NVIDIA has launched its Open Agent Safety Platform, bringing together software, compute and infrastructure controls designed to govern autonomous AI agents as more than 100 technology companies back or integrate with the initiative.
The platform combines NVIDIA OpenShell, Vera and Sentry to secure how agents execute tasks, access systems and interact with data across software and hardware environments.
For channel partners, the launch also creates a new infrastructure and services opportunity as vendors including Cisco, Dell Technologies, HPE, Lenovo, Microsoft and Oracle Cloud Infrastructure support the platform’s technologies.
NVIDIA builds agent safety across the technology stack
The platform enables full-stack governance and control across software that runs agents, hardware and compute layers that power their work, and the robotics systems that execute tasks in the physical world.
Elements of the platform can be deployed by organizations, depending on their unique requirements.
“AI’s extraordinary potential for society will only be realized if we solve AI safety,” said Jensen Huang, founder and CEO of NVIDIA. “As we continue to discover the frontier of AI capabilities, we must accelerate discovery at the frontier of AI safety. Safety and security require full-stack engineering.”
OpenShell, Vera and Sentry provide layered controls
The platform includes NVIDIA OpenShell, which sets boundaries for agents running on CPUs. It provides a secure runtime boundary for controlling how autonomous AI agents execute tasks across open and closed models.
OpenShell delivers this protection on NVIDIA Vera, a purpose-built CPU for agentic AI. OpenShell and Vera work together to enable agents to operate securely while completing their work as quickly as possible.
Further, the platform also features NVIDIA Sentry, an out-of-band watchdog that runs on NVIDIA BlueField-4 DPUs to continuously monitor agent behavior. Sentry provides in-silicon security enforcement and is built on NVIDIA DOCA software.
DOCA provides programmable capabilities that Sentry uses to inspect agent requests and responses, provide attested telemetry, verify agent identity, and enforce granular, zero-trust access policies for data, tools, application programming interfaces, and services.
Anthropic, Salesforce and SAP integrate agent safeguards
NVIDIA has collaborated with Anthropic to bring additional layers of security and control to the agent stack.
“Companies are giving AI agents more of their most important work, and they need to direct and verify what those agents do, especially in sensitive environments,” said Paul Smith, chief commercial officer, Anthropic. “Claude Managed Agents give companies a clear view of what each agent is doing, and NVIDIA’s platform adds another layer of governance and control across hardware and software.”
Other organizations enhancing agent security with NVIDIA include:
- SpaceXAI: Utilizing NVIDIA Open Agent Safety Platform for Cursor coding agents and Grok models.
- Scale AI: Working with NVIDIA to incorporate NVIDIA Open Agent Safety Platform technologies into the agentic infrastructure layer of Scale GenAI Portfolio.
- Salesforce: Integrated OpenShell with Slack, enabling teams to manage OpenShell agent activity directly from Slack, giving teams greater visibility and human oversight as agents work.
- SAP: Embedding OpenShell with Joule Studio runtime, part of the SAP Business AI Platform, to pair business oversight with runtime security. It is also contributing engineering work to OpenShell and working with NVIDIA to advance interoperability standards through the Open Secure AI Alliance.
“NVIDIA Open Agent Safety Platform brings together industry, researchers, and public-sector organizations to share best practices, align on evaluation methods and foster international cooperation,” said Huang. “Together, we can raise the bar for global AI safety.”
More than 100 organizations back NVIDIA’s safety platform
Further, among the over 100 organizations working with NVIDIA Open Agent Safety Platform technologies are: Accenture, Armadin, Cadence, Cognition, CrowdStrike, Cisco, Dassault Systèmes, Deloitte, EY, Hugging Face, IBM, Irregular, Perplexity, Microsoft, Netskope, SAP, Scale AI, SentinelOne, ServiceNow, Siemens, Synopsys, 1Password, OpenClaw, Palantir, and Palo Alto Networks.
“Today, you can’t responsibly hand consequential work to an AI agent without knowing three things: what it’s doing, whose authority it’s acting on, and whether the limits on that authority actually hold,” said Tomer Weingarten, Co-founder and CEO, SentinelOne. “Those guarantees can’t rest on assuming a model following instructions. We’ve watched that fail too many times. The safeguards have to build into infrastructure itself.”
Open platform use-cases target financial services, energy, and infrastructure
Additionally, organizations across robotics, financial services, infrastructure software, and energy are working with NVIDIA technologies to ensure agent safety.
Figure, Geck Robotics, and Skild AI are among robotics organizations building with OpenShell to embed agent safety controls into autonomous systems that take physical world action.
Among financial services organizations, Citi and JPMorganChase are collaborating with NVIDIA on shared open source agent safety technologies.
Hitachi Energy, EPRI, NextEra Energy, Quanta Services, SPP, Schneider Electric, Siemens Energy, and Worley are working with NVIDIA Open Agent Safety Platform technologies.
Canonical, SUSE, and Red Hat are integrating the platform into widely used software operating systems; Red Hat runs OpenShell and DOCA on Red Hat AI Factory with NVIDIA.
Lastly, NVIDIA partners such as Baseten, Cisco, CoreWeave, Dell Technologies, GMI Cloud, HPE, HP Inc., Irregular, Lenovo, Microsoft, Nebius, Oracle Cloud Infrastructure, Supermicro, and Together AI are offering AI infrastructure solutions that use and support the platform’s technologies to help customers run AI agents more securely.
NVIDIA also recently announced the formation of AEMA, an alliance focused on data center power consumption. Read more about why it formed the coalition with Google and Emerald AI.