Meta is building a cloud business to monetize excess AI compute
Meta Platforms is drawing up plans to sell excess artificial intelligence computing power and hosted AI models to outside customers, Bloomberg reported on July 1, 2026. The proposed cloud business sits inside Meta Compute and could help offset hundreds of billions of dollars in data-center spending while giving meta stock investors a clearer path to returns on AI infrastructure bets. The report sent META shares higher in premarket trading while pressuring neocloud rivals CoreWeave and Nebius.
Key Takeaways
- Meta is weighing two models: AI model hosting similar to AWS Bedrock, or raw GPU capacity like CoreWeave.
- Plans remain in development; a Meta spokesperson declined to comment to Bloomberg.
- META shares jumped more than 6% in premarket trading; CoreWeave and Nebius fell over 6%.
- CEO Mark Zuckerberg signaled in May that monetizing excess compute is "definitely on the table."
- The effort is led by Santosh Janardhan, Daniel Gross, and President Dina Powell McCormick.
Why Is Meta Building a Cloud Business Now?
Meta has been racing to secure costly data centers, chips, and other infrastructure to support its own AI ambitions, including CEO Mark Zuckerberg's push toward artificial intelligence "superintelligence." Investors have questioned how the company will earn returns on spending that Bloomberg said runs to hundreds of billions of dollars. A cloud business that monetizes spare capacity could offset those costs, according to people familiar with the matter who spoke to Bloomberg on condition of anonymity because the plans are not public.
The idea aligns with comments Zuckerberg made during a shareholder call in May. He said selling excess computing infrastructure or an API-based AI service—where customers pay per token—is "definitely on the table," and that outside companies regularly ask whether they can buy compute from Meta at a premium to what the company paid.
What Would Meta's Cloud Service Actually Offer?
Bloomberg outlined two potential offerings under discussion. One would let developers access AI models hosted on Meta's infrastructure, similar to Amazon Web Services Bedrock. Meta would operate the data centers and chips powering models, including its Muse Spark models, and charge developers for access.
The second option would sell raw computing capacity akin to neocloud providers such as CoreWeave. Both approaches would place Meta in more direct competition with hyperscalers including AWS, Microsoft Azure, and Google Cloud—a shift worth tracking across our Future Tech & AI Wonders coverage.
How Did Markets React to the Meta Cloud Report?
Markets split on the news. Meta stock rose more than 6% in premarket trading Wednesday after the Bloomberg report, according to Seeking Alpha. Amazon shares turned negative following the disclosure, Seeking Alpha noted.
Specialized infrastructure providers took a harder hit. Investing.com reported that CoreWeave and Nebius Group shares tumbled more than 6% amid concerns about intensified competition in the GPU-accelerated AI cloud market, where CoreWeave has built a strong foothold.
Who Is Leading Meta's Compute Expansion?
The cloud effort sits inside Meta Compute, an internal initiative focused on building and managing Meta's AI infrastructure. Bloomberg reported the project is led by infrastructure chief Santosh Janardhan, Daniel Gross of Meta Superintelligence Labs, and Meta President Dina Powell McCormick.
Company officials stressed that plans remain in development and could change. Until Meta confirms a launch timeline or pricing, the report amounts to a strategic signal—not a finished product roadmap.