Uber Technologies Inc. has announced plans to build a fleet of 100,000 autonomous vehicles powered by Nvidia Corp.’s technology, marking a major step towards making robotaxis more affordable for consumers.

According to Nvidia, the expansion is set to begin in 2027. The move builds upon a partnership formed earlier this year, under which Uber agreed to share driving data to help enhance Nvidia’s artificial intelligence models and chip systems used by carmakers to develop self-driving vehicles.

At its GTC conference in Washington, DC, Nvidia unveiled its new platform, Nvidia Drive AGX Hyperion 10, which enables automakers to integrate advanced hardware and sensors compatible with autonomous-driving software.

As part of this collaboration, Stellantis NV will be among the first manufacturers to deliver at least 5,000 Nvidia-powered robotaxis for Uber’s operations in the United States and internationally. Uber will manage end-to-end fleet operations, including remote assistance, charging, cleaning, maintenance, and customer support. Stellantis will partner with Foxconn for hardware and systems integration, with production scheduled to begin in 2028 following pilot testing in the US.

The commitment will allow Uber to steadily expand the number of robotaxis available on its platform, potentially reducing operating costs and accelerating commercialisation of the technology. Uber already collaborates with over a dozen autonomous vehicle developers and has invested in several as it envisions a future where human drivers and robotaxis operate side by side.

Currently, Uber offers autonomous rides in Austin and Atlanta through partnerships with Waymo, and with WeRide Inc. in Abu Dhabi and Saudi Arabia. However, the current fleet remains relatively small compared with Uber’s millions of human drivers and couriers.

The Nvidia alliance is expected to help lower long-term costs and increase availability of autonomous vehicles. Uber’s future and existing robotaxi partners, including Avride, May Mobility Inc., Momenta, Nuro Inc., Pony.ai, Wayve Technologies Ltd., and WeRide, will be able to use Nvidia’s technology to contribute to the 100,000-vehicle fleet target. This figure also includes 20,000 Lucid Gravity and Nuro vehicles that Uber previously committed to deploying over the next six years.

Additionally, Uber and Nvidia are developing a “robotaxi data factory” to support autonomous vehicle research. Uber will collect more than three million hours of robotaxi-specific driving data for AI training and model validation, while Nvidia will supply processors, AI models, and simulation tools.

“Together, these capabilities form a powerful data engine—spanning ingestion, labelling, scenario mining, synthetic data generation and large-scale training—that aims to shorten the path from pilot to profitable autonomy deployment,” Uber said in a statement.