NVIDIA's Jensen Huang: OpenClaw Will Be Like Windows — the "New Computer" Every Enterprise Needs

Lead

At GTC 2026, NVIDIA founder and CEO Jensen Huang (黄仁勋) told a global audience that every company must adopt an "OpenClaw strategy" — a plan to build and deploy AI agent systems that he called "the new computer." Huang argued OpenClaw, an open‑source agent project that has sparked global interest, provides the software stack enterprises need now to create personal and corporate intelligent agents.

OpenClaw compared to Windows, Linux and HTML

Huang lauded OpenClaw’s potential, likening its industry role to Windows for personal computing and to other foundational technologies such as Linux, Kubernetes and HTML. "OpenClaw makes it possible to create personal agents — the meaning of that is incredible," he said, adding the project supplies "what the industry needs at the moment." It has been reported that OpenClaw’s founder was recruited by OpenAI; regardless, Huang emphasized the open‑source project will continue to evolve and enable broad innovation.

Security, NemoClaw and hands‑on evangelism

Security, Huang warned, is the project's central challenge. To address it, NVIDIA unveiled NemoClaw — a company‑controlled build of OpenClaw that adds privacy and safety controls, including what Huang described as a "network guardrail" and a "privacy router" to prevent agents from executing unsafe commands inside enterprises. To drive adoption, NVIDIA staged a developer "make‑a‑claw" session at GTC so attendees could prototype custom agents on stage.

Hardware bets and geopolitical context

Huang paired the OpenClaw push with heavy hardware bets. He announced a new inference system integrating Groq technology and reiterated prior commercial moves — it has been reported that NVIDIA previously agreed to a reportedly $20 billion deal with Groq — and forecast that demand for Blackwell and Rubin chips could hit $1 trillion by 2027. But adoption will not occur in a geopolitical vacuum: Western export controls and broader US‑China tech tensions could shape where and how these agent stacks and accelerators are deployed, especially in enterprise and cloud environments that span jurisdictions.