Meta is exploring a new way to monetize the massive computing power it has built for artificial intelligence. According to Bloomberg, the company is considering a cloud infrastructure service that would offer access to AI compute and models, placing it in direct competition with major providers such as Amazon Web Services, Google Cloud, and Microsoft Azure.
The idea reflects a broader shift in the AI industry: the value may increasingly lie not only in model performance, but also in who controls the underlying data center capacity. SpaceX has already moved in a similar direction through xAI, highlighting how infrastructure itself is becoming a strategic asset.
Meta has invested heavily in this area. By the end of the first quarter, the company had committed $182.9 billion to AI infrastructure over the coming years, including large-scale projects in Louisiana and Ohio. The Ohio site, described by Mark Zuckerberg as being comparable in scale to Manhattan, is expected to go live this year.
At the same time, Meta's own AI products have not yet emerged as a major standalone revenue stream. That makes a cloud-style business model especially relevant, as the company looks for new ways to turn its infrastructure into a direct commercial advantage.
Bloomberg also reported that Meta is weighing the sale of hosted access to several AI models, including its closed-weight model Muse Spark. The initiative is said to be developing under the internal name Meta Compute, with leadership drawn from the company's infrastructure and AI teams.
If the plan moves forward, Meta could help define a new phase in which AI platforms are valued as much for their computing backbone as for their software. That shift may shape the next generation of digital services and the global cloud economy.
