Yes the driver is a big cost factor but so is the vehicle. If you can eliminate the driver and produce the vehicle and parts you will be able to produce and repair cars at rates lower than anyone without the same manufacturing capability.
This only makes sense if Uber driver's net income is pure profit. Instead I hear they have a set of expenses - fuel, insurance, tires, maintenance as well as servicing payments on the car. Some of these Uber is likely to negotiate down, mainly insurance. But some (cleaning, maintenance, parking, car washes, tires) are currently externalised to the driver and will have to be assumed by the company.
This is even before we get into the details of who manufactures the cars and how Uber comes into the possession of the vehicle (lease? outright purchase? purchase backed by a loan? etc.)
In what way? Will they manufacture their own cars? That seems expensive, and from the past experience of GM and Chrysler Fiat not exactly a money-printing machine. Will they partner with manufacturers and buy their cars to add to Uber fleet? That will be a new expense that was externalised to the driver before, but now affects the company's bottom line. Will they just become a dispatch software licensee like Flywheel? Enterprise buyers tend to squeeze margins and negotiate tough.
What's the scenario where either their top line increases in a massive way or expenses are trimmed way, way down from where they are today?