Beijing — If Nvidia CEO Jensen Huang were a university student in 2025, he says he would pursue the physical sciences over software, anticipating the coming era of “Physical AI.”
Speaking to journalists in Beijing on Wednesday, Huang reflected on what his academic focus might be if he were a 20-year-old graduate today. He explained that while software shaped his own career, the future of innovation will demand expertise in physics, chemistry, astronomy, and other fields within the physical sciences.
Huang’s own academic path began with an electrical engineering degree from Oregon State University in 1984, followed by a master’s degree from Stanford University in 1992. Just one year later, in April 1993, he co-founded Nvidia over a meal at a Denny’s restaurant in San Jose, California. Under his leadership, Nvidia has evolved into the world’s most valuable company, recently surpassing a $4 trillion market capitalization.
While Huang did not elaborate in Beijing on why he would choose physical sciences today, his recent speeches have emphasized what he calls the “next wave” of artificial intelligence — Physical AI. Over the last decade, Huang has described AI development in phases: the early “Perception AI” era, sparked by breakthroughs like AlexNet in 2012; the rise of “Generative AI” capable of creating images, code, and text; and the current stage, “Reasoning AI,” which enables problem-solving beyond pre-programmed scenarios.
The next leap, Huang says, will involve AI systems with physical reasoning — understanding the laws of physics, inertia, and cause-and-effect relationships. This evolution will enable AI to predict outcomes, manipulate objects with precision, and interact safely in dynamic environments.
“When you take that physical AI and put it into a physical object called a robot, you get robotics,” Huang noted. He envisions a future where robotics plays a central role in manufacturing, addressing global labor shortages.
Over the next decade, Huang expects factories worldwide — including a new wave of plants in the United States — to be built with high levels of automation. These robotic systems, powered by physical AI, could transform industrial productivity and redefine how humans and machines collaborate in the workplace.
