The Nvidia CEO compared companies that lack an agentic AI plan with those that lacked a web site in 1998 and the tech community is paying attention.
Nvidia CEO Jensen Huang appeared on stage at the annual GTC developer conference last week and gave what could be the most urgent message of the company since its inception: the era of the autonomous AI agents has come, and those who fail to adapt to it will face potentially lethal consequences.
Anyone who is not considering what an independent agent might achieve in their operations, Huang said, is the same person who did not have a web site in 1998.
It was a stinging analogy - one calculated to shake boardrooms out of complacency. Similarly to how the internet revolution required every company to establish an online presence or face extinction, Huang is currently making the case that artificial intelligence agents are equally big wave technology.
OpenClaw is in the centre of the vision of Huang, a free and open-source platform to build and deploy autonomous AI agents. OpenClaw quickly spread in the Silicon Valley area and was featured by Huang as a characteristic platform of the present agentic AI phase.
GTC Huang also introduced NemoClaw, an open-source, chip-agnostic platform to build and deploy AI agents - and he made his stance absolutely clear: "Every company in the world should have an agentic system strategy. This is the new computer now." Huang made a brash historical analogy, making comparisons between OpenClaw and Windows in the personal computing age - it only took OpenClaw three weeks to become as important as Windows had been in thirty years. Like any company used to require an internet strategy and a cloud strategy, Huang said, every single firm in the world now requires an OpenClaw strategy.
The prediction by Huang is appalling. Huang foresaw that each engineer will be having a hundred agents on the last day of GTC 2026 on the All-In Podcast. He went to an extent of indicating that engineers who do not adopt AI tools are lagging behind and not saving money. He said he would be greatly alarmed if that engineer with a salary of half a million dollars did not eat at least a half million dollars worth of tokens. According to Huang, tokens are no longer a cost to reduce in order to facilitate AI. Their ability to be maximized.
The passion is not just a rhetoric. Nvidia is now projecting at least $1 trillion worth of purchase orders of its Blackwell and Vera Rubin technology by 2027, which is twice as large as the estimations of the past 12 months. Nvidia already collaborates with companies such as Adobe, Palantir, and Cisco, to improve agentic capabilities on their platforms using the Nvidia Agent Toolkit. Huang also introduced Cosmos, a new foundation model meant to build physical AI, which is systems that do not exist in the physical world like autonomous vehicles and robotics. He foresaw that the robotics business would be the first multi-trillion-dollar industry in robotics, independently.
Huang viewed the present as one of three colossal platform shifts occurring at the same time - the first in twenty-five years since the emergence of Moore Law. He said: Most of the industries of the world have not yet begun to make use of agentic AI, and they are about to. A survey of McKinsey in November 2025 identified 62 percent of organizations were experimenting with AI agents, at least, which Huang thinks will soar quickly in the coming years.
The message by Jensen Huang of GTC 2026 is clear and far-reaching, autonomous AI agents are not a future tendency, but rather a current reality. Firms that do not develop an agentic strategy in the present time will find themselves becoming the online equivalent of the company that bypassed the internet altogether.
The analogy of 1998 can be dramatic. However, with Nvidia having a trillion-dollar order book and the adoption of platforms, such as OpenClaw, exploding, Huang may be right.
Visit Karostartup for more insights into the intersection of technology, policy, and the future of India.
Quick Share
