[chipotle_coyote]

As valuearb has pointed out in other comments, it's not at all clear that Uber needs to be in the self-driving car game to the point of making self-driving car technology, any more than they need to be in the human-driven taxi game they're in now to the point of buying an automobile manufacturer.

It's not at all unreasonable to think that the way to save Uber is to ruthlessly cut costs and focus on the business of fleet and driver management, without banking on the assumption that the drivers are all going to go away within a decade. They should certainly still be watching autonomous cars, and perhaps find someone to partner with the way Lyft has with Google/Waymo, but I'm not at all convinced that's a business they should actually be in.

(I also suspect that the belief that we're going to have "level five" autonomous cars within the decade is far too optimistic. Also, an awful lot of discussions apparently assume the vast majority of consumers will, as soon as it becomes practical, stop buying cars and switch to autonomous taxis for all transportation needs. At least among American consumers, I'm going to stamp that with "[citation needed]" until further notice.)

[xyzzy_plugh]

Why would anyone who is working on their own products partner with Uber (or Lyft etc.)? It's pretty clear that a genuine self-driving product would demolish the human driving equivalent. Forget network effect, it's going to be way cheaper.

[ghaff]

Because, if they're still in business long-term, 1.) they're potentially a very big customer and 2.) They can be a great source of data and broader insights about exactly how people use/want to use taxis in different locations. What are the patterns? Pickup and dropoff points. Etc.