There are fundamental limitations on how dense automobile development can be.
But that begs the question--what is that limit? The highest density recent residential development I've seen in transit-less locations was an apartment complex with five residential stories over a garage podium. This, along with the success of the Gateway in Downtown, SLC seems to indicate that there is no limit--once land values rise high enough, developers simply start building parking structures and underground parking.
Taking this to its theoretical limit and it begin to approach the sort of city that early highway designers imagined--towering structures with comparatively tiny building footprints, with the rest of the land used for roadways. (Why it didn't turn out that way)
This suggests that the transit emphasis on density is to some degree a red herring--density alone does not make for a transit supportive environment. Perhaps a different calculation is needed. Density is typically estimated using households and jobs per acre. But one of the notable characteristics of sprawl is 'openness'. Sprawl development is sparse. This suggests that a simple figure ground drawing would be a powerful tool is assessing actual density. But building height is important, so perhaps a combination figure ground/contour map would be would be the best tool. A 'proximity' could then be calculated, dividing an area's gross square footage by its gross acreage. Doing so would separate areas where there are a lot of places in proximity to one another from areas where there is one thing in proximity to nothing.
In the end, the real limit to automotive density may be more social then economic--People like to walk. Not too far, and they like to have lots to look at and watch while they do it, but they like to walk. How else can the Vegas Strip be explained?