The Race to Build Self-Driving Trucks Has Four Horses and Three Jockeys

Over the last five years, as it’s become clear that self-driving cars will take longer than expected to arrive on most American streets, some of the biggest players in the industry have turned their attention to long-haul trucking. Alphabet Inc.’s autonomous vehicle unit Waymo, the leader in robo-taxis, launched its trucking division Via in 2017 after a handful of start-ups, including Otto, Starsky Robotics, and TuSimple Holdings Inc., had entered the field. Aurora Innovation Inc., one of Waymo’s most formidable competitors, recently decided to focus its efforts on bringing a trucking product to market before branching into ride-hailing. The logic behind the pivot is twofold: highways are easier to navigate than city streets and cargo is less demanding than human passengers. If a robo-truck drives extra cautiously on its way to a big box store, as Aurora co-founder and CEO Chris Urmson put it when I spoke to him earlier this year, “the roll of toilet paper doesn’t care.”

There are also advantages on the business side. In ride hailing, robo-taxis need to outperform and undercut a large, flexible and relatively cheap pool of gig workers. Truck drivers, on the other hand, are in short supply, a problem that only promises to get worse as e-commerce continues to boom. In one possible version of the autonomous future — a model known as “depot-to-depot”—robot drivers would cover the long and relatively simple stretches of interstate driving and leave the trickier surface streets to human drivers who would take over at highway off ramps. The robots could operate for hours on end without running afoul of service time rules and needing to stop only for fuel, while truckers would still have jobs and could sleep in their own beds at night. TuSimple CEO Cheng Lu estimates that driverless trucks running more hours on the road can reduce hauling costs by 50%.