Meta is reportedly preparing a significant expansion into the cloud computing market by developing a commercial artificial intelligence infrastructure business that could allow developers to purchase AI computing power and access hosted AI models.
According to a Bloomberg News report citing people familiar with the matter, the initiative remains under development and its final strategy could evolve. However, the report immediately drew attention across financial markets, with investors interpreting the move as a potentially transformative shift in Meta's long-term business model.
Meta declined to comment on the report, while Reuters said it could not independently verify the details.
Meta Compute Could Become a New AI Platform
The proposed initiative, internally referred to as Meta Compute, is reportedly being developed under the leadership of Santosh Janardhan, Meta's Head of Infrastructure, together with Daniel Gross of Meta Superintelligence Labs.
According to the report, Meta is evaluating two possible business models.
The first would allow developers to access Meta-hosted AI models—including the company's Muse Spark model family—through cloud-based APIs. Under this approach, Meta would operate the data centres and AI hardware while charging customers for model usage, similar to how Amazon Web Services (AWS) offers managed AI services through Bedrock.
The second option would involve offering customers direct access to raw GPU computing capacity, positioning Meta closer to emerging AI cloud providers such as CoreWeave.
Both approaches would represent a significant departure from Meta's traditional advertising-driven revenue model.
Zuckerberg Previously Hinted at Commercial Compute Services
Although the report surprised investors, Meta Chief Executive Officer Mark Zuckerberg had previously suggested such a move was under consideration.
During the company's May earnings call, Zuckerberg disclosed that outside organizations had repeatedly approached Meta seeking access to its AI infrastructure.
He noted that companies had requested both hosted API services and access to excess computing capacity, adding that while Meta currently intends to use most of its infrastructure internally, commercializing surplus capacity could become an option if future investments exceed internal demand.
Meta has significantly increased spending on AI infrastructure, raising its projected 2026 capital expenditure to between $125 billion and $145 billion.
Current projects reportedly include:
- A 1-gigawatt AI data centre in the American Midwest.
- A 2,250-acre hyperscale campus in Louisiana known as Hyperion.
These investments are designed to support Meta's expanding artificial intelligence ecosystem.
Investors Welcome the Strategy
Financial markets responded positively to the report.
Meta's shares rose by more than 10 percent, reflecting growing investor confidence that the company may have identified new ways to monetize its enormous AI infrastructure investments.
The gains also helped reverse concerns over Meta's substantial capital expenditure, which had weighed on investor sentiment throughout the year.
In contrast, several AI-focused cloud providers experienced sharp declines.
Companies including CoreWeave and Nebius, both of which currently supply AI computing capacity to Meta under large commercial agreements, saw their shares fall following speculation that Meta could eventually compete directly with them.
According to the report:
- CoreWeave maintains a $21 billion supply agreement with Meta.
- Nebius has a contract valued at up to $27 billion.
Analysts believe Meta's potential entry into the AI cloud market could reduce its future dependence on third-party compute providers.
AI Infrastructure Is Becoming a Competitive Battleground
Industry observers view Meta's reported plans as part of a broader shift among artificial intelligence companies seeking sustainable returns on the enormous investments required to build next-generation AI infrastructure.
If Meta launches its cloud platform, it would join established hyperscale providers including:
- Amazon Web Services (AWS)
- Microsoft Azure
- Google Cloud
Unlike these companies, however, Meta would be entering a market where enterprise relationships, cloud operations, and commercial support networks have been built over many years.
Success would therefore depend not only on computing capacity but also on pricing, reliability, developer adoption, and enterprise trust.
Diversifying Beyond Advertising
The reported cloud initiative coincides with broader efforts by Meta to diversify its revenue streams.
The company has also been linked to reports of:
- testing lower pricing for consumer AI services,
- exploring premium subscription offerings,
- and expanding commercial AI products aimed at businesses and developers.
Taken together, these initiatives suggest Meta is pursuing multiple monetization strategies while continuing to invest aggressively in AI infrastructure.
Whether Meta Compute ultimately launches in its current form remains uncertain, but the market's immediate reaction highlights how significant such a move could be for the rapidly evolving AI ecosystem.
Shunyatax Global Insight
Artificial intelligence has shifted from being primarily a software race to an infrastructure race. Companies are investing hundreds of billions of dollars in data centres, advanced chips, networking, and energy resources to power increasingly sophisticated AI systems. Meta's reported cloud ambitions indicate that the next phase of AI competition may revolve not only around building the most capable models but also around becoming a leading provider of AI infrastructure. As demand for enterprise AI continues to expand, the distinction between technology platforms, cloud providers, and AI developers is becoming increasingly blurred.
Stay connected with Shunyatax Global for trusted coverage of artificial intelligence, cloud computing, technology policy, cybersecurity, digital transformation, and global innovation.