Meta commits $21 billion to CoreWeave, reshaping the AI infrastructure race

The AI infrastructure market has just set a new milestone. CoreWeave and Meta announced an expanded cloud computing agreement worth $21 billion, extending their partnership through December 2032.

The deal builds on a previous $14.2 billion contract signed in September 2025, bringing Meta’s total commitment to CoreWeave to more than $35 billion. It represents one of the largest infrastructure agreements in the AI era.

CoreWeave has rapidly emerged as a leading GPU-focused cloud provider. Its data centers are purpose-built for AI training and inference workloads, using high-density systems powered by NVIDIA hardware. The new agreement also includes future Rubin-based systems deployed in infrastructure dedicated to Meta.

The move highlights a broader trend: Meta is investing heavily in external AI compute capacity, suggesting demand for infrastructure is growing faster than companies can build internally. CoreWeave is capitalizing on this opportunity with multi-year contracts that provide unprecedented revenue visibility for a relatively young company.

CoreWeave’s portfolio already includes major commitments, such as $22.4 billion with OpenAI and $6.3 billion with NVIDIA, reinforcing its role as a key backbone of the emerging AI economy.

More broadly, specialized AI cloud providers are beginning to capture share from traditional hyperscalers such as Amazon Web Services, Microsoft Azure, and Google Cloud for the most compute-intensive workloads. A new power structure is emerging in global computing, one where dedicated GPU capacity becomes the most strategic asset.

The Meta–CoreWeave agreement is more than a contract. It’s a signal that the AI race is increasingly being fought at the infrastructure level.