Small vehicles will never solve traffic, and will never be cheap, no matter the level of automation they implement. Labour cost is just one part of the equation.
Clean, effective, prompt, affordable, quickly constructed mass transit however, does seem to work.
I'm not anti-robo-taxi or pro-mass-transit. Instead, if the value is fewer cars and less carparks, mass transit with the aforementioned properties has been shown to work.
Sadly the United States has been unable to hit those notes with their projects (LA's metro, for example, is still constructing approximately the Prop A system approved by voters, in 1980, 44 years ago: https://upload.wikimedia.org/wikipedia/commons/thumb/6/6f/Lo...) so in that specific case, it's murkier.
Where do those cars go to charge outside peak demand?