Starlink will have trouble with tunnels and crowded downtown skyscraper areas, etc, similar to satellite TV/radio/etc. Starlink also will have some limit to the number of subscribers it can handle per square mile (or whatever area metric you prefer) without bandwidth starvation (similar to terrestrial cellular networks). So if Tesla uses Starlink in the future, it is likely Tesla would still include a LTE/GSM/whatever cellular service as a backup when Starlink is unavailable and/or underperforming for a given location.
With a clear view of the sky, and not in a crowded (in terms of Starlink subscriber) area, Starlink would provide high bandwidth, low latency (the planned LEO and VLEO constellations will in some cases have better latency than terrestrial options) coverage, and likely do so at a highly competitive rate, even without having to give Tesla special pricing.
After all, typical cellular data service is usually some combination of limited bandwidth, limited data quota, or "unlimited" with scare quotes in which it eventually goes from fast to slow depending on your usage. But all these pricing schemes are just to extract more money from subscribers, none of the options guarantee a minimum speed (at most a maximum speed), so it would be more honest to have a flat rate. If Starlink doesn't participate in the pricing shenanigans, it could be cheaper just by being more honest.