Uber Deploys Fleet of 500 Cars to Collect Data for Autonomous Taxis

Uber has unveiled a prototype of the Hyundai Ioniq 5, modified to collect real-world driving data. The vehicles are equipped with 14 cameras, 8 lidars, and 9 radars, with data processing handled by the Nvidia Drive Thor platform. A partnership with Roush Performance enabled the adaptation of these production electric vehicles for autonomous system needs.

A fleet of 500 such vehicles is planned for global deployment this year, with the first 50 hitting the roads by summer. The collected data will be synchronized and structured to train self-driving vehicle algorithms, including those of Uber’s partners—Avride, Waymo, and WeRide. Uber aims to create the most geographically diverse dataset for autonomous driving.

The AV Labs initiative is part of Uber’s strategy to advance autonomous technologies. Previously, the company collected data from thousands of vehicles with external cameras and hundreds of Lucid Air electric cars in the U.S. and Europe. The new project will expand AI testing and training capabilities, bringing mass adoption of robotaxis closer.