NVIDIA is working on an open-source platform for AI agents called NemoClaw, with an enterprise focus. The platform allows access even for systems not using NVIDIA chips. It comes amid concerns over the security and unpredictability of such autonomous tools.
NVIDIA is reportedly developing its own open-source AI agent platform, according to a Wired report. The platform, named NemoClaw, is being pitched to enterprise software companies and emphasizes accessibility for users whose agents do not run on NVIDIA hardware.
The company has an enterprise-oriented approach for the initial rollout. NVIDIA has contacted firms including Salesforce, Cisco, and Google to form partnerships, though none have issued official statements. This development precedes NVIDIA's annual developer conference next week.
AI agent platforms like NemoClaw enable autonomous operation on computers to handle complex tasks with minimal human oversight, differing from traditional chatbots. However, their use in enterprise settings remains controversial due to potential unpredictability and security issues. Some tech companies have instructed employees to avoid tools like OpenClaw on work devices. For instance, a Meta employee recounted an AI agent that went rogue and deleted numerous emails, highlighting risks when such agents access enterprise networks.
To address these concerns, NVIDIA is incorporating extra security features into NemoClaw to appeal to business users. The trend traces back to Clawdbot, now known as OpenClaw, created by Peter Steinberger. Steinberger recently joined OpenAI, where CEO Sam Altman praised him: “Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting with each other to do very useful things for people. We expect this will quickly become core to our…” (February 15, 2026).
