This is generally the first thing that everyone thinks in response to seeing Uber's shocking loss numbers and lack of profitability in the model where drivers needs to be subsidized greatly to increase supply in unprofitable areas.
Since the idea of self-driving cars as an Uber savior is so common, it is one of the most deeply analyzed, particularly by Naked Capitalist [0].
The outlook doesn't look good unless some very specific and IMO unlikely conditions are met for _any_ ride hailing service.
Most ride-hailing services can only profit from self-driving technology if they get it to a fully completed and commercially viable final product. Partial solutions that could power robotics at, say, a seaport or warehouse, likely wouldn't work well unless the business pivots away from ride-hailing. The ride-hailing part, though, requires a full end-to-end solution before the losses Uber currently has could ever begin unwinding or leading towards a profit.
But the problem here is that many competitors in the self-driving tech space are perfectly able to profit from shorter term or partial solutions. Waymo and GM in particular. And if they can build from incremental successes, rather than needing the whole shebang before they can use it for profit, it suggests they will have a large advantage when claiming patents, understanding how the technology works for customer needs, legal requirements or licenses, and many other things.
In short, Uber would basically need to beat all of those competitors to self-driving tech, create an unassailable patent war chest from it, _and_ get the commercial implementation all the way to the end state where it is ready for consumers in large-scale taxi situations (including all the issues with governments, safety, and so on)... all _before_ any competitor seriously develops even partial business models on components of overall self-driving systems.
I just think it's such an unrealistic ask. But I do grant that somehow, maybe, if Uber is exceptional enough at self-driving cars, it is a possibility they could become a full on monopoly of the tech and that would be their way to capture the market and have room to charge higher prices.
But essentially _any_ other outcome in which self-driving tech is commoditized, so that either Uber leases it out from someone else like Waymo or GM, or where multiply players all have separate commercially viable systems, then Uber would gain no cost advantage.
Basically, if Lyft, Via, Uber, local taxis, etc., are all just licensing the same self-driving cars, then everybody can eliminate the cost of drivers, and now suddenly you're in the same price war you were in with drivers. The total price of a ride might be lower, but you're still pressured to drive it to the marginal price that the wear and tear on the self-driving vehicle costs, because all the other players can do this too.
So it means Uber's expenses would go down, but their revenue per ride should go down more or less exactly the same, eating away any excess revenue gained from eliminating the need for a driver. Sure, it won't be perfect, but it still immediately begs the question, how can Uber prove this would be profitable? So far, there is no evidence to suggest it would be, unless Uber has an insane monopoly of self-driving for ride-hailing.
Finally, the second thing everyone thinks of after they realize that self-driving doesn't change the fundamental profitability of taxi services if it's commodity, is they think about Uber-based gig economy stuff like Uber Eats. But there again, Naked Capitalism has looked into it, and the results are dire.
The best I could say is that maybe Uber can find a way to use partially completed or partially approved self-driving tech to hammer on their logistics business. But again, any way you slice it, to be profitable, it literally requires a pivot away from general purpose ride-hailing. The business model is intrinsically not profitable for what Uber claims to do. A subsidy to increase ride stock in regions where ride stuck is fundamentally unprofitable is, well, unprofitable. It's just a VC subsidy. Unless they produce an innovation whereby they actually reduce the in-built cost of inconvenient and under served transit relative to all other ride providers, then they don't have a business there, just more losses.
[0] https://medium.com/@michaelhaupt/this-is-such-an-important-r...