In pre-conference briefings, Nvidia pitched NemoClaw to enterprise firms including Salesforce, Cisco, Google, Adobe and CrowdStrike. These partnerships aim to accelerate adoption and demonstrate agent-driven automation across sales, security and design workflows. Nvidia plans to reveal additional details at its upcoming San Jose developer conference, including release dates, pricing and regional availability. The company will also unveil a new inference chip designed in collaboration with startup Groq. Nvidia is developing NemoClaw, an open platform that enables autonomous AI agents to execute enterprise tasks without constant supervision. The software is designed to integrate with existing workflows and includes safety protocols to mitigate risks from unpredictable agent behavior. NemoClaw represents Nvidia’s pivot from proprietary CUDA systems toward open-source AI tools. The platform is hardware-agnostic, allowing businesses to deploy agents on non-Nvidia GPUs and on-premises servers.