Gate News Report, March 17 — NVIDIA announced the release of the Agent Toolkit open platform at GTC, featuring the open-source secure runtime OpenShell. This provides policy-based security, networking, and privacy safeguards for autonomous AI Agents. During the launch, Jensen Huang stated, “Claude Code and OpenClaw have sparked a turning point for Agents, extending AI from generation and reasoning to action. Employees will be enhanced by teams composed of cutting-edge, specialized, and customized Agents. The enterprise software industry will evolve into a specialized Agent platform, and the IT industry is on the brink of the next major expansion.”
The Agent Toolkit also includes the open-source AI-Q Blueprint, co-developed with LangChain. It employs a hybrid architecture with cutting-edge models for orchestration and Nemotron open models for research, reducing query costs by over 50%. Agents developed using NVIDIA’s AI-Q Blueprint are currently ranked first on both DeepResearch Bench and DeepResearch Bench II leaderboards.
In terms of security, NVIDIA is collaborating with Cisco, CrowdStrike, Google, Microsoft Security, and TrendAI to make OpenShell compatible with their cybersecurity and AI security tools. CrowdStrike also released the “Secure-by-Design AI Blueprint,” integrating Falcon platform’s protection capabilities directly into NVIDIA’s AI Agent architecture.
17 software platform vendors have integrated the Agent Toolkit: Adobe, Amdocs, Atlassian, Box, Cadence, Cisco, Cohesity, CrowdStrike, Dassault Systèmes, IQVIA, Palantir, Red Hat, SAP, Salesforce, Siemens, ServiceNow, and Synopsys. Among them, Salesforce will run Agentforce Agents via Slack as the main interface and orchestration layer, while Siemens has launched the Fuse EDA AI Agent based on Nemotron for full automation of chip and PCB design processes.