VAST Data signed a $1.17 billion commercial agreement with CoreWeave to serve as the primary data foundation for its AI cloud. The deal deepens a long-standing collaboration between the two companies and positions VAST’s AI Operating System (AI OS) as a cornerstone of CoreWeave’s infrastructure for training and inference workloads.
CoreWeave, which operates one of the largest GPU-accelerated cloud platforms, will expand its use of the VAST AI OS across its global footprint. The platform enables instant access to massive datasets and delivers consistent performance across large-scale, distributed environments. Together, the companies aim to optimize data pipelines and enable seamless scaling of AI models, from research to deployment.
The partnership will also introduce advanced data services for joint customers, integrating storage, compute, and orchestration capabilities into a unified system. The companies describe this architecture as key to supporting continuous training and real-time inference for mission-critical AI applications.
• VAST Data signs $1.17B commercial agreement with CoreWeave
• VAST AI OS becomes CoreWeave’s primary data foundation for AI workloads
• Joint effort focuses on optimizing large-scale data pipelines and accelerating AI infrastructure
• Designed to enable real-time inference, continuous model training, and global scalability
• Extends a multi-year strategic collaboration between the companies
“At VAST, we are building the data foundation for the most ambitious AI initiatives in the world,” said Renen Hallak, Founder and CEO of VAST Data. “By aligning our roadmaps, we are delivering an AI platform that organizations cannot find anywhere else in the market.”
🌐 Analysis:
Founded in 2016 by Renen Hallak, Jeff Denworth, and Shachar Fienblit, VAST Data has emerged as a significant data infrastructure innovator in the AI era. The company began by redefining storage economics through its Disaggregated, Shared Everything (DASE) architecture—an approach that separates compute from storage while maintaining a single global namespace. This eliminates traditional trade-offs between performance, scalability, and simplicity that constrain legacy NAS and object systems.
The VAST AI Operating System extends this foundation beyond storage, combining data services, compute orchestration, and real-time reasoning capabilities into one software-defined platform. Its architecture leverages NVMe-over-Fabrics, stateless front-end servers, and flash-optimized data structures to deliver exabyte-scale performance with the resilience and cost profile of object storage. Central to its IP portfolio is its global namespace and similarity-based data reduction engine, which enables deduplication and caching efficiencies across petascale datasets.
Major milestones include:
• 2019 – Launch of the first VAST Universal Storage platform with DASE architecture
• 2021 – Reached unicorn status following a $3.7B valuation led by Tiger Global
• 2023 – Introduction of the VAST Data Platform and the first iteration of its AI OS
• 2024 – Integration with NVIDIA DGX SuperPODs and CoreWeave’s GPU cloud
• 2025 – The $1.17B commercial expansion with CoreWeave marks its largest deal to date
VAST now positions itself as a foundational layer for large-scale AI systems—enabling model training, data-intensive simulations, and next-generation reasoning workloads. The company competes in a fast-evolving sector alongside storage and data orchestration leaders such as Pure Storage, WEKA, and NetApp, but differentiates itself through its unified architecture and focus on AI-native infrastructure.
🌐 We’re tracking the latest developments in AI infrastructure. Follow our ongoing coverage at: https://convergedigest.com/category/ai-infrastructure/
