That's just the current cost. The long term cost structure should be based on cars that come out of the factory with all the right stuff pre-installed. There's a BOM for some extra components; many of which you might already find in some cars. Otherwise it's just another EV. So, long term the extra cost per mile relative to a driver driving the same car should be cents rather than dollars per mile. And of course if there is no driver, some components like manual controls, dashboards, mirrors, etc. actually become redundant as well. So the total BOM might actually be lower long term.

The driver cost is of course the big saving. And they need breaks as well and don't drive 24x7. A robo taxi only has down time when there are no rides, or for charging and maintenance/cleaning.

Mostly what Waymo is doing currently with customized vehicles is not actually super scalable. But it helps at their relatively small current scale. You wouldn't design a custom vehicle + factory for their current growth rate. That becomes more interesting when they start scaling beyond tens of thousands of new vehicles per year. They are probably in the lower thousands currently.

I think they raised close to 20-30B so far. They say they are doing 500K rides per week. At 15$ per ride that adds up to ~390M/year. That's revenue, not profit. But if they could 100x that by rolling out to more and more cities and larger and larger areas, it's going to add up to annual revenues that add up to more than what they raised. That's not going to happen overnight, obviously. But they seem on a path where they are scaling, optimizing, reducing their cost, and growing.

The risks here are mainly that they won't have the market to themselves. Others are doing robo taxi's too and if any of them starts scaling faster and cheaper, Waymo could hit some growth issues. Also, with multiple companies competing, prices per ride would eventually go down. The next five years are going to be interesting.