NVIDIA Rolls Out NemoClaw To Enable Privacy-Focused, Always-On AI Assistants

In Brief

NVIDIAhas launched the NemoClaw stack for the OpenClaw platform, providing a single-command solution to deploy secure, privacy-focused, and scalable autonomous AI agents that can operate locally or in the cloud.

NVIDIA, a leading technology company, announced the launch of the NVIDIA NemoClaw stack for the OpenClaw agent platform, enabling users to install NVIDIA Nemotron models and the newly introduced NVIDIA OpenShell runtime in a single command

The platform incorporates privacy and security controls designed to make self-evolving, autonomous AI agents, referred to as claws, more reliable, scalable, and accessible.

Jensen Huang, founder and CEO of NVIDIA, described OpenClaw as opening “the next frontier of AI to everyone” and noted that it had become the fastest-growing open source project in history. He compared OpenClaw to operating systems for personal computers, stating that it represents an operating system for personal AI and marks “the beginning of a new renaissance in software.”

OpenClaw facilitates closer interaction with AI and enables a future in which individuals can maintain their own AI agents NVIDIA, along with the broader ecosystem, is establishing both the capabilities and the safeguards necessary to allow anyone to develop secure and powerful AI assistants.

NemoClaw Enhances Security And Accessibility For Autonomous AI Agents

The NemoClaw stack leverages NVIDIA’s Agent Toolkit software to optimize OpenClaw with a single command. It installs OpenShell, which provides access to open models and an isolated sandbox environment that enforces data privacy and security for autonomous agents. This infrastructure layer underpins claws, granting them the computational and operational access needed to perform tasks while maintaining policy-based security, network, and privacy controls.

NemoClaw is compatible with any coding agent. For open agents, it can utilize local models, including NVIDIA Nemotron, running on a user’s dedicated system. Through a privacy router, agents can also access frontier models hosted in the cloud. The integration of local and cloud-based models establishes a framework for agents to acquire new skills and complete tasks while adhering to defined privacy and security protocols.

The platform is designed to support always-on agents that require dedicated computing resources for software development, tool creation, and task completion. NemoClaw for OpenClaw can operate on a variety of dedicated platforms, including NVIDIA GeForce RTX PCs and laptops, NVIDIA RTX PRO-powered workstations, and high-performance AI systems such as NVIDIA DGX Station and NVIDIA DGX Spark supercomputers. These resources provide the local computing capacity necessary for autonomous agents to operate continuously.