Robotaxi
A robotaxi is a taxi that operates without a driver and steers its own route. It's booked via an app, and the technology behind it consists of sensors, maps, and software that has learned from sample drives.
A robotaxi is a taxi without a driver. You order it via an app on your phone, it stops at the curb, you get in the back, and you’re taken to your destination. No one sits in the driver’s seat, and the steering wheel moves on its own. The car decides for itself when to brake, turn, or yield to a pedestrian. This is made possible by numerous measuring devices on the roof and bumpers, as well as software that has learned from millions of kilometers driven. Robotaxis don’t operate everywhere today, but only in precisely defined city areas that have previously been mapped in great detail.
What’s at stake for cities and automakers with the driverless taxi
In a normal taxi ride, the biggest cost factor is the human behind the wheel. If this cost disappears, a ride could become significantly cheaper. This is exactly what the companies pouring billions into the technology are counting on. Whether the math works out remains open: sensors, maintenance, and remote monitoring also cost a lot of money.
For cities, more is at stake than just the fare. If rides become very cheap, some people will switch from the bus to the robotaxi. This could result in more traffic instead of less. Other experts expect the opposite: those who no longer need their own car free up parking spaces. Which effect predominates will only be known after years of real-world operation.
A third point is safety. Humans cause accidents through distraction, fatigue, and alcohol. A computer doesn’t have these problems. However, it does make mistakes that a human would never make, for instance when an unusual situation never occurred during training. That’s why authorities scrutinize robotaxi accident figures very closely, often individually and publicly.
Sensors, maps, and the human in the control center
A robotaxi must first know what’s happening around it. For this, cameras, radar, and usually lidar are mounted on the vehicle. Lidar is a laser scanner that scans the surroundings and creates a three-dimensional image from it. Radar measures distances and speeds using radio waves, even in rain and fog. The software merges all these measurements into a single situational picture.
In addition, there is a highly precise map. It is far more detailed than a navigation app and knows every stop line and every curb down to the centimeter. The car constantly compares what its sensors see with this map. This way, it knows very precisely where it is. Because such maps are elaborate to create, robotaxis are limited to individual cities.
The final step is the decision: brake, swerve, wait. This part was largely learned from recorded drives, similar to how a language model learns from texts. In tricky situations, such as a construction site with hand signals, the car contacts a control center. There, a human remotely provides guidance on which route the vehicle should take. So they don’t steer along with it, but rather give an approval.
Waymo, Tesla, and why hardly any are driving in Germany
In the US, robotaxis have become an everyday sight. Waymo, a subsidiary of the Google Group, operates in San Francisco, Phoenix, and Los Angeles and reports hundreds of thousands of paid rides per week. Tesla launched its own service in Austin in 2025, initially with safety personnel on board. In China, Baidu and Pony.ai operate similar fleets in several megacities.
In Germany, by contrast, one sees almost only test vehicles. This is due less to the technology than to the regulations and the weather: snow and dense fog cause problems for sensors. Legally, there has been a framework for driverless vehicles since 2021, but the permits so far apply only to small areas. Initial projects are running in Hamburg and Munich, for example.
In business news, the term mainly comes up in connection with stock prices. At Tesla, Uber, or Waymo’s parent company Alphabet, part of the valuation hinges on the hope for driverless fleets. A common misconception here: an assistance system that keeps a car in its lane on the highway is not yet a robotaxi. There, a human must be able to intervene at any time, whereas in a robotaxi, no one sits behind the wheel anymore.