Volkswagen and MOIA: Why the German robotaxi project could be coming to an end

It is a peculiar scene that can be observed in Hamburg: An electric VW van drives through city traffic, recognizes traffic lights, cyclists and pedestrians, stops at intersections and makes its way through the streets. Only one thing is still missing for the grand vision of the Volkswagen Group: the certainty that the whole thing can also be run as a business. A safety driver is still on board. And while MOIA tests the operation with selected passengers, in Wolfsburg apparently there is consideration about whether the company should even have a future within the group.

WirtschaftsWoche reported at the end of September, citing an internal strategy paper, that Volkswagen is examining an exit from the robotaxi business. Heise also picked up the investigation. On the table are, according to reports, various possibilities, from bringing in new investors to a sale to liquidation. A final decision is not known. But even the fact that such a withdrawal is being considered tells a story about the difficulties of the German automotive industry to turn technological promises into a viable business model.

From automaker to mobility provider

MOIA should be more than just another subsidiary with an app. Volkswagen wanted to show that a car manufacturer not only sells vehicles but also can organize mobility. The service started in 2018 in Hamburg and initially relied on ride pooling: Several passengers share a ride, the software bundles their routes. The company thus gained experience in dispatching, fleet operation and dealing with fluctuating demand.

Der nächste Schritt sollte den Fahrer überflüssig machen – zumindest perspektivisch. Volkswagen Nutzfahrzeuge entwickelte dafür den ID. Buzz AD, eine für autonome Fahrdienste ausgelegte Version des elektrischen Bulli. Ursprünglich sollte der kommerzielle Einsatz bereits 2025 beginnen. Als Technologiepartner war zunächst Argo AI vorgesehen, an der Volkswagen und Ford beteiligt waren. Im Oktober 2022 zog sich Volkswagen aus der Finanzierung des Unternehmens zurück. Argo AI wurde anschließend abgewickelt. Für das Robotaxi-Projekt bedeutete das einen empfindlichen Rückschlag. Volkswagen setzte die Entwicklung mit Mobileye fort. Das israelische Technologieunternehmen liefert das System, mit dem das Fahrzeug seine Umgebung erfasst und Fahrentscheidungen umsetzt. MOIA wiederum sollte die Software für die Buchung und Disposition, das Flottenmanagement und den operativen Betrieb beisteuern. Aus Auto, Sensorik und App sollte ein fertiges Mobilitätsangebot werden.

Das Konzept klingt plausibel. Es ist aber anspruchsvoller, als einen Bus mit zusätzlichen Kameras und Rechnern auszustatten. Ein autonomer Fahrdienst muss auch dann funktionieren, wenn ein Fahrzeug blockiert wird, ein Fahrgast Hilfe benötigt oder eine Baustelle die geplante Route unbrauchbar macht. Was bislang ein Fahrer erledigt, verteilt sich nun auf Technik, Leitstelle und gegebenenfalls Personal vor Ort. Das System muss nicht nur fahren, sondern einen Betrieb organisieren.

Technology is moving ahead; the business can't keep up.

In July 2026 MOIA in Hamburg reached an important milestone. Selected, previously registered passengers have since been able to book rides with autonomous ID.-Buzz vehicles as part of the ALIKE research project. The offer is geographically limited and remains part of a test operation. A safety driver remains on board.

The launch occurred with substantial delay. According to NDR, initially ten autonomous minibuses were on the road in Hamburg, fewer than the originally planned 25. The originally targeted removal of the safety driver by the end of 2026 also now appears questionable. NDR also reported a test ride in which the safety driver had to intervene because the vehicle, when turning in rain, did not adequately account for a crosswalk with two pedestrians.This is not proof that autonomous driving generally does not work. It shows, however, how large the gap can be between a successful demonstration and a reliably operating system in the complex urban traffic. A human driver compensates for uncertainties with experience, attention and intuition. An autonomous system must technically detect, assess and safely handle such situations.

In addition, there is the economic question. The vehicles must be developed, approved, maintained and insured. Sensors and computers cost money, as does the infrastructure for monitoring and deployment planning. As long as people sit in the vehicle to intervene in an emergency, a substantial part of the hoped-for cost advantage is lost. Only when a fleet can be deployed reliably and at a larger scale without a safety driver does the economics change fundamentally.

Billions for an Open Bet

Volkswagen is under cost pressure. The group must prioritize its investments, solve its software problems, and at the same time finance the transition to electric vehicles. In this environment, autonomous driving is a particularly difficult bet: the expenditure is incurred today, while a commercially viable market has yet to emerge.

Following reports about internal exit scenarios, MOIA needs billions more to scale the business to a larger fleet. At the same time, Volkswagen has reportedly been seeking potential investors. According to media reports, talks with Uber and Lyft recently proved unsuccessful. Whether and in what form additional partners will join remains open.The problem is not only about financing. An external partner would have to be willing to contribute capital and technology without losing control over its own business. Volkswagen, in turn, would have to accept that a partner would have influence over vehicle selection, software, and operating model. Whoever bears a large part of the investment would, after all, want to have a say in how the fleet is deployed and how the costs are amortized.

The calculation is tricky. An autonomous ride-hailing service needs many vehicles to spread development and operating costs over enough trips. At the same time, demand in a region must be large enough for the vehicles to be fully utilized. A small fleet can function technically, but may remain economically unattractive. A large fleet, in turn, requires capital, permits, and an operation that runs reliably in everyday use.