VAST Data has launched a fully CUDA-accelerated artificial intelligence (AI) data stack developed in collaboration with NVIDIA Corp., one of several deals announced Wednesday.
The move transforms the VAST AI Operating System into a unified platform capable of managing the entire AI lifecycle, from data ingestion and analytics to inference and autonomous agent runtimes, within a single software environment.
Linchpin to the news is the introduction of VAST CNode-X, a new generation of NVIDIA-Certified Systems. In a departure from traditional architectural silos, the VAST AI OS now runs directly on NVIDIA-powered servers. This shift allows the platform to orchestrate complex AI pipelines, including vector search and Retrieval-Augmented Generation (RAG), as first-class citizens within the infrastructure.
By embedding NVIDIA software libraries directly into the VAST DataEngine and DataBase, the system achieves lower latency and higher efficiency for real-time SQL analytics and AI inferencing.
“Ten years ago, we set out to build a system that could continuously refine data into intelligence,” said Renen Hallak, CEO of VAST Data. “By accelerating both compute and data paths with NVIDIA, we are providing a faster, simpler way to move AI from pilot projects into durable, production-grade systems.”
NVIDIA CEO Jensen Huang said the partnership aims to provide AI agents with a “persistent memory.” The CUDA-accelerated layers allow agents to tackle complex problems over extended periods without losing context, which Huang described as the “next frontier of AI.”
Recognizing security challenges inherent in production-scale AI, VAST announced a strategic partnership with CrowdStrike Inc. The integration connects VAST’s native data-layer governance with the CrowdStrike Falcon platform.
The arrangement is designed to protect the AI lifecycle by offering continuous threat detection and automated response, securing the environments where models are trained and operated. As enterprises shift toward autonomous systems, this layered defense ensures that the data and operational states remain uncompromised.
“This partnership with CrowdStrike extends continuous threat detection and coordinated response into AI workflows and runtime environments, helping customers operate enterprise AI with greater confidence as they scale,” Hallak said.
To streamline the transition for enterprises, Supermicro Inc. is partnering with VAST to launch CNode-X Solution. The integrated stack combines Supermicro’s GPU and storage servers with the VAST AI OS and NVIDIA microservices, offering a turnkey infrastructure for rapid deployment.
Together, these developments signal a shift away from fragmented AI experiments toward a cohesive, secure, and highly accelerated foundation for the next generation of enterprise intelligence, according to VAST.
“At the end of the day, we are compute agnostic,” said John Mao, vice president of global business development at VAST. “Every C-level executive says they have an AI initiative… that will eventually make its way to the enterprise.”
While autonomous agents are still a work in progress, Mao acknowledged, nine-figure AI projects have significantly picked up in the last year.
