Meta, like SpaceX, looks to turn excess AI compute into cash

Meta, having invested billions in artificial intelligence development and the expansion of its data centers, is reportedly shifting gears to capitalize on its resources. According to a recent Bloomberg report, the tech giant is exploring the launch of a cloud infrastructure business that would offer access to both AI computing power and models. This strategic move positions Meta to compete directly with major players in the cloud services market, including Amazon Web Services, Google Cloud, and Microsoft Azure. This pivot comes shortly after SpaceX, through its xAI initiative, announced similar intentions. Earlier this year, SpaceX secured a deal with Anthropic to fully utilize its Colossus 1 data center's computing capacity. Additionally, partnerships have been formed with Google and Reflection AI, indicating a growing trend among tech companies to monetize their excess computing resources. Meta’s strategy suggests that success in the AI sector may hinge not just on creating superior models and services, but on owning the substantial data center infrastructure that supports them. However, the sustainability of this approach depends on continued demand for computing power and the lasting value of data centers in an evolving market. Concerns have been raised regarding the rapid expansion of AI infrastructure, with some experts warning of a potential bubble fueled by quickly depreciating technology. There is also skepticism about whether AI firms can generate sufficient revenue to justify their massive investments. Despite these challenges, Meta remains committed to enhancing its AI capabilities, with plans to allocate $182.9 billion towards AI infrastructure over the coming years. This includes significant projects in Louisiana and Ohio, with the latter being described by CEO Mark Zuckerberg as comparable in size to Manhattan and scheduled to go live this year. Unlike competitors such as Google and OpenAI, Meta has not yet experienced significant demand for its AI models and services. The company has not disclosed specific revenue figures for its AI initiatives, focusing instead on internal applications of its technology. To recoup its substantial investments, Meta might adopt a model similar to that of CoreWeave, offering direct access to its computing resources. Additionally, Meta is considering following in the footsteps of AWS by providing access to various AI models, including its recently introduced closed-weight model, Muse Spark, through its cloud infrastructure. This new venture is expected to be part of an initiative named Meta Compute, spearheaded by key leaders in the company. This development aligns with Zuckerberg’s earlier assertions that establishing a cloud computing division is a viable strategy for maximizing returns on their AI investments. TechCrunch has reached out to Meta for further comments on this emerging business strategy.