Nvidia picks OpenShell to tame AI gents
- On Wednesday, Nvidia CEO Jensen Huang introduced the OpenShell Reference Architecture at GTC Berlin, a security framework designed to govern AI agents within clearly defined operational boundaries.
- Corporations deploying autonomous agents face risks including data leaks and unauthorized actions, prompting Nvidia to design a framework enabling organizations to predefine permitted agent scope and monitor activity.
- The architecture utilizes Vera Rubin infrastructure for models and Vera CPUs for agent software, while BlueField-4 DPUs deploy Nvidia Sentry to monitor behavior and isolate anomalies within milliseconds.
- Major cloud providers including Amazon Web Services, Google Cloud, and Microsoft Azure are supporting the initiative, while Anthropic and SAP integrate the architecture for their agent services.
- Developed at Nvidia's Israel research center, the platform aims to integrate security directly into computing architecture, with cybersecurity firms including Cisco and Palo Alto Networks already adding support.
11 Articles
11 Articles
The Register
IT Security News - cybersecurity, infosecurity newsNvidia picks OpenShell to tame AI gents
Give your agents freedom to do useful work by locking them down
Nvidia OpenShell gets a hardware stack for Anthropic and OpenAI agents
Nvidia has released a design for running AI agents inside fixed limits, from enterprise infrastructure to edge machines. Anthropic, OpenAI, Meta, Microsoft, Google Cloud and AWS are among the companies working with it, Nvidia said. Nvidia announced the OpenShell Reference Architecture at its GTC Berlin conference on Wednesday. It is available now. Last week, Nvidia […] This story continues at The Next Web
NVIDIA Packages OpenShell Into a Full-Stack Safe-Agent Reference Architecture at GTC Berlin
At GTC Berlin on Oct. 21, 2026, NVIDIA put OpenShell, Vera Rubin, Vera CPUs and BlueField-4 Sentry into one reference architecture for building, containing and enforcing safe AI agents — and said it is available now.
Open AI models promise independence. However, with OpenShell and special processors, Nvidia wants to limit the scope of action of agents.
Nvidia GTC Berlin 2026: Nvidia Launches OpenShell Reference Architecture for Safe AI Agents
Summary: The Nvidia OpenShell Reference Architecture delivers enterprise governance, isolation and security for building, containing, observing and enforcing safe agents across enterprise infrastructure and edge systems Built on the foundation of Nvidia OpenShell secure runtime software, the architecture brings together Nvidia Vera Rubin infrastructure for model inference, Nvidia Vera CPUs for agent harnesses and tool […]
Coverage Details
Bias Distribution
- 75% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
