Apple has enough money that it doesn't matter if this research is not successful. At least they want to try this, if the result is good enough they may eventually release a car. If not, they will just consider as a lost research cost.
If Apple doesn't expect to achieve a satisfactory return on investment they shouldn't enter a new business area, regardless of their liquid assets ('money reserves').
If expected ROI is too low they should return the money to investors (share buyback or dividends) and let investors choose a different investment with better prospects by themselves.
If Apple made decisions that way, I suspect they wouldn't have done as well as they did. While I'm sure they did plenty of research beforehand, their biggest successes don't strike me as products where 'a satisfactory return on investment' was certain or predictable enough to optimize for.
