Jensen Huang Ties Nvidia’s Future to OpenClaw and Agentic A.I.

At GTC, Nvidia CEO Jensen Huang unveiled NemoClaw, a secure proxy platform based on OpenClaw. Photo by Benjamin Fanjoy/Getty Images

Nvidia is betting that the future of enterprise AI will run on OpenClaw and that the company is the one to make it happen. At the GTC conference yesterday (March 15), CEO Jensen Huang unveiled NemoClaw, a new enterprise proxies platform built on the open source Viral Platform OpenClaw, positioning Nvidia as a safe way for businesses to adopt proxies.

NemoClaw is designed to bring OpenClaw “tentacles,” or autonomous AI agents, into enterprise environments with greater security and governance. The prepackaged software suite installs OpenClaw alongside Nvidia’s Nemotron models and a new runtime layer, adding sandboxing, privacy controls, and policy-based firewalls so agents who need full access to files and data can work more securely.

OpenClaw’s popularity has boomed since its launch four months ago, bringing standalone clients to work He took over their social platform Fueling the fan base that includes the New York City conference and increasing adoption in China. Huang called it “the most popular open source project in human history” and compared its role in AI to the Windows and Linux operating systems of earlier computing eras. “Now, OpenClaw has made it possible for us to create personal agents,” he said.

Nvidia worked closely with OpenClaw creator Peter Steinberger, an Austrian computer scientist known for his prolific development tools, to shape NemoClaw, Huang said. Steinberger launched OpenClaw in November 2025, and has since been hired by OpenAI to lead its personal agent division, even as the software remains an independent, open source project.

Nowhere has the rise of OpenClaw been more dramatic than in China, where companies like Tencent, Alibaba, and ByteDance have rushed to launch toolkits built around it. But regulators have been far more cautious, with Beijing warning companies of the security risks. Similar concerns have arisen in the United States; For example, Meta asked employees not to install OpenClaw on work machines after one agent Go rogue and delete the user’s email en masse.

Huang chose NemoClaw as an answer to these concerns. He said Nvidia worked with “the world’s best security and computing experts” to build the security and privacy controls that organizations expect. “Every company in the world today needs an OpenClaw strategy and a proxy system strategy – this is the new computer,” he told the GTC audience, and he has reportedly been promoting the platform to companies including Salesforce, Cisco, Google, Adobe and CrowdStrike.

Betting on OpenClaw comes from a position of tremendous power. Thanks to insatiable demand for its graphics processing units, which power most large-scale AI systems, Nvidia has become the world’s most valuable public company with a market capitalization of about $4.5 trillion. Huang, who originally founded Nvidia as a gaming chip maker before pivoting to artificial intelligence, told the audience he expects the company to sell $1 trillion worth of advanced graphics processing units through 2027 as demand for computing soars. Last year, he expected orders to reach $500 billion by the end of 2026.

Huang’s OpenClaw push and bullish sales outlook were the headline of this year’s GTC, often called the “Super Bowl of AI.” Speaking at a packed hockey arena in San Jose, he also previewed upcoming GPU architectures called Vera Rubin and Feynman and introduced a new inference-focused chip created in collaboration with startup Groq.

Jensen Huang is betting on OpenClaw with Nvidia's new NemoClaw Agent platform


Leave a Comment