Waymo, largely considered the leader in the self-driving space, announced it was pausing its trucking efforts, known as Waymo Via. They had previously had layoffs in this division earlier this year. At the same time, Aurora, which is run by former Waymo engineering head Chris Urmson, announced it had just raised $820M — $600M in a private placement — and was doubling down on its trucking division, which has long been its major focus. Aurora, which has no revenues and had lost much value since going public by SPAC previously, had triggered concerns about how much runway it had left before the raise.
Recently we’ve also seen a shutdown of U.S. operations by trucking startup tuSimple and the closure of Embark, another player in the space. Other players like Peloton and Locamotion in the platooning space also failed. Plus.AI is still operating but has not raised money since April 2021. Kodiak recently received military contracts to keep it afloat. Non long-haul companies Gatik and Einride appear to still be going well.
Waymo, as a unit of Alphabet, has access to vast resources, but the patience there is not infinite, particular with Larry Page, who created the project, no longer involved day to day. Many of those who have entered long haul trucking have done so because its commercial case is among the clearest in the field — the economics of trucking and the cost and issues with drivers are well understood. At the same time, while 2,000 die every year in truck accidents, trucking is seen mostly as an economic play — save money and make shipping more reliable — rather than also providing huge value through safety and by saving amateur drivers billions of hours of time.
Trucking has always presented a variety of good and bad trade-offs. Freeway driving is vastly simpler than urban driving — the roads are simple and gentle and generally well maintained. There are no intersections or oncoming traffic, and no vulnerable road users like pedestrians are allowed. It’s a much easier problem to solve. At the same time, the speeds are fast and dangerous, and when you start working with class 8 semi-trucks, the kinetic energy is enormous. Even a small mistake, like the sort that have caused minor fender-benders for Cruise and Waymo on city streets, could be a disaster. Even though such a mistake is much less likely, perfection is still not possible. You would rather your first crashes happen in small cars, not transport trucks.
This is also complicated by the employment question. The reality in trucking is that there are over 100,000 open job positions, and turnover in the job is over 100% some years — robots were not going to take anybody’s job for many, many years to come. Even so, people worry about it, because unlike cab driver, truck driver is a career for some people. They are also represented in some cases by the Teamsters, who are now actively working against self-driving tech and recently pushed a law in California that could forbid unmanned self-driving trucks in the state for some time.
This is one reason we’ve seen companies like Gatik and Einride focus on other aspects of the game, like middle-mile warehouse deliveries with smaller trucks in the former case, and electrified trucking in the latter.
Aurora might be annoyed by this announcement. With their top competitor pulling out, that’s good news if they can make it work, and might have pushed their valuation. Of course it might also have scared investors in general concerning the space. There are so many routes that companies were not likely to see competition for many years, even if they all went gangbusters.
While Waymo is not literally “doubling down” as they suggest. They will probably mostly conserve funds rather than double the budget of Waymo One. They do say capital saved will go into Waymo one. They do have a planned expansion to Los Angeles to do fairly soon, though not date is released. They are hoping to get their expansion in the Bay Area soon if they can get past the CPUC’s issues on emergency encounters, though those mostly involve Cruise. Waymo has been invited to comment on any budget increase for Waymo One.
Currently, Waymo One charges about $2/mile, similar to Uber, in Phoenix. To simply duplicate Uber is not the plan, though, but it is worth noting that long haul trucks don’t charge much more than this — about $1.80 to $2.40 for van service able to do 11 hours/day. They charge more if there are two drivers that can go 22 hours/day. A robot can do almost 24 hours/day, needing only to refuel, once they get that working. But running a truck is vastly more expensive per mile than a taxi, though you get 3-4 times as many miles/hour. Even so, the taxi can be a more lucrative service. (Trucking used to be a fair bit less in the 2010s.)
Of course, just as Aurora also is working on cars — in the background — Waymo’s technology really isn’t that difference between a car and a truck. All the work being done to get their taxis to work on freeways — which they still don’t do in public service, but will do for employees — will apply to any return to trucks in the future. They only say they have “pushed back the timeline” but they have also cut back almost all technical development for that unit, and presumably the business development and operations as well. For now they will literally “laser focus” on ride hailing. Truck drivers make about $30/hour when all is done, but the people riding in taxis usually gain back time worth more than that.