This week, Lyft announced that it is going to be selling its autonomous vehicle division to Toyota for some $550 million. (Apparently $200 million of this will be paid upfront, with the remaining $350 million paid out over a five year period.) This is notable because Uber did the exact same thing last year when it sold its autonomous vehicle business to Aurora (which happens to be working with Toyota), and because the reasons for selling seem clear: getting to full autonomy is going to cost a bunch more money and both Uber and Lyft are determined to reach profitability sooner rather than later.
The other thing that you might be able to glean from these announcements is that neither company seemingly feels like they need to fully own/control the autonomous piece. Presumably the thinking is that someone else can spend the money on developing full autonomy and they'll just stick to building out their ride-hailing network. Once we have autonomous taxis, they'll need a network to run on anyway, right? I guess. But wouldn't this dramatically undermine the network effects of Uber and Lyft?
If you go back to Uber's S-1, there was a diagram that explained Uber's "liquidity network effect." See above. It starts with more drivers and more supply (1), because more cars driving around means that wait times and fares are lower (2) and so more people are likely to use Uber (3). Network size matters. But if you no longer have drivers -- only autonomous vehicles -- isn't it relatively easy to add more supply to any network? I suppose this partially depends on how the ownership structure will end up working for these autonomous taxis. Still, I wonder about the barriers to entry under this scenario.