Nvidia unveils OpenShell architecture for secure AI agent deployment

Nvidia (NASDAQ: NVDA) announced the OpenShell Reference Architecture at GTC Berlin on Oct. 21, 2026, a full-stack infrastructure framework designed to provide governance, isolation and security for AI agents deployed across enterprise and edge systems.

The architecture builds on Nvidia's OpenShell secure runtime software and integrates several hardware and software components: Nvidia Vera Rubin infrastructure for model inference, Nvidia Vera CPUs for agent harnesses and tool use, and Nvidia BlueField-4 data processing units running Nvidia Sentry for in-silicon policy enforcement. According to the company, Nvidia Vera Rubin NVL72 systems deliver 30x higher throughput per megawatt for agentic workloads, and Nvidia Vera CPUs deliver 2x the performance of x86 processors. Nvidia Sentry is said to boost agent capacity by up to 20%.

The architecture extends to edge deployments via Nvidia RTX Spark, Nvidia DGX Spark and Nvidia DGX Station systems, and runs on enterprise operating systems including Canonical, Red Hat, and Microsoft Windows with Microsoft Execution Containers. It is also compatible with cloud platforms from Amazon Web Services, Google Cloud and Microsoft Azure.

A broad group of companies is working with the OpenShell Reference Architecture. Cybersecurity firms including CrowdStrike, Cisco, Palo Alto Networks and Zscaler are integrating support into their agent security products. Enterprise software companies including SAP, ServiceNow, IBM, Siemens and Dassault Systèmes are building agent platforms on the framework. System manufacturers Dell Technologies, HPE, Lenovo and Supermicro are building compatible infrastructure.

ServiceNow announced general availability of ServiceNow Arc, a long-running AI agent governed at runtime by its Super Intelligence Control Tower and enforced by Nvidia OpenShell and Nvidia Sentry.

European banks BBVA and Santander are exploring the architecture as a framework for agent safety and governance.

"As enterprises prepare to deploy billions of AI agents, safety and control must be built into the computing architecture," said Jensen Huang, founder and CEO of Nvidia.

The Nvidia OpenShell Reference Architecture is currently available, according to the company's statement.

You May Also Be Interested In

• Needham reiterates Buy rating on Nvidia with $300 target
• Cushman & Wakefield retires 2028 notes, topping $800M in debt repayment
• N-able and SentinelOne expand cybersecurity partnership