Nvidia reportedly releases hardware-based controls for AI agents

Nvidia has released the Open Agent Safety Platform to constrain AI agents through controls outside the models themselves, Paul Nugent reported. The system combines OpenShell, which runs on CPUs and limits agents' permitted actions, with Sentry, which runs on network chips and monitors their behavior at the infrastructure level. Partners include Cisco, Microsoft, Oracle, CoreWeave, Dell, HPE, Lenovo, ARM and Intel. Nvidia is also working with Anthropic to integrate managed agents with OpenShell. Nvidia's enterprise AI vice president, Justin Boitano, says model-level safeguards alone cannot govern what agents access or do.