21:56 06-01-2026

Nvidia maps robotaxi timeline, Mercedes-Benz demo, and new chips at CES

Nvidia, which in recent years has become one of the chief beneficiaries of the AI boom, now clearly wants to channel that clout into the auto industry. At CES in Las Vegas, CEO Jensen Huang said a robotaxi service based on Nvidia’s technologies could launch with a partner as early as 2027. The roadmap stretches further: between 2028 and 2030, the company intends to bring its autonomous driving solutions into consumer vehicles from multiple brands.

Tellingly, one of Nvidia’s flagship demos is tied to Mercedes-Benz. On the eve of the show, the new CLA completed an urban route in San Francisco, identifying signs, traffic lights, right-of-way rules and pedestrians. On the longer drive, a safety driver still had to intervene — a reminder that even with rapid progress, city streets remain the toughest arena for self-driving systems.

Nvidia’s bet isn’t on a single sensor, but on a blend. In dense traffic you need more than a clear picture; resilience to edge cases is crucial — the system must tell apart a person idling on the curb from someone about to step into traffic. In the CLA demo, the setup reportedly used 10 cameras and 5 radars. For robotaxis, the company also emphasizes lidar, which scans the scene and helps read surrounding geometry more reliably. Against that backdrop, Tesla’s camera-only philosophy looks increasingly like a lone stance across the industry.

Competition around robotaxis stood out at CES as well. Uber showed future autonomous EVs based on Lucid that it plans to deploy around San Francisco, while Amazon’s Zoox is already testing vehicles without a steering wheel or pedals in Las Vegas. By scale, Waymo is still regarded as the market leader, with thousands of driverless taxis operating in several U.S. cities.

Separately, Nvidia used CES to unveil its next chip generation: the start of production for the Vera Rubin platform, which the company describes as significantly more efficient than the previous Blackwell. That matters because autonomy hinges not only on software but also on the computing muscle available right on the car.