Meta shares jump 10% on plans to sell AI computing power

Meta shares jump 10% on plans to sell AI computing power

The company is considering selling access to hosted AI models and raw computing capacity as it looks to generate returns from its massive infrastructure investment.

Meta shares jumped more than 10% Wednesday after reports that the company is developing a cloud infrastructure business that would sell access to artificial intelligence models and computing power.

The stock traded near $623 at press time, marking its largest intraday gain since April.

The proposed business would put Meta in more direct competition with Amazon Web Services, Microsoft Azure and Google Cloud, which already generate billions of dollars by renting computing infrastructure and software to outside customers.

One plan under consideration would allow developers to access AI models hosted on Meta infrastructure, including the company’s Muse Spark models. Meta would operate the data centers and chips powering the models and charge customers based on their usage, following an approach similar to the services available through AWS Bedrock.

Advertisement

Meta is also considering renting raw computing capacity to customers, placing it in competition with specialized AI cloud providers such as CoreWeave and Nebius.

The plans are being developed through Meta Compute, an internal initiative responsible for managing the company’s expanding AI infrastructure. The group is led by infrastructure chief Santosh Janardhan, Meta Superintelligence Labs executive Daniel Gross and Meta President Dina Powell McCormick.

A cloud business could provide Meta with a path to generate revenue from the hundreds of billions of dollars it has committed to data centers, chips and other infrastructure needed to pursue advanced AI systems.

The company has also signed major computing agreements with CoreWeave, Google and Oracle to secure additional capacity. Those deals have positioned Meta as one of the largest buyers of AI infrastructure, but the proposed cloud business could eventually turn the company into a supplier as well.

Shares of CoreWeave fell as much as 14% following the report, while Nebius dropped as much as 17%, reflecting concerns that Meta could become a powerful competitor to companies that currently provide it with computing capacity.

Mark Zuckerberg previously said selling excess infrastructure or offering access to Meta models through an API was under consideration. He said companies regularly approach Meta seeking access to its computing resources, although the company had continued using that capacity for its own AI development.

Entering the cloud market would require Meta to build more than data centers. The company would also need software platforms, enterprise sales operations and customer support services capable of competing with established providers.

Still, the business could help Meta monetize unused infrastructure, diversify its revenue beyond advertising and justify the scale of its AI spending as demand for computing capacity continues to grow.

Disclosure: This article was edited by Estefano Gomez. For more information on how we create and review content, see our Editorial Policy.

Meta shares jump 10% on plans to sell AI computing power

Meta shares jump 10% on plans to sell AI computing power

The company is considering selling access to hosted AI models and raw computing capacity as it looks to generate returns from its massive infrastructure investment.

Meta shares jumped more than 10% Wednesday after reports that the company is developing a cloud infrastructure business that would sell access to artificial intelligence models and computing power.

The stock traded near $623 at press time, marking its largest intraday gain since April.

The proposed business would put Meta in more direct competition with Amazon Web Services, Microsoft Azure and Google Cloud, which already generate billions of dollars by renting computing infrastructure and software to outside customers.

One plan under consideration would allow developers to access AI models hosted on Meta infrastructure, including the company’s Muse Spark models. Meta would operate the data centers and chips powering the models and charge customers based on their usage, following an approach similar to the services available through AWS Bedrock.

Advertisement

Meta is also considering renting raw computing capacity to customers, placing it in competition with specialized AI cloud providers such as CoreWeave and Nebius.

The plans are being developed through Meta Compute, an internal initiative responsible for managing the company’s expanding AI infrastructure. The group is led by infrastructure chief Santosh Janardhan, Meta Superintelligence Labs executive Daniel Gross and Meta President Dina Powell McCormick.

A cloud business could provide Meta with a path to generate revenue from the hundreds of billions of dollars it has committed to data centers, chips and other infrastructure needed to pursue advanced AI systems.

The company has also signed major computing agreements with CoreWeave, Google and Oracle to secure additional capacity. Those deals have positioned Meta as one of the largest buyers of AI infrastructure, but the proposed cloud business could eventually turn the company into a supplier as well.

Shares of CoreWeave fell as much as 14% following the report, while Nebius dropped as much as 17%, reflecting concerns that Meta could become a powerful competitor to companies that currently provide it with computing capacity.

Mark Zuckerberg previously said selling excess infrastructure or offering access to Meta models through an API was under consideration. He said companies regularly approach Meta seeking access to its computing resources, although the company had continued using that capacity for its own AI development.

Entering the cloud market would require Meta to build more than data centers. The company would also need software platforms, enterprise sales operations and customer support services capable of competing with established providers.

Still, the business could help Meta monetize unused infrastructure, diversify its revenue beyond advertising and justify the scale of its AI spending as demand for computing capacity continues to grow.

Disclosure: This article was edited by Estefano Gomez. For more information on how we create and review content, see our Editorial Policy.