Nvidia showcases next AI leap with Vera Rubin

Nvidia CEO Jensen Huang unveiled its newest AI server system, called Vera Rubin, at the Consumer Electronics Show in Las Vegas—accelerating the global AI race.

The system, named after US astronomer Vera Rubin, goes on sale in the second half of 2026. It is designed to handle the enormous computing demands of advanced AI training and simulations.

Nvidia says AI models now require far more computing power to think and respond in real time. 

Vera Rubin can train massive models with up to 10 trillion data parameters in a month using just a quarter of the chips needed with Nvidia’s previous generation, Blackwell, while delivering a 10-fold cost reduction for inference tasks.

Nvidia also introduced new software libraries, connectivity tools, and memory-storage tools to speed up computing and support robotics and autonomous vehicles, which Huang calls “physical AI.”

Analysts call Vera Rubin a generational leap, and the early unveiling signals the system will reach the market quickly. Despite this, Nvidia’s stock ended nearly flat.

Industry experts say the early launch shows Nvidia’s urgency to stay ahead as AI demand continues to surge worldwide.