Nvidia built its empire on gaming chips, but now it’s coming for your work tasks. The company plans to launch NemoClaw, an open platform for AI agents that handle jobs without constant hand-holding. The chipmaker has been pitching this software to enterprise companies ahead of its developer conference next week in San Jose.
Sources say Nvidia reached out to potential partners including Salesforce, Cisco, Google, Adobe, and CrowdStrike. Companies can access NemoClaw even if they don’t run on Nvidia hardware.
Why Nvidia is betting on autonomous agents
The move taps into a growing trend around “claws,” open-source tools that run locally and execute tasks without much supervision. Earlier this year, a tool called OpenClaw captivated Silicon Valley by running autonomously on personal computers.
OpenAI ended up acquiring that project and hiring its creator. But bringing these agents into workplaces carries risks. WIRED reported that some tech companies, including Meta, asked employees to avoid OpenClaw due to unpredictable behavior.
Last month, a Meta employee shared a story about an AI agent going rogue and mass deleting emails. That incident explains why Nvidia is emphasizing security tools as part of NemoClaw.
A strategic shift beyond chips
For Nvidia, NemoClaw isn’t just another software product. It’s part of a broader effort to court enterprise clients by offering extra protection for AI agents. The platform also marks Nvidia’s embrace of open-source models, a shift from its proprietary CUDA system that locks developers into Nvidia GPUs.
The timing matters. Leading AI labs now build their own custom chips, threatening Nvidia’s hardware dominance. By offering an open platform that works across any infrastructure, Nvidia positions itself as the trusted software layer regardless of what silicon runs underneath.
What to watch at Nvidia’s conference
Nvidia hasn’t publicly confirmed this platform yet. The company didn’t respond to comment requests. Representatives from Cisco, Google, Adobe, and CrowdStrike also stayed quiet. Key details remain unclear, including release date, pricing, and regional availability.
Next week’s developer conference in San Jose should bring answers. Nvidia also plans to reveal a new inference chip system there, incorporating design from startup Groq.
Between the hardware news and potential NemoClaw announcements, the conference will show how Nvidia plans to reshape the future of work.