A MobilEye robotaxi prototype, branded with Moovit, which will deploy in Munich by 2022, Intel says.
MobilEye
At the IAA Mobility conference in Munich Intel
INTC
UBER
This joins a variety of robotaxi pilot projects in the USA, China and now Europe and Israel. Details are few, including when the rides will begin with no safety driver in the vehicle — a key stepping stone which was recently made legal in Germany once a process is complete. There is earlier coverage here on Forbes.com but it’s worth discussing the bigger trends behind this announcement.
Not long ago, there were broad claims that we were in a robocar winter — that the technology was still a decade away, if not more, and car OEMs were all pulling back. Most of them did pull back, but the companies dedicated to Robotaxi, like Waymo, Cruise, Zoox, MobilEye, AutoX, Baidu, WeRide, Motional, Argo, Aurora and others, as well as the trucking firms, barely noticed any winter.
Right now, companies keep their true performance close to the vest — except Waymo which has been willing to publish real statistics on their pilot operations, though only some of them. As such, what matters, and shows how far along they think they are, is what they are willing to do that’s risky. This includes steps like offering service to the general public, and of course operating with no safety driver on board. To a lesser extent, it includes announcing firm dates for when you will do this, but dates slip notoriously. (Indeed, nobody believes dates from Elon Musk any more at all.)
Also important to factor in are how difficult the driving territory is, and if hard parts have been excluded. Waymo’s launch in San Francisco will exclude downtown; their first pilot in Chandler is about the easiest place to drive there is. AutoX keeps to a fairly calm suburb of Shenzhen, though it is more complex than Chandler to be sure. Baidu uses a very small area and keeps safety drivers on board. We don’t yet know what territory MobilEye will serve, though almost all of Tel Aviv is challenging to drive.
Germany is an odd first choice. While they have passed a law that will allow this, it generally a place with more bureaucracy than the other contenders. It is less clear how quickly the safety drivers will be removed, even if it is legal. The taxi lobbies in Germany have had success against Uber and will not be enthused with this project either.
Car Replacement
Operating an Uber-like service is a worthwhile milestone, and offering it to the general public is particularly important. Once it’s offered to the general public is a sign the team feels much more ready, and is willing to give up more control. Members of the public will see your warts, and video them and put them up on YouTube. Waymo got a lot of derision for this embarrassing performance in a construction zone but the reality is, we saw this because they were letting the public ride. Other teams that keep things more under wraps are almost surely doing worse.
Removing safety drivers in theory is a minor step, because by the time you do it, you need a safety record so good that they are not doing anything any more. Removing them shows you’ve reached that point, and reached it a while ago. In some cases, team leave cars under constant remote monitoring (ie. with one remote operator per vehicle.) A better milestone happens when each remote operator is handling many vehicles.
But an Uber-style ride-hail service is also just a stepping stone. The robotaxi revolution doesn’t happen when you can make a modestly cheaper Uber. (Some even wonder how viable a business Uber really is, given the amount of investor subsidy it has needed.)
To make a working robotaxi service, you have to reach the level of car replacement. That people decide to give up car ownership because of your service. They don’t need to use your service exclusively. The car replacement will include regular human driven taxis, daily rent-a-car, transit, e-bikes, minimobility and other tools. And the first car replacement will be taking 3 car families down to 2, and 2 down to 1 which are much easier than getting somebody to zero. Still, it’s only when the robotaxi service covers a large enough area at a good enough price that somebody decides to sell a car that we move into the world-changing department. Nobody is yet there, though Europe might be the sort of place that can happen sooner.
The use of Sixt and Moovit is unclear here. While Moovit as an app is one people use to make trips, it’s not obvious that people using it will be the ones to suddenly want to summon a robotaxi with it. It’s more for people in the transit market with a focus on low cost at present. It is not clear that early robotaxi companies gain a lot by partnering with existing transportation apps, even Uber or Lyft. While Uber may be the most popular app people pull up to find a taxi-like ride, dedicated apps for novel new services like this will work fine for now. Only in the future, when a proper car replacement service is ready, will people quest for a multi-use app. Sixt is known only for car rental. They will do fine at maintaining the fleet and operating it, but their value to the customer is less strong. It’s really Intel/MobilEye who are the key provider in this game.
