Nvidia CEO Jensen Huang: We've been thinking about robotics for a long time

Watch on YouTube ↗  |  January 06, 2026 at 13:30  |  7:41  |  CNBC
Speakers
Jensen Huang — CEO, NVIDIA
Jon Fortt — Host, Fortt on Tech

Summary

CNBC's Jon Fortt interviews Nvidia CEO Jensen Huang at CES 2026. Huang discusses Nvidia's open AI strategy, partnerships with xAI and Google, the Vera Rubin chip generation, AI factory economics, and the physical AI/robotics opportunity. He argues that Nvidia's broad ecosystem, falling token costs, and new chip integration support strong value creation. The interview also touches on long-term research areas like drug discovery and ray tracing.

  • Jensen Huang says Nvidia works with every AI company and runs every domain of science.
  • Huang highlights Vera Rubin as a 4x training leap over Blackwell with 10x lower token costs and 10x higher throughput.
  • Nvidia's gains come from integrating six new data-center chips across CPU, GPU, networking, switches, and DPU.
  • Huang describes AI factories that produce tokens, with lower token costs expected to increase usage.
  • Huang sees physical AI and robotics reaching an enabling-technology moment, citing action-generation models.
  • The interview mentions Nvidia partnerships with xAI and Google Gemini and long-term research in drug discovery and ray tracing.
Ideas
Jensen Huang CEO, NVIDIA 0:55
Open strategy broadens Nvidia ecosystem
Nvidia's Vera Rubin generation is a 4x training leap over Blackwell, cuts token costs by 10x, and increases AI factory token throughput by 10x, enabling customers to train faster, reach market sooner, or lower costs. The gains come from integrating innovations across six new data-center chips, including CPU, GPU, networking, switches, and DPU, making the value proposition compelling despite only 1.7x more transistors.
Jensen Huang CEO, NVIDIA 2:11
Physical AI moment is now
The enabling technology for physical AI and robotics is now ready because generative models can produce action tokens, such as steering-wheel activation or finger articulation, just as they generate language or video. Jensen sees this as a long-pursued moment whose time has come, with autonomous vehicles and humanoid robots as examples.
Up Next

This CNBC video, published January 06, 2026, features Jensen Huang discussing NVDA, ROBO, Physical AI. 2 trade ideas extracted by AI with direction and confidence scoring.

Speakers: Jensen Huang  · Tickers: NVDA, ROBO, Physical AI