Meta Reportedly Plans AI Cloud Business to Sell Excess Computing Capacity

Meta is launching a new cloud business to sell its excess AI computing power and model access to outside customers, moving to monetize its massive infrastructure investments.

The business, internally called MetaCompute, would position the social media giant in direct competition with Amazon Web Services, Microsoft Azure, and Google Cloud. CEO Mark Zuckerberg first signaled the possibility during the company's Q3 2025 earnings call, telling investors it was "definitely on the table."

Meta has committed to spending up to $145 billion on capital expenditures this year, with investors expressing concern about returns on that spending.

The cloud offering would create a revenue stream to offset those costs while providing a safety valve if AI capacity exceeds internal needs.

According to Bloomberg, Meta is debating whether to sell access to AI models hosted on its infrastructure (similar to AWS's Bedrock) or sell raw computing power (similar to neocloud providers like CoreWeave). Shares surged nearly 9% on the news, reflecting investor relief at a potential path to monetization.

"Almost every week there are different companies that come to us from the outside asking us to both stand up an API service or asking if we have compute that they could buy from us at some premium to what we've bought it at." – Mark Zuckerberg, CEO of Meta

Meta's entry into cloud computing would introduce a powerful new competitor to the hyperscale cloud market.

Unlike traditional cloud providers that must fund data centers through service revenue, Meta has already financed its infrastructure for internal development, meaning any external sales would represent pure margin expansion.

This dynamic could put pressure on pure-play infrastructure providers like CoreWeave, whose shares fell nearly 15% following the news.

The Meta cloud computing strategy could generate tens of billions in revenue, following the precedent set by SpaceX's xAI, which has signed deals with Anthropic and Google for compute capacity.

Morgan Stanley analysts expect the neocloud option to be easier to execute than a full hyperscale product, potentially adding approximately $3 per share to Meta's 2028 earnings for every 250 MW of capacity leased.

The move comes as Meta's core advertising business remains strong, with 33% year-over-year revenue growth in the first quarter driven by AI-enhanced ad targeting.

However, the stock has fallen 26% from its August 2025 peak as investors worried about the AI spending cycle.

The cloud business offers a way to transform that spending from a cost center into a revenue opportunity, potentially reshaping how the market values Meta's AI investments.