I have hosted the models (mostly regression and random forest) WV use to predict missing part availability at their dealership in 2018-2020 (considering sales in the area, average fabrication/delivery time, likeliness of the part having to be replaced, probably others).

I guarantee you that even if I don't like their car, their dealership will very likely have the part you need around the time you need it. It's not the only car-adjacent company that does something like this (Valeo for sure does it too, i worked with them also), but I'm pretty sure it's the only one who has an internal data scientist team working on it.