Meta Eyes AI Cloud Business to Monetize Massive Infrastructure

Meta is exploring a new cloud computing business as it looks to generate revenue from its expanding artificial intelligence infrastructure, signaling a strategic shift after investing heavily in AI over the past year. Reports at the beginning of July indicate the company is seeking cloud service agreements that would allow it to sell excess computing capacity built through its multibillion-dollar AI investments.

Analysts estimate Meta’s total spending will reach approximately $145 billion in 2026, reflecting one of the largest technology investment programs in the industry. The company now appears to have additional AI computing capacity beyond its internal needs and is looking to enter the cloud services market.

The move could reshape competition among AI-focused cloud providers. Meta has previously purchased computing capacity from CoreWeave, but by offering cloud services directly, it could become a competitor to the same company that once supplied its infrastructure.

CoreWeave recently signed a major long-term agreement with Meta worth $14.2 billion to provide AI computing powered by Nvidia GB300 systems. The contract extends through December 2031 and includes an option to continue into 2032. In announcing the agreement last September, a CoreWeave spokesperson said, “The agreement underscores that behind every AI breakthrough are the partnerships that make it possible.”

Investors responded positively to reports of Meta’s cloud ambitions, with the company’s stock rising about 9% after the news, reversing declines seen in recent months. According to CNBC’s Jonathan Vanian, investors have been seeking new ways for Meta to generate returns from its substantial investments in AI infrastructure. While the company’s AI spending has largely strengthened its advertising business through improved targeting and creative tools, approximately 98% of Meta’s revenue still comes from digital advertising. Expanding into cloud computing represents one of CEO Mark Zuckerberg’s most ambitious efforts to diversify the company’s business.

The announcement also weighed on shares of Nebius, another AI-focused cloud provider founded in Amsterdam by former Yandex co-founder and CEO Arkady Volozh. Industry observers increasingly refer to companies such as CoreWeave and Nebius as “neoclouds,” describing cloud providers built specifically for artificial intelligence workloads rather than traditional enterprise computing.

Unlike broader cloud platforms such as Amazon Web Services and Microsoft Azure, which support a wide range of computing applications, neocloud providers focus primarily on AI infrastructure designed for large language models and other advanced AI systems.

Meta has also expanded its AI software efforts this year through its Muse Spark general-purpose model. Although public adoption remains unclear, reports indicate the model is primarily used internally to power AI capabilities across Facebook, WhatsApp, other Meta platforms, and the company’s Ray-Ban smart glasses.

Some analysts believe Meta’s cloud strategy may have been influenced by recent large-scale AI infrastructure agreements elsewhere in the industry. In May, reports emerged that xAI would lease the entire capacity of its Colossus 1 facility in Memphis to Anthropic. The arrangement includes approximately 300 megawatts of power and more than 200,000 Nvidia chips, with the customer expected to pay roughly $1.25 billion per month.

The broader AI infrastructure market continues to evolve rapidly. Separate reports have also indicated that OpenAI is seeking a significant investment from the U.S. government, underscoring the growing competition among major AI companies as they expand computing capacity and pursue new revenue opportunities.