NVIDIA is preparing to launch an open-source AI agent platform called NemoClaw, built for companies that want to deploy AI agents across their workforces.
Wired first reported the news on March 9, 2026, citing people with direct knowledge of Nvidia's plans. NVIDIA has not made a public announcement yet, but is expected to reveal the product at the annual GTC, its developer conference that will run from March 15 to 19 in San Jose
NemoClaw will let enterprise companies send out AI agents to carry out tasks on behalf of their employees. The platform comes with built-in security and privacy tools. It is also hardware-agnostic — companies can run it regardless of whether their products are built on Nvidia chips. NVIDIA's existing CUDA platform has long tied developers to NVIDIA's GPU ecosystem with little room to build outside it. NemoClaw will be able to run on any hardware.
Per the report, NemoClaw is expected to be open source, giving early partners the ability to build AI agents without relying on proprietary commercial APIs. NVIDIA has reportedly held conversations with Salesforce, Cisco, Google, Adobe, and CrowdStrike about early partnerships. None of those companies has confirmed a deal.
NemoClaw vs OpenClaw: What Is the Difference
OpenClaw—an AI agent previously called Clawdbot and Moltbot—captivated Silicon Valley earlier this year after it drew attention for handling real work tasks on personal computers with no human direction. OpenAI acquired the project and hired its creator.
OpenClaw was built for individual users. NemoClaw is built for companies. Wired previously reported that Meta restricted employees from using OpenClaw on work devices due to unpredictability and security concerns. A Meta employee working on AI safety publicly described an incident where an agent accessed her machine without instruction and deleted her emails in bulk.
What Nvidia Is Announcing at GTC 2026
CEO Jensen Huang is slated to deliver the keynote at GTC on March 16. NemoClaw is expected to be one of the headline announcements at the event.
Separately, the Wall Street Journal reported that Nvidia also plans to unveil a new inference chip system developed by Groq at the conference. NVIDIA and Groq finalised a multibillion-dollar licensing agreement in late 2025.
Damilare Odedina
