Meta, like SpaceX, is looking to turn excess AI computing into cash

Meta has spent billions of dollars developing artificial intelligence and building data centers to support it. But now, the company may be preparing to put those data centers to a more profitable purpose immediately.

Wednesday, Bloomberg reported Meta is developing plans for a cloud infrastructure business, selling access to both the power and computational power of AI models. The move will put it up against big cloud providers like Amazon Web Services, Google Cloud, and Microsoft Azure.

Meta’s decision to sell excess computing comes weeks after SpaceX xAI, announced similar plans. In early May, SpaceX signed a deal with Anthropic to buy all of the computing capacity at SpaceX’s Colossus 1 data center. SpaceX has since signed similar leases with Google and Reflection AI. The fact that Meta is doing the same is an indication that the winners in the AI ​​race may not be those who provide the best models and services, but rather those who own the data centers.

That is, if demand for computing continues, and if data centers maintain their value. Some skeptics have warned that the race to build AI infrastructure is creating a tending bubble Largely on chips that quickly depreciate in value. Others They wondered whether AI companies could generate enough end-user revenue to justify trillion-dollar bets.

These concerns have not stopped Meta from investing heavily in AI computation infrastructure. As of the end of the first quarter, it was dead Committed to spending $182.9 billion on AI infrastructure in the coming years, including ongoing mega projects in Louisiana and Ohio. Ohio Project, which Zuckerberg said it would be The size of ManhattanIt is expected to come online this year.

Unlike Google and OpenAI, Meta has not seen much demand for its AI models and services. Meta does not break out revenue from Meta AI or from Llama, the open-weight family of AI models, in its earnings, and executives have mostly emphasized companies’ internal uses of AI in public statements. This could mean that Meta’s AI endeavors do not yet represent a material, standalone revenue line.

To get a return on some of its massive spending, Meta may copy CoreWeave’s business model and sell access to “raw” computing capacity, according to Bloomberg. The outlet also reported that Meta is considering following in AWS’s footsteps and selling access to various AI models — including the recently launched closed weight model, Moses Spark – Hosted on its AI infrastructure.

The new business line will be part of a new initiative called Meta Compute, which is led by Infrastructure President Santosh Janardhan, Meta Superintelligence Labs President Daniel Gross, and President Dina Powell McCormick.

The report confirms Zuckerberg’s statements in May that Meta’s cloud computing business is a business “Definitely on the table.” As a way to get a return on some huge investments in its strategy to develop super artificial intelligence.

TechCrunch has reached out to Meta for comment.