnerships as it pursues what it calls superintelligence.

Chief Executive Mark Zuckerberg said the initiative will be co-led by Santosh Janardhan, Meta’s head of global infrastructure, and Daniel Gross. Janardhan will continue overseeing Meta’s technical foundations and data centre operations, while Gross will head a newly created group responsible for strategic capacity planning and business partnerships.

Both executives will work closely with Dina Powell McCormick, who recently joined Meta’s leadership team, Zuckerberg said in a post on Threads.

Meta Compute sits at the core of Meta’s aggressive push into frontier AI and so-called personal superintelligence—a theoretical stage where machines surpass human cognitive abilities. Zuckerberg said the company is investing heavily in data centres and the energy systems required to run them, noting that Meta plans to build “tens of gigawatts” of capacity this decade and potentially “hundreds of gigawatts or more” over the longer term.

Such computing ambitions would consume electricity on the scale of small cities or even countries, raising concerns about pressure on resources such as power and water. The move comes as Meta seeks to regain momentum in the competitive AI race following a lukewarm response to its Llama 4 model. The company has committed up to $72 billion in capital expenditure in 2025 alone.

Across the tech sector, rising AI workloads are driving a surge in U.S. power demand for the first time in two decades. To secure long-term energy supplies, Meta has signed 20-year agreements to purchase electricity from nuclear plants operated by Vistra and has partnered with two companies developing small modular nuclear reactors.