NVIDIA unveils architecture for secure enterprise AI agents – Engineering.com

NVIDIA unveils architecture for secure enterprise AI agents – Engineering.com

By Srabanti Chakraborty
Publication Date: 2026-10-26 09:24:00

The design combines Vera Rubin infrastructure, Vera CPUs and BlueField DPUs to monitor agent behavior and enforce security policies.

NVIDIA today announced the NVIDIA OpenShell Reference Architecture, which enables building, containing, observing and enforcing safe agents on enterprise infrastructure and edge systems.

The OpenShell Reference Architecture builds on the foundation of NVIDIA OpenShell secure runtime software, extending it into full-stack, industrial-scale infrastructure for agent development and operations.

It includes NVIDIA Vera Rubin infrastructure for running enterprise workloads, including closed and open models built with NVIDIA Nemotron; NVIDIA Vera CPUs for operating agent harnesses; open source NVIDIA OpenShell software; and NVIDIA BlueField DPUs for running the NVIDIA Sentry unified security software framework that provides continuous out-of-band monitoring of agent behavior and in-silicon policy enforcement.

The OpenShell Reference Architecture is extensible to agents running at the edge across industries such as retail, healthcare, manufacturing, robotics, software development and telecommunications with NVIDIA RTX Spark, NVIDIA DGX Spark and NVIDIA DGX Station systems.

Full-stack infrastructure built for agents

The OpenShell Reference Architecture features NVIDIA accelerated computing paired with software for secure agent operations.