In places where land is cheap, and there are single family houses, cars will be the most efficient way to commute for years to come. We'll switch to self-driving electric cars, but still.
The point of the boring company is to make excavation of tunnels an order of magnitude cheaper. [1] And if he succeeds, I suspect demand for the tunnels will push them out to the burbs.
Nor, presumably, transfer from an autonomous taxi or bus to a subway train at that station.
Maybe he's right? Think of this as a subway for cars!
Also people per hour is not really the metric the average person cares about. It's travel time to destination.
[1] http://www.lamag.com/driver/l-s-subway-cars-will-definitely-...
Also, each LA subway car can fit over 50 people seated and probably three times that standing. $3M divided by the price of the cheapest Tesla is about 85. So the number doesn't seem high at all.
Actually I'm not. I just want to point out that public trains cost huge amounts of money for investment, they need special infrastructure which many US cities don't have, and the only way they work feasibly is if the population density is high enough so the taxes will support them.
In other words, there's a huge upfront cost to even considering building a system that most cities/governments will choose roads over rail.
A bus on the other hand is $300,000, holds around 60 passengers, and uses the same infrastructure the city already has, and doesn't incur the need to have right of way fights with land owners.
What would you realistically do given the choice?
If I had the choice of spending $3 million to buy 10 buses, or to do nothing at all, I'd save the $3 million and wait until it's feasible to develop something that the city will keep wanting to fund.
That seems high to me, full-stop. The London Underground S-Stock order came to around £1 million per car (though was a much, much larger order); even smaller orders of Bombardier Movia family trains have come far closer to that than the $3 million per car.