Imagine a small single occupant autonomous car not much bigger than an enclosed motorcycle. You call it with a cellphone app. It figures out where you are and comes to pick you up, and it already knows where to go because you entered that in the app. Imagine them so small and sleek that they are less than 100 wh/mile to operate. A 4kWh battery will take you 40 miles, there is no significant city bus route longer than that. The 4kWh battery could be as small as 60 pounds, the whole vehicle could easily be less than 500 or 600 pounds. In between passengers they can go find a place to charge or battery swap if needed before the next passenger. Intelligent routing will send you a vehicle that has enough charge to get you to where you need to go. When more cars exist than rider demand, the excess cars find a place to charge or swap or just park and wait in some out of the way location. The average bus ride is on the order of 5 miles, not 40, so you would actually choose a battery size that lets the entire fleet continuously operate through peak load ( rush hour ) without having to stop to charge or swap. Average speed during rush hour is likely no more than 25mph, so 40 mile range would let them operate for 1.6 hours, which is probably plenty. At 100 wh/mile your fuel operating cost is 1 penny per mile. Compare that to a 40 seat city bus that costs $350,000 and gets somewhere around 6 mpg. If the average number of riders on that bus is 7, then its the equivalent of 42 mpg. At $4 per gallon for diesel thats 10 cents per mile. If the bus is actually full then its 240 mpg but that is 1.67 cents per mile. If the city bus is $350000 and serves an average of 7 riders, then each of those riders is riding around in $50000 worth of bus. The tiny single occupant EV with a 4kWh battery could likely be made for less than $10000 each with todays technology. So you could have 35 of them instead of one bus. Since the single occupant EVs are more flexible than buses you would actually need significantly less total capacity. A properly built battery electric bus may be able to compete on efficiency with the single seat EV if it is full, but it is likely to be a loser when you factor in that it needs a much larger battery to be able to continuously circle the route ( which it continues to drive even when empty ) and cant just go charge or idle when there is no passenger demand. The BEV bus is also likely to be much more expensive than the ICE city bus, and thus it would be much cheaper to just have 40 single seat EVs. On top of all the efficiency arguments, the autonomous car will take you exactly where you need to go and pick you up from wherever you are. Vastly more time efficient than any current mass transit. It is unlikely that any form of bus or train can compete on cost ( operational or capital ) with small single occupant autonomous EVs unless you are talking about some kind of super high speed travel.