Nobody seems to be contesting that Levandowski took a bunch of documents from Google (Waymo). If they want to, they can almost certainly make Uber choose between stopping working on self-driving cars, and firing or transferring Levandowski.
There's also the possibility that Levandowski goes to jail over this (I don't think his trying to plead the Fifth was just gamesmanship).
Bottom line: My best guess is that Uber is able to continue, but they have to throw Levandowski under the bus to do so. I don't know to what degree that will hurt them.
My point is that, IMO, these kind of "IP" cases, never impede an actually successful business. And that if a business fails, and also has IP suits against it, it is for other business reasons that it failed.
I'm no IP lawyer ( that's probably helpful here ), but looking at Oracle vs Google, Samsung v Apple, IIRC, that actually more legitimate business didn't get impeded and just kept going.
My prediction is that if Uber falls it won't be IP suits that deal the fatal blow.
I think this is generally because courts move slower than the world, and both sides are, usually, "essentially matched" since they can both afford solid legal teams.
All this sort of leads to the conclusion that IP has mostly defensive, and capital benefits ( owning IP can deter people entering market or suing, and IP can add to value of transactions ), but it doesn't have the power to prevent other people using your invention ( as I think the intent of patents originally was, exclusive use ).