Key Takeaways

  • Meta has received offers from businesses seeking to rent its AI infrastructure, although it does not yet operate a commercial cloud service.
  • Amazon, Alphabet, Microsoft, and Meta plan roughly $725 billion in AI-focused capital spending during 2026.
  • Strong cloud growth is helping Amazon Web Services, Microsoft Azure, and Google Cloud make the financial case for sustained AI investment.

Second-quarter results from Alphabet, Microsoft, and Amazon have sharpened the clearest argument yet for Big Tech’s extraordinary AI spending: cloud computing can turn costly infrastructure into recurring, high-margin revenue.

Now Meta is considering whether it can adopt part of that model.

During Meta’s latest earnings call, CEO Mark Zuckerberg said businesses interested in renting the company’s AI infrastructure had offered a “meaningful premium over what we paid for the compute.” Meta does not currently sell public cloud services, and Zuckerberg’s comment did not amount to a product launch or formal market-entry plan. Still, it indicated that outside demand for Meta’s computing capacity is becoming difficult to ignore.

The timing matters. Amazon, Alphabet, Microsoft, and Meta are expected to spend roughly $725 billion on AI-focused capital expenditures in 2026, about 77% more than the approximately $410 billion spent in 2025, according to figures reported by Tom’s Hardware. Much of that money is flowing into data centers, processors, networking equipment, power systems, and the physical facilities needed to train and operate increasingly large AI models.

That is a staggering bill, even for companies with substantial cash generation. Some Wall Street analysts estimate that industry spending could reach $800 billion to $900 billion during 2026 and exceed $1 trillion by 2027. Alphabet, Amazon, Meta, Microsoft, and Oracle have also added about $350 billion in debt over five years to supplement operating cash flow used for AI and cloud infrastructure.

Capacity sitting idle is expensive, while capacity sold to enterprise customers can become a highly lucrative business.

Amazon Web Services, Microsoft Azure, and Google Cloud already have the sales organizations, billing systems, partner networks, and service catalogs required to convert infrastructure into revenue. Their customers rent computing resources for databases, software development, storage, analytics, and, increasingly, AI model training and inference. Long-term contracts also offer revenue visibility that consumer-facing AI products generally do not.

Recent growth reinforces the point. Google Cloud revenue increased 63% year over year, according to Reuters, while cloud businesses across Alphabet, Microsoft, and Amazon expanded faster than many of their companies’ more mature operations. Across leading cloud providers, contract backlogs now exceed $2.3 trillion, up 16% quarter over quarter. That pipeline gives investors more reason to believe today’s infrastructure spending may produce revenue for years rather than quarters.

Could Meta become a fourth major public cloud provider? Not quickly. Building data centers is only one part of the equation. Enterprise cloud customers expect service-level agreements, technical support, compliance controls, regional availability, predictable pricing, and tools for moving workloads between environments.

Meta would also need to decide what kind of provider it wants to be. A broad competitor to Amazon Web Services, Microsoft Azure, and Google Cloud would require an enormous commercial expansion. A narrower service focused on AI training, inference, or access to specialized Meta infrastructure could be more practical. Oracle Cloud’s emergence as a hyperscale contributor shows that customers may consider alternatives when specialized capacity and pricing are attractive.

There is another wrinkle. Meta primarily built its infrastructure for internal products, including advertising, recommendations, and generative AI. Renting spare capacity could improve utilization, but committing too much infrastructure to outside customers might limit Meta’s flexibility when its own computing requirements surge.

Industry architecture may lower some barriers. The Georgetown Center for Security and Emerging Technology has examined how cloud investment supports AI development, while the Cloud Native Computing Foundation ecosystem around Kubernetes gives enterprises a more portable foundation for scalable workloads. Even so, operational maturity and customer trust take time.

For Meta, cloud sales could create an additional return on infrastructure it already plans to build. For investors, the possibility offers something equally valuable: a clearer path from AI expenditure to revenue. The offers described by Zuckerberg suggest demand exists. Whether Meta turns that interest into a durable commercial platform is now the more consequential question.