Meta weighs AI cloud infrastructure push

Meta Platforms is developing plans to launch a cloud infrastructure business that would sell access to artificial intelligence computing power and models to outside customers. The move would create a new line of business alongside Meta’s existing social media and advertising operations, centered on monetizing its large investments in AI infrastructure. The plans are still being developed and have not been finalized.

The initiative is emerging as Meta continues to build out data centers and specialized hardware to support its own AI ambitions. By selling access to capacity that is not fully used internally, the company is exploring ways to generate additional revenue from this infrastructure. A spokesperson for Meta has declined to comment on the details of the project.

One option under consideration is to provide developers with access to AI models that run on Meta’s own infrastructure. These would include Meta’s Muse Spark models, which would be hosted and managed by the company. Developers would connect to these models and pay for usage, similar to other model-hosting services in the market.

A second option being weighed is the sale of raw computing capacity to external customers. Under this approach, Meta would rent out GPU and other AI-focused resources in its data centers, allowing customers to run their own workloads. This neocloud-style model would focus less on specific models and more on flexible access to compute.

The company has not disclosed which approach it will prioritize, and it may ultimately pursue a mix of hosted models and infrastructure rentals. Pricing structures, service levels, and launch timing have not been made public.

The cloud initiative sits within a broader push by Meta to expand and commercialize its AI capabilities. The company has been securing data centers and infrastructure at significant cost to support its internal AI projects. In this context, a cloud business would allow Meta to offset part of these capital expenditures by selling excess capacity.

In May, CEO Mark Zuckerberg told investors that selling excess compute or providing an API-based AI service was “definitely on the table.” He noted that external companies frequently inquire about purchasing access to Meta’s infrastructure. These comments signaled openness at the top leadership level to turning Meta’s AI backbone into a revenue-generating product.

Reports detailing Meta’s work on an AI cloud infrastructure business prompted a swift response in financial markets. Meta shares rose sharply in early trading on the day the news was reported, reflecting investor interest in a potential new revenue stream tied to AI infrastructure. The move suggested that the market views monetization of excess compute as a meaningful strategic development.

While investors reacted positively, the plans remain in development and could still evolve. No formal product announcement, launch schedule, or detailed roadmap has been released. Until Meta provides further disclosures, the scope and scale of the prospective cloud business will remain an area of active focus for market participants and potential customers.