Nvidia is preparing to launch an open-source AI agent platform called NemoClaw, according to a WIRED report, as the chipmaker expands its push into enterprise AI infrastructure.

The platform is being pitched to enterprise software companies and would allow organizations to deploy AI agents capable of performing tasks for internal workforces. Sources told WIRED the system will work regardless of whether a company’s software runs on Nvidia chips.

The move comes ahead of Nvidia’s annual developer conference in San Jose next week. Nvidia has reportedly approached companies including Salesforce, Cisco, Google, Adobe, and CrowdStrike about potential partnerships around the agent platform.

Sources said partners could gain early access to the open-source system in exchange for contributing to development. Nvidia is also expected to integrate security and privacy tools aimed at making AI agents safer for enterprise environments.

Interest in autonomous AI agents has surged in recent months. These systems, often called claws, are designed to execute multi-step tasks with limited human supervision and can run locally on personal machines or inside enterprise workflows.

Earlier this year an autonomous agent project known as OpenClaw gained attention in Silicon Valley for its ability to run independently on personal computers and complete work tasks. OpenAI later hired the project’s creator and began open-sourcing the technology.

The push also comes as Nvidia continues expanding across the AI ecosystem. Earlier today the company announced a partnership with AI research startup Thinking Machines Lab, further deepening its involvement with AI developers.

Nvidia is also expected to unveil new inference hardware at its developer conference, including a system incorporating chips designed by AI startup Groq as part of a multibillion-dollar licensing agreement.