Nvidia is preparing to launch an open-source AI agent platform called NemoClaw to compete with OpenClaw, according to a Wired report. The company has reportedly been pitching NemoClaw to corporate partners including Salesforce, Cisco, Google, Adobe, and CrowdStrike ahead of its annual developer conference next week.
The chipmaker best known for the GPUs powering nearly every large model now wants a foothold in the software layer where those models actually do work. That ambition puts Nvidia in direct competition with OpenAI, Anthropic, Microsoft, and Google, and it signals that the fight over personal, always-on agents is entering its next phase.
Key takeaways
- Nvidia is building NemoClaw, an open-source agent platform, to rival OpenClaw, and has pitched it to Salesforce, Cisco, Google, Adobe, and CrowdStrike before its developer conference next week (per Wired). - OpenClaw, formerly known as Moltbot and Clawdbot, went viral in January for letting users run always-on AI agents from their own machines on top of any underlying model. - OpenAI hired OpenClaw creator Peter Steinberger last month "to drive the next generation of personal agents," and the project will run under an independent foundation with OpenAI's support. - Nvidia is following the open-source playbook Meta used with Llama: seed ecosystem adoption through openness, then monetize through hardware and enterprise services. - For brands, the platform war matters less than the outcome it accelerates: more autonomous agents deciding which products and sources to surface, which makes AI visibility (GEO) a distribution problem, not a marketing afterthought.
What NemoClaw actually is
NemoClaw, as the awkward name hints, would be a direct competitor to OpenClaw, the system that drew attention in January for letting people direct always-on agents from their personal machines using any number of underlying models. Rather than a single assistant locked to one vendor's model, OpenClaw-style platforms are orchestration layers: they let an agent run continuously, pull in tools, and act across apps.
By making NemoClaw open source, Nvidia is not trying to sell the platform itself. It is trying to make its hardware and services the default substrate underneath a broad ecosystem of agents. Wired's sources did not spell out what Salesforce, Cisco, Google, Adobe, and CrowdStrike would get from associating with the tool, which is a reminder that this is early positioning ahead of a conference reveal, not a shipped product.
Why Nvidia wants the agent layer
Three strategic moves are stacked into one announcement.
The first is ecosystem expansion. Nvidia is moving up the stack from chips and infrastructure into the software layer where agents run. That deepens customer lock-in and opens new revenue lines beyond silicon.
The second is the open-source strategy itself. By open-sourcing NemoClaw, Nvidia borrows Meta's Llama approach: build adoption through openness, then capture value through enterprise services and, in Nvidia's case, the hardware everything runs on.
