Fun fact: you can do Zürich Madrid with 3 transfers and 14 hours. The car will take 16 hours. Obviously, nobody is going to do that, and the plane is a better alternative. Your example does not get solved by autonomous cars, I am quite certain nobody is going to do this in a rideshare either.

Most people commute to the nearest city, move within a city, or go see friends in a neighboring city. Yes, US is much bigger, but I am quite certain that most people don't do more than 75km on a daily basis. The US still has plenty of medium and high density, where good public transit would make a lot of sense, but it is seriously lacking. It is a political decision, not a inevitable fact of life that the US has poor transit compared to other countries.

The correct question to answer would be: how do you move people efficiently in a 75km radius? Are there trips that could be solved with a good train, tram or bus? Do you provide streets with bike lanes such that people feel safe riding the bike? What about pedestrians? Are autonomous cars really solving the congestion problem? Most of these cars still need to be parked during the night when demand is lower, exactly like a railyard. However, the space taken by one person in one car is much higher than any other forms of transit.

My point here is: autonomous rideshares may displace taxi services and Uber. But most importantly, good street and city design, which is hard but not impossible, matters much more than solving the autonomous car problem.