To operate self-driving cars, they'd need to buy or finance the vehicles, which is a huge expenditure. They'd also need to hire people to clean and maintain them, since drivers wouldn't be doing that anymore. They'd have to bear the cost of fuel (whether gas or electricity) and insurance, which drivers currently pay for. Once they pay for all that, would it cost more or less than paying a driver who provides their own car? I certainly don't know.

Also, even if all the factors noted above balance in favor of self-driving cars in the long term, it may be a while before self-driving cars come down enough in price to compete with human drivers in conventional vehicles - it takes significant engineering work to evolve a design from a prototype to mass production.

The question is whether Uber can stay in business long enough to reap the benefits of self-driving cars. I don't think the answer is obvious.

Also, city and state regulators are not even close to allowing massive fleets loose on the roadways. Uber will need 5+ years of burn or positive cash flow before they can even begin to rely on autonomous cars and that is only in easy to navigate dense cities of the world.

If they raise prices significantly, someone else will come at a lower price with lower service quality (as already happens in Asia) and take a share of more price sensitive segment of the market.