Dutch car-sharing operator MyWheels has started introducing Tesla’s Full Self-Driving Supervised software into its shared fleet, taking an advanced driver-assistance system into an environment where different customers may encounter the same vehicle and software. The company placed its first equipped shared car into operation on 27 August and plans to introduce the system across at least 50 vehicles during 2026.

Despite the product name, the cars are not autonomous. Tesla’s FSD Supervised can perform multiple driving tasks, but the driver remains responsible and must continue monitoring the road, a distinction the Dutch vehicle authority RDW has emphasised since approving the technology for use in the Netherlands.

MyWheels says the rollout will be gradual, using today’s supervised system to build operational experience before technology and regulation permit higher levels of automation. The immediate programme is consequently closer to a fleet-management experiment than a robotaxi launch, with paying customers still responsible for the journey even as software undertakes a growing proportion of the driving task.

Shared vehicles create a different operating environment from privately owned Teslas. Drivers may encounter the technology for the first time in a hired car, vehicles pass between users with differing levels of experience, and the fleet operator has to manage instruction, software configuration, maintenance, support, incident handling, and customer expectations across assets it owns centrally.

Approval has not removed driver responsibility

The Netherlands has played an important role in Tesla’s European approval process. RDW assessed FSD Supervised for more than 18 months and says that work involved more than 3,000 hours of testing, over 1,000 test runs, and analysis connected with 1.8 million kilometres driven in Europe before the authority concluded that the system met the applicable requirements.

RDW nevertheless describes the software as an advanced assistance system rather than autonomous driving. The driver must remain engaged and ready to intervene, while attention is monitored and post-approval supervision continues. Regulatory approval therefore establishes the conditions under which the assistance technology can be used rather than transferring responsibility from the person behind the wheel to the car.

That limitation shapes the commercial proposition for MyWheels. A genuinely driverless shared vehicle could reposition itself between customers and operate without somebody travelling to collect it, potentially increasing utilisation and reducing the friction involved in finding a nearby car. FSD Supervised does not yet enable that model because a responsible driver still has to be inside the vehicle throughout use.

A shared fleet can nevertheless provide unusually concentrated evidence about advanced driver assistance. One operator can observe software behaviour across different drivers, routes, and usage patterns, while utilisation, customer-support queries, interventions, incidents, and maintenance can be examined across vehicles managed within the same service.

Mobility software meets fleet economics

MyWheels is placing the driving-software deployment alongside an existing effort to turn electric shared cars into flexible energy assets. The company says nearly 400 vehicles already support vehicle-to-grid operation in the Netherlands, allowing batteries to feed electricity back into the network during periods of high demand, with an ambition to increase that fleet to 1,000 vehicles during 2026.

At that scale, MyWheels estimates the V2G fleet could provide more than 5MW of flexible power. Combining shared ownership, grid services, and increasingly automated driving creates vehicles whose economic value depends not solely on transporting an individual customer but on how effectively their batteries, software, availability, and physical location can be coordinated during the day.

The driving component remains the least mature part of that model. MyWheels describes a future in which a customer could request a vehicle that arrives autonomously and later moves to the next user, but current approval does not permit that operation. Present-day experience can instead help the company understand how customers behave when automation becomes another configurable feature inside a shared service.

Customer instruction will be one of the less visible challenges. Someone buying a Tesla can spend time learning the system’s limitations, warnings, and intervention behaviour, whereas a car-sharing user may enter an unfamiliar vehicle for a journey lasting only an hour. Operators introducing advanced assistance need to determine how much explanation is necessary before the journey begins and how strongly interfaces must prevent users from confusing supervised automation with autonomous driving.

European regulation can also make fleet deployment uneven. Approval and availability can differ between markets, meaning a service operating across several countries cannot assume that an identical software feature will be usable under the same conditions throughout its network.

The first 50 MyWheels vehicles are therefore a modest deployment with a useful operating difference from Tesla’s conventional customer base. Software can undertake more of the driving task, but the fleet operator owns the car, the customer remains responsible behind the wheel, and the regulator continues to treat the system as supervised assistance. How those responsibilities work in a shared fleet will provide more practical evidence than the product name alone.