If you're going to send AP data up, and receive map data down, even if you magically knew which data to send down separately, you'd still be doing 2-way comms. And you'll need the same antenna regardless.
In fact, because the satellites are whipping across the sky, steering spot beams along the ground, in order to send or receive you must have yourself a steerable beam phased array antenna to receive much less send. Basically, you electronically "point" the antenna at the satellite, just as it is doing to you.
Because of constraints when no / poor line of sight (garages, built up downtown areas with skyscrapers, valleys, tunnels, etc), you'd need to keep cellular service anyways, plus there's only so many terminals that can be served in a given area (polarization (left circular / right circular) * some number of max frequencies per spot beam * some number of max multiplexing of time slots or whatever).
so while in theory they could save money (or at least pay another Musk entity) by using starlink preferentially to or in replacement of cellular service, you still need the cellular service for many scenarios.
It doesn't make sense to sell terminals to Tesla when they won't want to pay more than cellular, when they can instead sell the limited number per area to those they can charge more (such as cellular backhaul, or internet backhaul, big companies that want dedicated high bandwidth private links between offices, etc).
I expect few if any individuals to be sold the service in populous areas, though I expect them to price it very competitively in rural areas, and perhaps practically give it away free to places in Africa or wherever that don't even have utilities (send in a "care package" container/crate of solar panels, battery, water purifier, starlink terminal with wifi bridge, etc).