The money quote is here:
> But even if Uber can keep up in terms of autonomous vehicle technology—or find more willing partners in the established auto industry—there’s a much larger problem looming ahead.
> The companies that will profit from the transportation-as-a-service revolution will depend on an incredibly costly infrastructure—one that doesn’t yet exist. Someone has to buy and maintain all the autonomous vehicles, enough to replace all of the cars driven (and usually owned) by Uber and Lyft drivers, not to mention many of the cars driven by you, me and countless others, too.
> To date, Uber’s model has been built on it not owning vehicles. Owning an enormous fleet of them, which might be required to make transportation-as-a-service work, is in some ways antithetical to the business model that has given Uber a nearly $70 billion valuation, more than any other startup. That model depends on outsourcing vehicle ownership and maintenance to its drivers.
> Uber does own its prototype self-driving vehicles, but declined to comment on the extent to which its future would depend on owning vehicles.
> Auto manufacturers have a great deal of experience working with the complex web of dealers, financing companies and fleet managers—even car-rental agencies—that could potentially be repurposed to manage millions of self-driving vehicles.