VAST Data Launches Open Source Foundation Stacks to Accelerate NVIDIA AI Blueprint Adoption

VAST Data introduced VAST Foundation Stacks at NVIDIA GTC 2026 to bridge the gap between AI prototyping and production. This new open source library converts NVIDIA AI Blueprints into ready-to-use implementations on the VAST AI Operating System. Organizations can now deploy complex AI pipelines without the need for extensive manual infrastructure integration or bespoke platform assembly.

Scaling Enterprise AI

The initiative targets the common bottleneck where teams struggle to stitch together fragmented data and compute layers. By offering repeatable templates, VAST enables enterprises to operationalize proven AI patterns with minimal friction. These stacks unify data services, orchestration, and execution into a single, scalable environment that runs anywhere the VAST AI OS is deployed. The initial release features two primary stacks based on high-demand use cases like Video Search and Summarization (VSS). These tools allow users to extract semantic insights from massive volumes of live or archived video files through summarization and interactive Q&A. Another key stack utilizes NVIDIA AI-Q to build custom research assistants capable of synthesizing private enterprise data safely.

Unified Performance

Performance is driven by the VAST AI OS and its underlying Disaggregated Shared Everything architecture. This software can run in the cloud or on-premises through the recently launched CNode-X hardware platforms. By embedding NVIDIA Blackwell GPUs directly into storage nodes, VAST eliminates the latency typically associated with moving massive datasets between disparate clusters.

VAST Vice President John Mao highlighted that these stacks offer a faster path to production for agentic AI systems. NVIDIA executives echoed this sentiment, emphasizing that continuous data preparation is vital for accurate intelligent agents. This partnership ensures that AI workflows always have access to the most recent and relevant information across the entire organization.

Developers can access the Foundation Stacks today through a public GitHub repository. VAST plans to expand the library with industry-specific examples and interactive sandbox environments in the coming months. This move further solidifies the collaborative ecosystem between VAST and NVIDIA in the enterprise sector.