Volkswagen doffs cap to Waymo, hopes to catch up down the road

Google sister-company Waymo is getting used to backhanded complements from rivals, with Volkswagen CEO Herbert Diess latest to dish out the praise. He credited Waymo with a one to two-year lead over the field in autonomous driving, certainly over Volkswagen itself, while pointing out that the future would depend on how regulation pans out. There may have been a twinkle in his eye as he said that given how badly Volkswagen fell afoul of diesel emissions regulation in the scandal that blew up in September 2015, but also how unexpectedly quickly the group has recovered from that.

Just two months earlier the chief executive of Chinese start-up Pony-ai James Peng had been crediting Waymo with leading the pack of US self-driving companies which he said were still well ahead of his own country’s. He said that although China was striving to close the gap Waymo and to a slightly lesser extent General Motors were still ahead on all fronts including technology innovation, data collection and talent cultivation.

This largely equates to one simple metric – miles on the road. Waymo announced in October 2018 that its self-driving cars had driven 10 million miles on public roads unaided, way ahead say of Uber, which clocked 2 million miles before having to freeze public testing after that widely publicized fatality in March 2018. Waymo may also be in the lead in virtual miles travelled in simulation at 7 billion, but that is a less meaningful statistic because that can only go so far and real miles exposed to the unpredictable behavior and conditions of roads with human drivers is needed to refine the models. In fact, the two work in tandem because experience from real road driving is incorporated in the simulations to broaden their scope and make them more realistic.

While Waymo may be well ahead in distance travelled its 10 million miles does not sound much compared with the 3 trillion covered by drivers each year in the US alone. It is widely agreed that more like 100 million will need to be covered to iron out all the wrinkles associated with unexpected situations, or to make the correct decisions on the basis of rules incorporating ethical as well as practical considerations in almost all circumstances. At its current rate Waymo would not pass that mark for 15 years, so there really is still all to play for.

Meanwhile Waymo has a collaborative challenge since it is not an auto manufacturer itself. That is not a problem at this stage of the self-driving journey, involving deployments in the hundreds or low thousands of vehicles, but will be when the time comes to scale up to millions. Waymo has been trying to woo major car makers into longer term contracts leading towards full scaling up of manufacture but so far this has not gone well. The reason is that major car makers face an existential battle to retain their dominant position in the auto business, and fear that under partnership with Waymo they would become mere makers of the commodity components.

The OEMs are quite happy to partner with Waymo on a smaller scale, with Chrysler supplying 62,000 cars and Jaguar 20,000 for example. But longer-term discussions with both Ford and Honda have broken down with the latter instead deciding to collaborate with GM’s self-driving subsidiary Cruise, in which it has invested $750 million. Honda sees Cruise under GM’s ownership being on the same side so that collaboration is less risky while pooling resources for a common cause.

Waymo may end up having to acquire an OEM or at least manufacturing capacity, although an alternative approach might be to go the other way around and offer a major manufacturer equity in the company. That way the risks and future gains would be shared.

Waymo is unlikely though to seek a third way of accepting it is just a provider of a technology stack, which is the model being pursued by Aurora, a startup co-founded by Chris Urmson, former head of Google’s self-driving car program, along with former head of Tesla’s Autopilot program Sterling Anderson and Drew Bagnell who was a senior member of Uber’s self-driving team.

Aurora’s aim is to build a self-driving technology stack for licensing to the OEMs, which is ambitious enough for a start-up but hardly palatable to Waymo. It is appealing to OEMs since they retain control over their autonomous destiny and given Aurora’s pedigree it is not surprising it has secured some big partners including Hyundai and Volkswagen. That then is one avenue Volkswagen is taking to close the gap on Waymo.