Uber and Nvidia are partnering to build a fleet of 100,000 robotaxis. “This will be a new computing platform for us, and I expect it to be very successful,” said Ceo Jensen Huang at NVIDIA’s DC Edition of its much-anticipated GTC AI conference.
According to Nvidia, the fleet will begin to release in 2027, and 100,000 is the number of vehicles targeted, but there is no time yet when the companies will reach that level of measurement. It’s an ambitious goal, especially when you keep in mind that Waymo’s main competitor operates almost 2,000 robotaxis, according to filings from August. However, it is not even a very optimistic estimate for the autonomous vehicle industry. Elon Musk, the head of Robotaxi provider Tesla, wants to have “millions of teslas that work independently” in the second half of the next year.
The technology Uber will be using is based on Nvidia’s latest car computer, call AGX Hyperion 10, which Nvidia executives say will make the cars ready for automatic 4. That means the car will be able to take pride of place on its own, without a driver ready to take over, as long as it operates in designated areas. It’s only one level away from truly autonomous driving without boundaries, which the company hasn’t really achieved yet.
“Human robots are still developing, but at the same time, there’s one robot that’s clearly in the commentary space, and actually here, and that’s a robot on wheels,” Huang said.
“This entire manufacturer is historical so if any computer or sensor fails it is always a safe stop,” said Ali Kani, Phrice President of the car at Nvidia, said in a press conference. Huang boasts that “the sensor, the surrounding cameras, the radars, and the lidar enable us to achieve the highest level of cocoon around the coconut around security.”
Unlike other robotaxi providers such as Tesla and Waymo, uber will only operate as a network of independent operators and will not build a fleet. That task will fall to several automotive partners, including Stellantis, Mercedes-Benz, and electric car maker Lucid Motors.
“We created this system so that all car companies in the world can create cars. Cars can be sold, it can be a passenger,” said Huang.
Tech will also be supported by the Joint AI Data Factory built on Cosmos, a family of AI models released earlier this year that can be used to train humanoid robots.
Nvidia is trying to gain ground in the competitive landscape. Last week, GM’s strategy was revealed to return not only hands-off but also “Eyes-off” electricity in 2028, to the scrutiny of many investors and investor relations.
Earlier this year, at NVIDIA’s original GTC held in San Jose, California, Huang announced that GM will use the drive agx platform to develop its future fleet of self-driving cars.
Following GM’s announcement last week, Musk spent most of Tesla’s earnings assuring investors that the company’s artificial intelligence and autonomous driving efforts were rated “highly,” That Tesla has achieved “clarity” on full-scale self-driving, and that he now feels confident in expanding Tesla’s production as quickly as possible.
