Nvidia Unveils Vera Rubin AI Chips at CES 2026

Nvidia CEO Jensen Huang unveiled the Vera Rubin architecture at the Consumer Electronics Show in Las Vegas on January 6. The new computing platform is now in production and expected to ramp up in volume during the second half of 2026. This launch comes months ahead of the late-2026 timeline Nvidia had previously projected.

Huang stated that Rubin marks a leap in performance compared to the Blackwell model. The new architecture offers more than triple the speed, runs inference five times faster, and delivers significantly more inference compute per watt of energy. The platform is designed to address the skyrocketing amount of computation necessary for AI.

Alongside the chip announcement, Nvidia introduced Alpamayo, a new open AI model for autonomous vehicles. Mercedes-Benz cars powered by this system are expected to hit the road in Q1 2026. Furthermore, Boston Dynamics announced a partnership with Google DeepMind to integrate Gemini AI into its Atlas robot.

The Consumer Electronics Show (CES) kicked off on January 6 in Las Vegas, with Nvidia CEO Jensen Huang delivering a keynote presentation that introduced the Vera Rubin architecture. Huang officially unveiled the platform ahead of the show, announcing that it is already in production. The architecture is expected to ramp up in volume in the second half of 2026, following a blockbuster year for the Blackwell chip.

In a presentation at CES, Huang explained the motivation behind the new architecture. "Vera Rubin is designed to address this fundamental challenge that we have: The amount of computation necessary for AI is skyrocketing," Huang told the audience. The architecture is named after astronomer Vera Rubin, who discovered the existence of dark matter.

Nvidia's senior director of HPC and AI infrastructure solutions, Dion Harris, described the platform as "six chips that make one AI supercomputer." The architecture is designed to support more complex, agent-style AI workloads, as well as increased networking and data movement.

The performance gains over the previous generation are significant. Compared to the Blackwell model, Rubin offers:

• More than triple the speed
• Five times faster inference capabilities
• Significantly more inference compute per watt of energy

The Vera Rubin systems are already lined up for deployment across much of the cloud industry. Nvidia confirmed that major partners plan to use the new platform. These include Amazon Web Services, OpenAI, and Anthropic. Additionally, the upcoming Doudna system at Lawrence Berkeley National Laboratory plans to utilize the platform.

The accelerated launch follows a period of record financial performance for Nvidia. The company recently reported record data center revenue, which was up 66% from a year earlier. This growth was driven largely by demand for Blackwell and Blackwell Ultra GPUs. Those chips have become a benchmark for the current AI boom.

Global investment in AI infrastructure continues to grow. Huang has previously estimated that between $3 trillion and $4 trillion could be spent globally on AI infrastructure over the next five years. Nvidia stated that products and services built on the Rubin platform will begin rolling out from partners in the second half of 2026.

Nvidia also used the CES presentation to push further into the automotive sector. Huang unveiled Alpamayo, a new open AI model and toolset designed to bring reasoning to autonomous vehicles. The goal of Alpamayo is to use vision-language-action models to handle rare and unfamiliar driving scenarios.

Mercedes-Benz CEO Ola Källenius joined the presentation to discuss the partnership. Källenius stated that he recently tested the Nvidia-powered system on public roads, driving for more than an hour through heavy traffic. "It feels that the car is on rails. You're just driving, and it does everything," Källenius said regarding the experience.

Huang announced that Mercedes-Benz cars powered by the system are expected to hit the road in the first quarter of 2026. "Our vision is that someday, every single car, every single truck, will be autonomous," Huang stated. Alpamayo represents Nvidia's first full-stack push into self-driving technology.

Robotics was another major theme at the show. Boston Dynamics announced a partnership with Google DeepMind to integrate the Gemini AI into its robots. The collaboration involves the humanoid robot Atlas and the robot dog Spot. The partnership was unveiled with Boston Dynamics' parent company, Hyundai Motor Group.

The integration aims to let Atlas understand natural-language commands, adapt on the fly, and interact more naturally with people. Executives from Boston Dynamics noted that embedding a foundation model allows Atlas to move beyond preprogrammed actions. It enables the robot to navigate unfamiliar environments and identify and manipulate objects.

"Atlas is going to revolutionize the way industry works," said CEO Robert Playter. He described the development as the first step toward a long-term goal of creating useful robots that can walk into homes and help make lives safer and more productive.