By Franziska Müller
Published on
The decision was announced in Taipei, but it actually concerns a completely different city: in Munich, Uber plans to deploy autonomous robotaxis together with the AI company Autobrains.
Uber has picked the Bavarian capital: “Subject to regulatory approval, Munich will serve as the first launch city for the robotaxi programme,” the company said in a statement.
That could make Munich the first German city where passengers can order autonomous robotaxis via the Uber app. The companies involved unveiled the plans at the GTC technology conference in Taipei.
Munich as a test bed for autonomous driving
Uber justifies its choice of Munich with the city’s role as a European automotive hub, its dense urban traffic and Germany’s regulatory framework for autonomous driving. The aim is to develop a model that can later be used in other cities and on different vehicle platforms.
Unlike many previous robotaxi projects, the technology is not intended to be tailored to specially developed vehicles.
Instead, the partners are pursuing a so-called approach designed to be compatible with various carmakers. Industry observers see this as an attempt to bring down the high costs of autonomous fleets and, ideally, to scale up the technology more quickly, in other words to roll it out in additional cities.
So far, the companies have not said which manufacturers are involved, how large the fleet will be or when the first passengers might actually be carried.
What lies behind “agentic AI”
At the heart of the project is the driving software developed by the Israeli company Autobrains. Unlike many of its competitors, the company says it does not rely on a single large AI model that handles all driving tasks.
“Autonomous driving will not be scalable if you rely on a single model to solve every driving scenario,” Autobrains founder Igal Raichelgauz said.
Instead, the overall driving task is divided up among several specialised AI agents that assess different traffic situations and make decisions in real time. This is intended to make the system more robust in the face of complex and unpredictable situations. The approach is known as agentic AI.
The company, which has been working on solutions for driver-assistance systems since 2018, argues that multiple specialised systems can cope better with uncertainty than a monolithic approach.
The vehicles are to run on Nvidia’s DRIVE Hyperion platform, a computing and sensor architecture for Level 4 autonomous vehicles. This level of automation in principle permits driverless journeys within defined operating areas.
Uber banks on partnerships instead of its own robotaxis
For Uber, the Munich project is part of a broader strategy. The company now follows a platform model and no longer develops autonomous driving systems in-house. Instead, Uber works with various technology partners and integrates their vehicles into its existing ride-hailing network. Ride-hailing refers to ordering transport services via digital apps.
In recent months, Uber has already announced similar partnerships with other providers of autonomous driving technologies. Together with Nvidia, the company is planning in the longer term to deploy autonomous fleets in several dozen cities worldwide.
Nvidia’s vice-president for automotive, Ali Kani, told Euronews in January that partially autonomous driving would arrive later this year. “We need to move forward as quickly as regulation allows. And I believe it is opening up,” Kani said.
Competition in this market is increasing at the same time. In the United States, Google sister company Waymo already operates commercial robotaxi services in several major cities. Tesla, Mobileye and various Chinese providers are also working on autonomous fleet solutions. Tesla boss Elon Musk has predicted that self-driving cars could dominate road traffic in as little as five years.
Europe’s robotaxi market is only just beginning
A commercial robotaxi service would be a major step for Germany. While autonomous vehicles are already part of everyday life in cities such as San Francisco, Phoenix or Beijing, Europe is still at an early stage when it comes to large-scale deployment. Autonomous taxis in Madrid were originally due to be rolled out from 2026.
The companies are therefore presenting Munich as a potential springboard for a broader European expansion. Whether this actually turns into a regular robotaxi service, however, will depend not only on the technology, but also on regulatory approvals, proof of safety and the economic viability of the model.