NVIDIA says the next trillion-dollar AI wave will not have a screen.
By dedicating its flagship GTC event to physical AI, the chipmaker signaled that conversational bots are giving way to autonomous hardware.
In March 2026, NVIDIA opened its flagship GTC conference in San Jose with a definitive pivot. The keynote was not dominated by faster language models or better text generators. Instead, the focus was entirely on "physical AI" — unveiling the next generation of computing architecture designed explicitly to power autonomous factories, warehouse robotics, and edge devices.
For the past three years, the market assumed generative AI was primarily a software revolution. We obsessed over chatbots writing emails and agents navigating spreadsheets. But the chatbot was just a warm-up act. This follows the exact historical pattern of the internet: it began as text on a desktop screen before permanently embedding itself into our phones, cars, and thermostats. By shifting its primary processing narrative from server farms to warehouse floors, NVIDIA is signaling that the next massive market is not about AI talking to humans. It is about AI moving metal.
When the world's most valuable semiconductor company stops talking about pixels and starts talking about robotics, AI is officially leaving the screen.
Ask which physical processes in your business could benefit from embedded intelligence — not AI tools, but AI in the thing itself. The gap between smart software and smart hardware is closing.
Source: NVIDIA