Nvidia CEO Jensen Huang believes the next revolution in artificial intelligence won't happen on screens or in data centers, but in the physical world around us. During a keynote conversation with Adobe CEO Shantanu Narayen at the Adobe Summit, Huang outlined a bold vision for what he calls "Physical AI," a technology that enables computers to understand and interact with tangible, real-world environments.

Why Does the Physical World Matter More Than Digital Data?

While artificial intelligence has become remarkably skilled at processing digital information, Huang argues that the biggest industries and opportunities remain stubbornly physical. "The vast majority of the world is physical, and if we want to be able to apply computing for the first time to some of the largest industries in the world, whether it's life sciences or logistics and manufacturing or transportation, unless the computer can understand the physical world, there is no chance of enhancing it, no chance of automating again," Huang explained.

Huang

This insight reveals a critical gap in current AI development. Most generative AI systems, including large language models (LLMs), excel at pattern recognition and text generation but lack the ability to grasp how physical objects behave, interact, and change in real environments. Huang's argument suggests that unlocking value in industries like pharmaceuticals, supply chain management, and autonomous vehicles requires AI systems that can reason about the physical world with precision.

What Role Do Digital Twins Play in Physical AI?

At the heart of Huang's vision lies a concept he emphasizes repeatedly: the digital twin. A digital twin is a high-fidelity, three-dimensional virtual replica of a physical object, system, or process. Unlike approximate digital representations, these models must be exact and truthful in every detail.

"We need a high fidelity, truthful, in its most precise representation, which is 3D graphics, digital representation of the artifact. It could be a car, it could be a perfume bottle, it could be a person. Whatever it is. From there, we can then integrate it with generative AI and express our creativity through that," Huang stated.

Jensen Huang, CEO at Nvidia

Huang emphasized that this precision is non-negotiable for companies producing physical products. "Many of the people in this room are producing products, and they're marketing those products, but those products are very specific. That product needs to be precise. The brand identity has to be precise. The design is precise. It's not an approximate representation of the product," he noted.

Huang

How Can Companies Leverage Digital Twins and Physical AI?

• Manufacturing Simulation: Companies can test robotic systems and production workflows in virtual environments before deploying them on factory floors, reducing costly trial-and-error and minimizing downtime.
• Logistics Optimization: Digital twins of transportation networks and supply chains allow AI systems to identify inefficiencies and test improvements without disrupting real-world operations.
• Creative Integration: By merging generative AI with precise 3D digital artifacts, designers and engineers can use AI as a creative tool, generating variations and innovations based on accurate physical models.
• Life Sciences Applications: Precise digital representations of molecules, proteins, and biological systems enable AI to accelerate drug discovery and medical research.
• Risk-Free Testing: Virtual environments eliminate the need for expensive physical prototypes and real-world experiments, allowing companies to explore multiple scenarios quickly.

The practical implications are significant. A manufacturing company could build a digital twin of its entire production line, then use AI to simulate thousands of scenarios to find the most efficient workflow. A pharmaceutical company could create digital twins of molecular structures and use AI to predict how compounds will behave before synthesizing them in a lab. A transportation company could model its entire fleet and route network, then optimize it with AI without touching a single vehicle.

Huang's emphasis on digital twins reflects a broader shift in how Nvidia positions itself within the AI ecosystem. Rather than competing directly with companies building large language models or generative AI applications, Nvidia is positioning itself as the infrastructure provider enabling others to build physical AI systems. The company's graphics processing units (GPUs) and specialized chips are essential for rendering high-fidelity 3D models and running the simulations that digital twins require.

This vision also signals where Huang believes the next wave of AI value creation will occur. While generative AI has captured headlines and investment, the real economic impact may come from AI systems that can understand and optimize the physical world. Manufacturing, logistics, life sciences, and transportation represent trillions of dollars in economic activity. Even small efficiency gains powered by Physical AI could translate into enormous value creation.

The conversation between Huang and Narayen underscores a key challenge facing AI development today: moving beyond text and images to systems that genuinely understand physical reality. As companies race to integrate AI into their operations, those that master the creation and use of digital twins may gain a decisive competitive advantage.