Uber is sending 500 Ioniq 5s onto city streets, each packed with lidar, radar and camera arrays.
The cars, retrofitted by Uber’s AV Labs team, will record traffic patterns, pedestrian movements and road conditions while appearing as regular EVs. Uber plans to roll them out over the next twelve months, gathering the raw data needed to train its autonomous‑driving models.
The move sidesteps the need for expensive, purpose‑built test rigs and gives Uber a steady stream of real‑world inputs. If the data proves useful, it could accelerate the timeline for Uber‑branded driverless rides, but it also raises privacy questions about ubiquitous sensor fleets.
In short, Uber is betting on quantity over bespoke hardware, hoping volume will translate into usable machine‑learning material.