TrendAI, a company operating in the field of AI security for organizations, announced an expansion of its collaboration with NVIDIA to support NVIDIA OpenShell, a new open-source runtime environment for Agentic AI introduced at the NVIDIA GTC conference.
The solution enables organizations to deploy autonomous AI agents with a built-in governance mechanism, continuous risk visibility, and policy enforcement at runtime, thereby addressing key barriers to adopting Agentic AI in production environments.
Rachel Jin, Chief Platform and Business Officer and Head of TrendAI, stated: "Agentic AI changes the security equation. When AI systems are able to plan, act, and interact with other tools independently, the risk profile looks very different from traditional AI. Our collaboration with NVIDIA allows us to bring security directly into the architecture, so organizations can adopt Agentic AI with the level of visibility and control they expect."
Traditional AI security models were built for short interactions between users and models. Agentic AI changes this dynamic by operating continuously and performing actions across different environments.
NVIDIA OpenShell is an open-source runtime environment for long-term agents that evolve autonomously, with planning, memory, and tool execution capabilities. While these capabilities unlock significant potential for improving efficiency, they also introduce risks. These include unauthorized capabilities, hidden behaviors, prompt injection attacks, and unintended access to systems.
Pat Lee, Vice President of Enterprise Strategic Partnerships at NVIDIA, added: "Agentic AI opens the door to a new generation of applications that can plan, infer, and act. In collaboration with TrendAI, we help developers add visibility and controls to make the operation of autonomous AI agents safer."
TrendAI can transform Agentic AI from a high-risk experiment into an enterprise-ready architecture. Organizations gain the ability to define trust boundaries, enforce runtime policies, and maintain ongoing visibility into autonomous AI behavior, all while preserving the flexibility and power that make Agentic systems valuable. It adds an organizational security layer that manages how agents operate, which tools they can access, and how risks are identified and enforced, before, during, and after deployment.