Jensen Huang unveils self-driving AI and production of Vera Rubin platform

黃仁勳發布自駕人工智慧（AI）及Vera Rubin運算平台生產方案

TAIPEI (Taiwan News) — Nvidia CEO Jensen Huang (黃仁勳) on Monday unveiled a new autonomous driving AI technology with reasoning capabilities at CES 2026 in Las Vegas and announced that the company’s next-generation Vera Rubin AI supercomputer has entered full-scale mass production through a key partnership with TSMC.

“The ChatGPT moment for physical AI is nearly here,” Huang said during a packed livestream of his keynote, per CNA. He said autonomous technology can be applied across a wide range of robotic systems.

Huang showed a demonstration video of an autonomous vehicle navigating rush hour traffic in downtown San Francisco with fully hands-free route planning. He also showcased two robots equipped with Nvidia Jetson computers that are capable of walking independently on two legs.

Huang said the first autonomous vehicles powered by Nvidia technology will hit US roads in the first quarter of this year. They are expected to enter Europe in the second quarter and roll out in Asia in the third and fourth quarters.

He said the Vera Rubin platform, which consists of six new chips, is now in full production. Compared with the previous Blackwell platform, inference costs are reduced to one-seventh, while the number of graphics processing units needed to train mixture-of-experts models is cut to one-quarter.

While introducing the platform, Huang held up a chip manufactured by TSMC that was co-developed with Nvidia and named Coupe. He described it as an “integrated silicon photonics process technology.”

Huang acknowledged that competition in AI is extremely intense, with companies racing toward the next breakthrough. “And so therefore, all of it is a computing problem. The faster you compute, the sooner you can get to the next level of the next frontier.”