Dubai, United Arab Emirates — A Taxi Without Anyone at the Wheel

Dubai has begun putting a strange kind of passenger vehicle on ordinary city streets: a taxi that can drive without anyone behind the wheel. The emirate’s robotaxi program moved from supervised testing into commercial driverless service in 2026, initially in limited coastal districts including Jumeirah and Umm Suqeim. Riders can request vehicles through the relevant app, while cars equipped with cameras, radar, lidar and onboard autonomous-driving software manage their own navigation within approved operating areas. The experience looks familiar from the curb — a ride booked by phone, a vehicle pulling up, a trip across the neighborhood — except that the front seat does not require a driver.

The initiative is not a single experimental prototype. Dubai’s Roads and Transport Authority (RTA) worked with multiple autonomous-driving companies, notably WeRide and Baidu’s Apollo Go, and local fleet-management partners. A public WeRide pilot booked through Uber started in December 2025 with a safety specialist on board. On 30–31 March 2026, the RTA and WeRide/Uber announced the start of commercial trips without a vehicle operator aboard. Baidu Apollo Go also began driverless commercial rides in Dubai in early April 2026 using its RT6 platform and its own ride-hailing app. These are distinct operator programs under a common city-level autonomous-mobility policy, not interchangeable names for the same taxi fleet.

The machinery behind the experience is more complex than an empty driver’s seat suggests. WeRide’s Robotaxi GXR and Apollo Go’s purpose-built RT6 use overlapping categories of sensing and compute, but the systems are developed independently. Sensors build a picture of surrounding vehicles, cyclists, road geometry and pedestrians. Software estimates how the scene will evolve, selects a trajectory and controls braking, steering and acceleration. Mapping, remote fleet monitoring, vehicle health checks and clearly defined operating zones make the service manageable in a busy city. The RTA describes the RT6 platform as equipped with more than forty sensors, including lidar, radar and cameras. These claims describe the technology architecture; they do not, on their own, prove how it performs under all weather or traffic conditions.

The distinction between a demonstration and a real transport service matters. By 19 August 2026, the RTA reported a program fleet of 144 vehicles, of which 48 were in active operation, along with 7,613 completed passenger trips. Those figures mark progress, but also show that a visually dramatic launch does not automatically mean citywide scale. Operations remain subject to defined service areas, permits, fleet availability and practical constraints. The program is one contribution to Dubai’s stated policy target of making 25 percent of transport journeys autonomous by 2030; it is not evidence that this proportion has already been reached.

For a Smart City Visual Atlas, Dubai offers a provocative engineering question. The visible innovation is a driverless car, but the real city-scale system includes regulation, incident response, fleet economics, passenger trust, cybersecurity, liability, accessibility and integration with traditional taxis and mass transit. Robots may eventually take over the steering wheel; the city still has to decide where they can operate, what performance is acceptable, who is accountable when something goes wrong and whether the result improves mobility for residents. That is the difference between a technology showcase and an urban transport system.