Nvidia is planning about launching an open source platform for AI agents, people familiar with the company’s plans tell WIRED.
The chipmaker offers a product called NemoClaw to enterprise software companies. The platform will enable these companies to send AI agents to perform tasks for their own employees. Sources say companies will be able to access the platform regardless of whether their products run on Nvidia chips.
The move comes as Nvidia prepares for its annual developer conference next week in San Jose. Ahead of the conference, Nvidia reached out to companies including Salesforce, Cisco, Google, Adobe and CrowdStrike to partner on its agent platform. It is unclear whether these talks have resulted in official partnerships. Since the platform is open source, it is likely that partners will receive free, early access in exchange for contributing to the project, sources said. Nvidia plans to offer security and privacy tools as part of a modern open-source agent platform.
Nvidia did not respond to a request for comment. Representatives for Cisco, Google, Adobe and CrowdStrike also did not respond to requests for comment. Salesforce did not provide a statement prior to publication.
Nvidia’s interest in agents comes as people start using “talons,” open-source AI tools that run locally on a user’s computer and perform sequential tasks. Claws are often described as self-learning because they should automatically improve over time. Earlier this year, an artificial intelligence agent known as OpenClaw – first Clawdbot and then Moltbot – wowed Silicon Valley for its ability to operate autonomously on personal computers and perform work tasks for users. OpenAI eventually took over the project and hired the creator behind it.
In recent years, OpenAI and Anthropic have significantly improved the reliability of models, but their chatbots still require hand-holding. On the other hand, purpose-built AI agents or claws are designed to perform multiple steps without as much human supervision.
The employ of claws in corporate environments is controversial. WIRED previously reported that some tech companies, including Meta, have asked employees to refrain from using OpenClaw on their work computers due to the agents’ unpredictability and potential security risks. Last month, a Meta employee who oversees security and customization in the company’s AI lab publicly shared the story about an AI agent that went rogue on her computer and mass-deleted her emails.
In Nvidia’s case, NemoClaw appears to be part of an attempt to woo enterprise software makers by offering additional layers of security to AI agents. It’s also another step toward the company’s employ of open-source AI models, part of a broader strategy to maintain its dominance in AI infrastructure at a time when leading AI labs build their own custom chips. Nvidia’s software strategy so far has relied heavily on the CUDA platform, the eminent proprietary system that allows developers to create software for Nvidia’s GPUs and has created a key “moat” for the company.
Last month, the Wall Street Journal. reported that Nvidia also plans to showcase a modern inferential computing chip system at its developer conference. The system will contain a chip designed by the start-up Groq, with which Nvidia signed a multi-billion license agreement at the end of last year.
Paresh Dave and Maxwell Zeff contributed to this report.