Cars are an expensive industry to get established in with a lot of details to get right and major regulatory concerns. My guess is that they looked at that, the struggles the entire field is having (Google's been at it for ages and is nowhere near ready for a consumer product), and decided that the better outcome is to build better integrations with existing companies which are really good at building cars — I know multiple people who decided what to buy in part based on who did the best job with CarPlay.

AR is more complicated: there are some neat ideas but the form-factors are brutal in terms of hardware design and there is weak consumer demand. Apple has one of the most popular AR products now (AirPods) but that took a while to mitigate the compromises forced by the physical limits. I'm expecting some additional hardware from them soonish based on the references in recent developer builds but will not be at all surprised if that takes a while to find popular adoption. I think it's easy to underestimate just how hard it is to deliver an application which is compelling enough to get people to pay non-trivial amounts for. VR has been educational in that regard going from something which was too hard / expensive to fairly available but not yet having found mainstream demand.