New York, USA, November 6, 2025- GPU cloud provider CoreWeave has signed a USD 1.17 billion commercial agreement with VAST Data, selecting the Nvidia-backed storage company as its primary data platform for AI workloads across its expanding cloud infrastructure.
VAST confirmed the deal in an official statement, describing it as a multi-year strategic partnership aimed at providing CoreWeave with a unified data platform to handle the extreme demands of generative AI and large language model training. The agreement builds on a relationship first announced in 2023 and significantly expands the scope of VAST’s role in CoreWeave’s infrastructure.
“CoreWeave is building one of the most powerful AI clouds in the world, and VAST Data is proud to serve as its data foundation,” VAST co-founder and CEO Renen Hallak said in the release. “This partnership validates our vision that next-generation AI infrastructure must unify compute and data, so organizations can train and serve models at massive scale.”
The contract is among the largest commercial agreements in the data infrastructure sector this year. It represents one of the most significant endorsements yet of VAST’s scale-out storage software, which enables faster access to high-bandwidth data needed by GPU clusters used for AI model development.
CoreWeave operates a fast-growing AI cloud platform powered by tens of thousands of Nvidia GPUs and has been securing long-term technology partnerships to sustain its expansion. The company’s selection of VAST is seen as an effort to streamline data movement and improve training efficiency for large-scale models hosted on its platform.
The deal also underscores how data infrastructure companies are becoming central to AI economics. VAST, which achieved a reported USD 200 million annual recurring revenue earlier this year, is one of a handful of firms redefining the storage layer for accelerated computing. The company counts Nvidia as an investor and collaborator, with its systems already optimized for the GPU giant’s data pipeline architectures.
Neither company disclosed specific financial terms beyond the total value of the contract or its duration. However, industry analysts view the multi-year scope as a sign of how AI infrastructure providers are locking in long-term ecosystem partners to support a global wave of GPU cloud buildouts.
.jpg)