Photo: ed br / Pexels
Jensen Huang says physical AI revenue could rocket from $10 billion to $100 billion in a decade, dwarfing even Nvidia's current generative AI windfall
Nvidia’s CEO Jensen Huang is placing a massive bet on robots, autonomous vehicles, and industrial machines, predicting that the market for physical AI will eventually be ten times larger than its digital counterpart. The company currently generates roughly $10 billion annually from physical AI, and Huang expects that figure to hit $100 billion within the next decade.
To put that growth trajectory in perspective, Nvidia’s entire data center business, the profit engine that turned the company into a multi-trillion-dollar behemoth, generates over $115 billion per year. Physical AI, if Huang’s projections hold, would approach that scale on its own.
The gap between ambition and reality #
Right now, physical AI contributes less than 3% of Nvidia’s total revenue. The data center segment, powered by insatiable demand for chips to train and run large language models, accounts for over 89%.
Huang has estimated the global market for industrial robotics and physical AI at roughly $50 trillion. PwC’s Strategy& division has offered a more conservative but still eye-popping estimate of around $490 billion by 2030.
The comments surfaced across several public appearances between late July and late August 2026, including a notable sit-down on the Y Combinator podcast. At Nvidia’s GTC events during this period, the company showcased over 100 robotic systems.
How Nvidia plans to get there #
The stack includes Jetson edge computers, compact processors designed to run AI models on robots and drones rather than in distant data centers. There’s Cosmos, a suite of world foundation models that help machines understand how physical environments behave. Isaac GR00T handles robotics-specific AI training. And Omniverse provides a simulation platform where companies can test robotic systems in virtual environments before deploying them in the real world.
Huang has emphasized that advances in realistic video generation are a critical bridge between digital and physical AI. If you can generate photorealistic simulations of how objects move, fall, break, and interact, you can train robots in simulation millions of times faster than you could in the real world.
The competitive landscape and open questions #
Partnerships with robotics firms and automakers are already forming, though navigating geopolitical headwinds adds complexity. Export restrictions on advanced chips to China, one of the world’s largest robotics markets, create a ceiling on how much of the physical AI opportunity Nvidia can actually capture.
A tenfold revenue increase over ten years implies a compound annual growth rate north of 25%, sustained for a full decade.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our