Several studies show that Millennials prefer car access over ownership. The trend is clear that ridesharing services are increasing in popularity. If I could hail a ride and watch a movie or work or eat or sleep, especially if I lived in a urban center where parking is a pain in the butt and expensive, I would try it out for several weeks to see if I like it. I would do it for the convenience, not so much the savings.
The transition won't be immediate or complete. I am sure there will still be car ownership. It's just that each robotaxi can replace the utility of several personal vehicles and so the global vehicle fleet won't be nearly as numerous. Therefore the phase out of ICE vehicles will be more rapid.
So many different environments around the US and the rest of the world.
As you noted, I can see robotaxis being a success and working very well for places like urban centers. I can easily imagine robotaxis taking over for the work of taxis/ubers/lyfts, so they make sense in areas where those services already have a MAJOR presence and profitability.
Also, it feels like you have to either limit users or have a massive number of cars in that city when the service first turns on. If a customer turns on the app, and no robotaxis are available, they'll just shift to another app. So, it feels like if/when Tesla starts producing robotaxis, they might have to ship them in big numbers to one city to get started there. Then another wave to another city later.
In very walkable cities, or areas with great mass transit systems, robotaxis would be less busy and less useful and less profitable.
In many typical suburbs, the entire layout of homes/work-places/shopping places/schools/etc. really makes life VERY inconvenient if you don't own and use your own car. Without a major change in society, a taxi network (robo- or otherwise) will only work for a small percentage of trips in these areas. All the moving-about of people is so concentrated into the morning and evening commutes that a robotaxi couldn't really serve two different people's commutes. And, the taxi wouldn't have much to do in the mid-day or night-time hours as most people are either at-work or sleeping. I also don't imagine many of these folks are going to want to rely on a taxi service of any sort to get their kids to/from activities, or to haul a week's worth of groceries either. A personally-owned robot chauffer makes good sense...but I just don't see economic viability in a full robotaxi (or any other kind of taxi) service in places like this.
And, in more rural areas where everything is more spread out and remote, I don't see robotaxis having much use at all.
So, overall, it feels like the focus on robotaxis that can be massively profitable as a service will limit the market to certain big city areas. Cars sold as personally-owned vehicles will still be effectively required in the suburbs and rural areas, whether or not they can drive themselves. Even though, over its life, a robotaxi in an optimal area might be more profitable, I certainly hope that Tesla doesn't leave out huge swaths of the population for whom a taxi service just doesn't make sense. I want them to push for massive scale as proposed in Master Plan Part 3!