Jensen Huang called it “the most important software release probably ever.” Now his company wants a piece of it.
Nvidia is preparing to launch an open source AI agent platform called NemoClaw, according to a report citing people familiar with the company’s plans. The firm has been pitching the platform to corporate partners ahead of its annual developer conference next week.
Salesforce, Cisco, Google, Adobe, and CrowdStrike are among the companies said to be in talks for those partnerships. What specific benefits those companies would receive for associating with an open source tool remains unclear.
What NemoClaw Is Competing Against
The platform would be a direct competitor to OpenClaw — formerly known as Moltbot and Clawdbot — the system that drew widespread attention in January for letting users direct “always-on” AI agents from personal machines using any number of underlying models. Last month, OpenAI hired OpenClaw creator Peter Steinberger “to drive the next generation of personal agents,” as founder Sam Altman put it. The OpenClaw project itself will be run by an independent foundation with OpenAI’s support.
The sudden interest in OpenClaw has also driven a run on Mac Mini hardware with unified memory well-suited to running the tool — a sign of just how much corporate and consumer appetite exists for persistent agent software.
Why Nvidia Needs This
NemoClaw will reportedly run on machines without Nvidia‘s own GPUs. That detail matters. The company’s dominance in AI hardware is under pressure, with rivals developing chips and models that reduce dependence on its products. Close involvement with a widely adopted agent platform could help direct corporate partners back toward Nvidia’s own hardware and services.
There is also a straightforward business logic to agentic AI for the chipmaker. Tools like NemoClaw allow AI agents to work continuously on a project for hours or days at a time — exactly the kind of sustained, GPU-intensive workload that benefits Nvidia regardless of which company’s software is running it.
Security will be a deliberate feature. The report says Nvidia plans to include “security and privacy tools” in the platform — a response, at least implicitly, to the well-documented risks that come when OpenClaw is granted unfettered access to user data.
Meanwhile, the company recently halted production of its H200 AI chips intended for the Chinese market after China moved to curb those imports in favor of locally manufactured alternatives. The push into open source agent software comes as that hardware channel closes.
Photo by Andrey Matveev on Pexels
This article is a curated summary based on third-party sources. Source: Read the original article
