Meta plans cloud business to monetize $182.9B AI infrastructure bet

Meta Platforms is preparing to sell access to its vast AI compute infrastructure, turning its $182.9 billion data center bet into a direct revenue stream.

Meta Platforms is building a cloud business to sell excess AI compute capacity, a strategic pivot that could turn its $182.9 billion infrastructure investment into a direct revenue stream and pit it against Amazon Web Services, Google Cloud and Microsoft Azure.

"The new business line will be part of an initiative dubbed Meta Compute," people familiar with the plans told Bloomberg, confirming Chief Executive Officer Mark Zuckerberg's May statements that a cloud business was "definitely on the table."

Meta has committed to spending $182.9 billion on AI infrastructure in the coming years, including massive data center projects in Louisiana and Ohio — the latter described by Zuckerberg as the size of Manhattan. The company raised its full-year 2026 capital expenditure guidance to between $125 billion and $145 billion at its April earnings call, nearly double what it spent in all of 2025. Meta has also shifted 7,000 employees toward AI work and reassigned roughly 8,000 positions as part of the reorganization.

The move follows a similar strategy by SpaceX, which in May signed a deal with Anthropic to buy out all compute capacity at its Colossus 1 data center. Google, facing its own capacity crunch with a $460 billion backlog of undelivered cloud contracts, agreed in June to pay SpaceX $920 million a month for access to roughly 110,000 Nvidia GPUs. For Meta, the cloud business offers a path to monetize infrastructure that has yet to generate material standalone revenue from its Llama AI models or Meta AI assistant.

Google's Capacity Squeeze Forced Meta's Hand

The urgency behind Meta Compute became clear in March, when Google told Meta it could not supply the full amount of Gemini computing capacity Meta wanted to purchase, according to the Financial Times. Google's cloud division, which posted quarterly revenue above $20 billion for the first time — up roughly 63 percent year over year — was itself rationing capacity as its contract backlog swelled to approximately $460 billion. Meta had been using Gemini for content moderation, scam detection and coding workflows, finding it more capable than its own Llama models for certain tasks.

The restriction forced Meta to encourage employees to use AI tokens more efficiently and delayed some internal AI projects. Meta has since shifted more workloads to Muse Spark, a proprietary model developed inside its Superintelligence Labs division, partly to reduce dependence on a rival's infrastructure.

Power, Not Chips, Is the New Bottleneck

Meta's infrastructure buildout now faces a constraint that money alone cannot solve. The Electric Reliability Council of Texas, which operates the state's isolated grid, has watched its interconnection queue for large new power users swell to roughly 226 gigawatts of requests as of November 2025, up from 63 gigawatts barely a year earlier. Lead times for high-voltage transformers and switchgear — equipment representing under 10 percent of a data center's total construction cost — now stretch 18 to 48 months.

Meta has signed a string of Texas power agreements this year to secure renewable energy, including a 220-megawatt deal with Sabanci Renewables and a 298-megawatt PPA with RWE for the Rabbit's Foot Solar project. Yet the company's El Paso data center, expected to begin operations around 2028, will rely on a $551.8 million, 366-megawatt natural gas plant for its first five years — a project the City of El Paso has formally opposed.

"Silicon is the binding short-term constraint. Power is the binding long-term constraint," Stephen Sopko, a semiconductor and deep-tech analyst at HyperFrame Research, told Data Center Knowledge.

Meta Compute will be led by head of infrastructure Santosh Janardhan, Meta Superintelligence Labs leader Daniel Gross, and president Dina Powell McCormick. The unit plans to sell access to both raw compute capacity — following the CoreWeave model — and access to AI models hosted on Meta's infrastructure, including its recently launched closed-weight model Muse Spark.

The cloud business could provide a high-margin revenue stream that validates Meta's massive capital spending, which has drawn scrutiny from investors concerned about free cash flow pressure. Meta shares rose more than 6 percent in premarket trading after Bloomberg reported the plans, while Amazon shares turned negative. Amazon Web Services, Google Cloud and Microsoft Azure together control roughly two-thirds of the global cloud market, according to Synergy Research Group data. Meta's entry threatens to add capacity to a market already straining under supply constraints, potentially pressuring pricing for compute services even as demand surges.

This article is for informational purposes only and does not constitute investment advice.