Nvidia chief executive Jensen Huang has said the next wave of artificial intelligence will depend on machines that understand the physical world, and that it will be felt most in the world's largest industries.
Speaking at the Hill & Valley Forum in Washington, Huang said today's AI systems lack basic physical common sense. The video is old and now recirculating.
Jensen Huang: The next wave of AI is physical AI and it's coming for the world's largest industries. “Most AI today doesn't understand the laws of physics, friction, inertia, or cause and effect.” “These are physical reasoning abilities that children and even our pets have.” “When you put physical AI into a physical object, you get robotics.” “Hopefully, in the next 10 years, this new generation of factories will be highly robotic.” “They'll help us deal with the severe labor shortage around the world.”
— Veee (@vikktorrrre) October 18, 2026
"The next wave requires us to understand things like the laws of physics, friction, inertia, cause and effect," he said.
He argued that this is the capability that separates current AI models from robots able to work in factories and warehouses.
"These are physical reasoning abilities that children and even our pets have," Huang said, adding that most AI systems do not have them.
He illustrated the gap with a simple example. A ball rolled off a kitchen counter, Huang said, may appear to an AI to have vanished, while a dog understands that it is still there and runs around the table to fetch it.
That grasp of object permanence, he said, is something a robot must learn, along with the fact that it cannot walk through a table and has to reason its way around it.
Huang called this field "physical AI". "When you take that physical AI and then you put it into a physical object called a robot, you get robotics," he said.
He also tied the technology to labour. "Hopefully, in the next 10 years, this new generation of factories will be highly robotic," Huang said. "They'll help us deal with the severe labor shortage around the world."
ALSO READ | Not All AI Chips Are Equal: Memory Is Bleeding, But Nvidia Is Wobbling Less
Nvidia's Robotics Push
The remarks reflect a strategy Nvidia has pursued for some time. At its GTC conference in 2025, the company unveiled GR00T N1, a foundation model for humanoid robots.
Huang has also described robots as "AI immigrants" that could fill worker shortages, an argument he has made at CES.
Nvidia's technology is increasingly being used in industrial robotics outside the US.
ALSO READ | Jensen Huang's Net Worth Crosses $200 Billion; Becomes World's 7th-Richest Person
Limits Of Today's Models
Huang's comments also point to a gap that researchers have long flagged. Large language models are trained mostly on text and images, which gives them fluency but little grounded understanding of how objects behave.
Closing that gap is the aim of physical AI, which relies on simulation and synthetic data to teach machines how the real world works.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.