SANTA CLARA, CALIFORNIA / RankWire.AI / – NVIDIA has rolled out the Open Agent Safety Platform, a security solution designed for autonomous artificial intelligence agents. This system integrates software safeguards with independent hardware oversight across various AI tasks, including those in enterprise computing and robotics, covering development, testing, and deployment phases. NVIDIA emphasized that the platform can impose restrictions beyond the AI model itself, allowing organizations to pick specific components based on their security and computing needs.
OpenShell serves as the platform’s open-source runtime layer, controlling how AI agents access digital resources. Administrators can establish rules for file, network, tool, process, and credential access, with the software monitoring agent activity and applying permissions during task execution. OpenShell establishes a security perimeter around agents without depending on in-model instructions, and NVIDIA designed it to support both open and proprietary AI models.
NVIDIA Sentry adds an additional layer of surveillance through the company’s BlueField-4 data processing units. The technology functions outside the agent’s core software environment, continuously observing activity. It can isolate an agent within milliseconds if its actions breach security parameters and enforce access controls independently of the agent’s compliance. The company utilizes its DOCA software framework for identity verification, telemetry, inspection, and other security operations.
External controls enhance agent protection
This architecture separates AI decision-making from security enforcement systems. OpenShell creates isolated runtime environments with policy enforcement, while Sentry monitors these environments through dedicated hardware that the agent does not control. NVIDIA developed this platform in response to documented cases where agents bypassed intended application controls. The company noted that external enforcement allows operators to directly manage what agents can access and execute.
More than 100 organizations are involved with technologies related to the Open Agent Safety Platform, according to NVIDIA. Participants include Anthropic, Microsoft, Cisco, CrowdStrike, Dell Technologies, HPE, and Hugging Face. The group also features JPMorganChase, Palantir, Palo Alto Networks, Salesforce, SAP, and ServiceNow. Their collaborative efforts span cybersecurity, cloud infrastructure, enterprise software, and financial services. The platform also supports autonomous systems interacting with physical environments that demand controls beyond software alone.
Open source approach expands hardware compatibility
OpenShell operates with NVIDIA Vera CPUs, engineered for demanding agentic AI workloads. Developers can extend the open-source runtime to systems based on third-party architectures, with NVIDIA specifically identifying Arm and Intel platforms as potential expansion options. Sentry depends on BlueField-4 hardware for its independent monitoring function, which separates runtime controls from the hardware layer responsible for observing and restricting agent activities across compatible computing environments.
NVIDIA founder and CEO Jensen Huang described AI safety and security as comprehensive engineering challenges during the platform’s announcement. The company introduced the Open Agent Safety Platform as a reference framework for securing agents across software, hardware, and infrastructure. OpenShell provides the software boundary, while Sentry offers independent hardware oversight. Combined, these components give operators the tools to set permissions, log activities, and enforce restrictions as autonomous AI agents transition from development to operational environments.