I'm assuming at least part of the thinking is eventually they'll be more available and cheaper than human rideshare drivers? I don't think that is currently the case as they're still new and high-tech and also getting a bit of the branded premium. And I don't know if it'll come out this way, but I can see a future where just like with so many other automations, they end up cheaper than alternatives.
What I'm thinking is - technically (even though I bet Uber/Lyft screw over the drivers) you need ~ 3 human drivers for 24 hour coverage(or figure for day only coverage across 2 days or whatever timeframe, I think this all holds true) with no "backup". You probably need 4-5 in reality to cover sick/vacation time etc. (While the drivers for rideshare aren't employees, they still as independent contractors need to cover their own potential sick time or time off for vacation or whatever, so at least implicitly they're including that in their pricing they're willing to accept.) With current independent contractors that means you're also implicitly covering the cost for 4-5 cars as well. Things like Waymo need 2-3 cars for the same coverage, to enable charging/fueling, repairs, and cleaning. But it's still 1-2 less cars for the same coverage.
So that's less capital to amortize. They also have no labor cost and no need for tips from consumers. (For the driving, probably still need technicians etc, but it'll be like any fleet mechanics or bus cleaners etc - way smaller numbers across a bunch of cars vs paying each owner/operator to do all the tasks on a 1 to 1 basis).
In the long term, I could absolutely see it becoming like many consumer goods where the more automated versions are rather cheap and the handmade versions are very expensive. I guess sort of like paying at a laundarymat and paying a maid to do your laundry or whatver.