Meta Turns Its Surplus Computing Power Into a New Revenue Stream

From Massive Spending to Generating Returns

Meta has spent billions of dollars developing artificial intelligence technologies and building massive data centers to support them. Yet the company now appears ready to transition into a more profitable phase, as it works on plans to establish a commercial arm in the cloud infrastructure space — one that will sell AI-dedicated computing capacity and offer external customers access to its large language models.

A Direct Challenge to Cloud Giants

According to Bloomberg, this move will place Meta in direct competition with major cloud service providers, including Amazon Web Services, Google Cloud, and Microsoft Azure. The initiative goes beyond selling raw computing power — it extends to offering hosted AI models running on Meta's own infrastructure, similar to what Amazon does through its AWS platform.

This new initiative — known internally as "Meta Compute" — is set to be overseen by three senior executives:

• Santosh Janardhan, the company's Head of Infrastructure.
• Daniel Gross, Head of Meta Superintelligence Labs.
• Dina Powell McCormick, President of the company.

The SpaceX Precedent and a Shifting Landscape

Meta is not alone in pursuing this direction. SpaceX's xAI led the way when it struck a deal in early May with Anthropic to sell the full computing capacity of its "Colosseum 1" data center, with further deals following with Google and Reflection AI. This shared approach between two companies of Meta and SpaceX's caliber points to a fundamental shift in how the AI race is being understood — the winner may not be whoever builds the best models, but whoever controls the infrastructure that runs them.

Massive Investment, Elusive Returns

Through the end of the first quarter of this year, Meta has committed to spending more than $182 billion on AI infrastructure in the years ahead. Among the most notable projects is a data center in Ohio that Mark Zuckerberg described as being the size of Manhattan island, with operations expected to begin later this year.

Yet this enormous spending has not yet been matched by tangible returns from direct AI products. Meta does not break out standalone revenue figures from Meta AI or its open-weight Llama models in its financial reports, and executive statements have largely focused on internal AI use cases. This means AI projects have not yet formed a truly independent revenue stream in any meaningful sense.

Legitimate Concerns Amid the Optimism

The picture is not without serious concerns. Some analysts warn that the frenzied race to build AI infrastructure could give rise to an economic bubble, particularly given the heavy reliance on chips that depreciate rapidly in value. Others question whether AI companies can generate sufficient revenue from end users to justify investments measured in the trillions.

For his part, Zuckerberg stated in May that entering the cloud computing business is "seriously on the table" as a means of recouping a portion of the vast investments the company has poured into its pursuit of what it calls "superintelligence."

Conclusion

Meta's move toward monetizing its surplus computing capacity signals a new chapter in the evolution of the AI industry — one in which control over infrastructure is becoming no less important than innovation in models and algorithms. The question that remains open: will the value of these massive facilities hold up against market volatility and the relentless pace of technological change?

✦ بقلم فريق دروب أيديا