Meta is working on plans to sell access to AI computing power and its own models, Bloomberg reports. Although comparisons are often drawn with AWS, Azure, and Google Cloud, the new Meta Compute initiative sounds more like a new option for AI companies alongside CoreWeave, SpaceX, Nebius, and others.
The move is intended to help recoup the massive investments in its own infrastructure. One option would give developers access to AI models running on Meta’s infrastructure, similar to Amazon Bedrock. Meta would then manage the data centers and chips that power the models, including its own Muse Spark models. In addition, the company is exploring the sale of pure computing power, a model used by so-called “neocloud” companies such as CoreWeave.
Part of Meta Compute
The initiative falls under Meta Compute, an internal program focused on building and managing AI infrastructure. According to Bloomberg, the project is led by infrastructure chief Santosh Janardhan, Daniel Gross of Meta Superintelligence Labs, and Meta President Dina Powell McCormick. A Meta spokesperson declined to comment; the plans are still in development.
It is noteworthy that, until now, Meta has primarily been a major consumer of AI capacity, not a seller. The company previously signed a $14.2 billion contract with CoreWeave and a $27 billion deal with the Dutch company Nebius for additional computing power.
Pressure to recoup investments
Meta previously announced a $600 billion investment program for data centers through 2028. Unlike the hyperscalers, Meta does not have a public cloud environment that could leverage those investments.
Investors have long been asking about the return on those expenditures, and a cloud business could help offset the costs. Zuckerberg indicated back in May that selling excess computing power or an API service was “definitely on the table,” as external companies regularly ask if they can purchase access to Meta’s infrastructure.
Nevertheless, another change in Meta’s cost structure is taking center stage. In May, the company began laying off 8,000 employees to make room for even more AI investments. Meta expects capital expenditures of at least $115 billion by 2026.