Meta has signed a new $21 billion agreement with CoreWeave to secure additional AI cloud capacity, expanding an earlier $14.2 billion deal as it accelerates investment in artificial intelligence. The newly announced contract will run from 2027 through 2032, extending Meta’s long-term access to high-performance computing infrastructure amid intensifying competition in AI.
CoreWeave operates data centers packed with Nvidia GPUs, which are critical for training and running advanced AI models. The company already supplies infrastructure to major players including Google, Microsoft, and OpenAI. While Meta is building its own data centers, including a planned $10 billion facility in Texas, it continues to rely on external providers to meet surging compute demand.
The investment aligns with Meta’s broader capital expenditure plans, which it recently projected at $115 billion to $135 billion this year, nearly double its 2025 spending. The company is scaling infrastructure to support its AI ambitions, including new model development through its Superintelligence Labs unit and the recently announced Muse Spark model.
For CoreWeave, the expanded partnership reduces reliance on Microsoft, previously its largest customer, and strengthens its position in the fast-growing AI infrastructure market. The deal reflects a broader industry trend, where even large technology companies are increasingly turning to specialized cloud providers to secure the computing power needed for next-generation AI systems.
