NVIDIA has announced a major push to accelerate the next phase of AI-driven work by launching an open platform for building autonomous enterprise AI agents. The initiative focuses on improving agent safety, security and efficiency, while enabling a new generation of intelligent software systems.
At the center of this effort is the NVIDIA Agent Toolkit, a collection of open-source models, tools and software that help developers create AI agents capable of independently completing complex tasks. A key addition is NVIDIA OpenShell, an open-source runtime designed to enforce policy-based security, privacy and network controls, making autonomous agents safer for enterprise deployment.
According to Jensen Huang, AI is evolving beyond simple content generation into systems that can take action. He noted that businesses will increasingly rely on teams of specialized AI agents to boost productivity, signaling a major transformation in both software and knowledge work.
The Agent Toolkit combines several components, including NVIDIA Nemotron models, AI-Q agent blueprints and tools like cuOpt, allowing developers to build agents that can perceive, reason, and act. These agents can analyze enterprise data, choose appropriate information sources, and deliver context-aware insights. The AI-Q architecture also helps reduce operational costs while maintaining high accuracy.
Security remains a core focus. OpenShell is being developed in collaboration with major cybersecurity providers such as Cisco, CrowdStrike, Google and Microsoft to ensure compatibility with advanced security tools and guardrails.
The platform is also gaining traction across the enterprise software ecosystem. Companies such as Adobe, Atlassian, Salesforce, SAP and Siemens are integrating the toolkit into their platforms to enhance AI-driven workflows. Use cases range from creative and marketing automation to semiconductor design and enterprise resource planning.
Other organizations, including Box, Palantir and Red Hat, are leveraging the toolkit to build secure, scalable AI systems tailored to their business environments.
Developers can access the Agent Toolkit and OpenShell through NVIDIA’s platform and GitHub, with support for deployment across cloud providers like Amazon Web Services, Google Cloud and Microsoft Azure. The system can also run on NVIDIA-powered hardware, from RTX-based workstations to high-performance DGX systems.
Overall, NVIDIA’s open agent platform aims to redefine enterprise software by enabling autonomous AI systems that can operate securely at scale, marking a significant step toward the future of intelligent, agent-driven work.
.webp)