CoreWeave has agreed to provide up to $14.2 billion of AI computing power to Meta Platforms, according to a Form 8-K regulatory filing released today. The deal, which extends through December 14, 2031, includes an option for Meta to “materially expand its commitment through 2032” for additional cloud computing capacity. 

Under the agreement, CoreWeave will provide Meta with access to Nvidia’s latest GB300-class GPUs. The arrangement underscores Meta’s aggressive buildout of infrastructure to support its AI-driven initiatives, ranging from generative AI models to recommendation engines powering its social platforms. 

This pact follows a series of major customer wins for CoreWeave. Earlier this month, the company expanded its agreement with OpenAI by up to $6.5 billion, bringing the total contract value with Microsoft-backed OpenAI to approximately $22.4 billion. Microsoft had previously been CoreWeave’s largest customer, but the new Meta deal further diversifies its client base. 

The scale of the Meta contract highlights the broader arms race among hyperscalers and AI-first platforms to secure reliable, long-term access to scarce high-end GPU capacity. With Nvidia supply still constrained, partnerships like this ensure that companies such as Meta can sustain model training and inference at scale while also providing CoreWeave with multi-year contracted revenue visibility.