The business model appears to be to get established and then use a combination of network effects and monopolistic pricing techniques to be profitable. The problem is though, you need to establish yourself as the dominant player, and jack up prices to start being profitable. That's a fine business strategy. The problem is that Uber has an outside clock ticking. The first player to economically produce a fleet of self-driving cars is likely to undercut Uber on price so violently that Uber's business model goes up in smoke pretty quick.

So either Uber wins the race to autonomous vehicles (which I think most people now regard as unlikely) or they have a hard limit on when their profitability is going to go away. So the valuation needs to reflect something along the lines of "How much money can uber make between now and 2025" and every year that goes by without profit, is a year towards the day that uber's business model goes poof!