I think a major reason for these models is that the more that the car becomes a computing device, the more that it’ll require regular patches and optimizations. Being connected to the servers and using services that route through it lets them gather usage data, offer some extra features that can functiom from anywhere, and update security and functionality (which would possibly involve full time developers I suppose).
It does seem greedy (way overpriced), but this isn’t the same as disabling hardware that you need to sub to activate (a la seat warmers). Plus it’s all still pretty cutting edge tech atm and I usually tell people that means you’re choosing to fund its early development (and being a beta tester) over using more standard and tested products.
I think e-cars are more computer-like as they’ll prioritize optimizing as much of the system as possible to maximize battery mileage; performance/riding experience as a live service; DRM; probably pretty hackable.
Driverless autonomy could also potentially turn pedestrian cars into part of the public transport system if people can have their idle cars work like taxis (not sure if this would involve things like smart contracts), but unfortunately it seems like the actual last piece of the puzzle that car companies aren’t gonna crack any time soon.