quantifying impact of variable through variation

Hi all, I have a problem im trying to solve and I was hoping for some guidance on the approach. I have converted the real life problem to an easy to understand word problem:

You have 100 of the exact same car and lets say you have theoretically controlled for every single factor that could cause difference in life of car except 1, driver difference. If you have life results from those 100 cars and their 100 drivers, how would you predict the expected life of the car? The driver impact and difference between drivers is not quantifiable as a stand alone variable. Can the variation in life be used to predict true life expectancy if car?