Call me pessimistic, but "purchase 24,000 of its XC90 SUVs between 2019 and 2021 to form a [Uber] fleet of autonomous vehicles" sounds rather unrealistic and more like an over optimistic marketing story to me. I know, a lot of good engineers are working on it and automatic vision, sensors and AI are improving constantly. Nevertheless, the Uber use case IMHO requires a solution for the most difficult level of autonomous driving - urban traffic. And without any kind of driver/ safety person/ driver awareness. For 2019 I could imagine autonomous Uber on preselected highway routes, but not in inner city during rush hour with hundreds of people chaotically crossing the street during heavy rain.