Aurora’s Plan to Catch Waymo in Robot Trucks, Then Taxis
On Feb. 11, about a month after closing on the acquisition of Uber’s Advanced Technologies Group, or ATG, the self-driving-car startup Aurora held an all-hands meeting. Like every other company meeting since the pandemic began, it took place via videoconference, and included the simulacra of in-person office interactions. When managers handed out awards, there was no applause from co-workers. Instead, Aurora co-founder and Chief Executive Officer Chris Urmson hit a button that produced canned cheers and clapping. “I am so looking forward to not having that button,” Urmson said after setting it off for the third time.
The weeks leading up to the meeting had been full of upheaval at Aurora Innovation Inc. Uber Technologies Inc. had essentially paid it to take ATG, forking over $400 million for a stake in the combined enterprise, which was valued at $10 billion. The deal allows Uber to unload a unit that was hemorrhaging cash while keeping a foothold in autonomous vehicles. Aurora, in return, adds almost 1,000 employees, more than doubling its workforce to 1,600 and bolstering its bid to become a credible competitor to Alphabet Inc.’s Waymo. The deal also gives Aurora what could be a bigger prize: the right to provide robo-taxis to Uber’s ride-hailing network.
“Having a strategic relationship with Uber is an incredible advantage,” Urmson says. Over the past two decades he’s done more than perhaps anyone to push the development of autonomous-driving technology. Now he’s in a leading position to be the first to truly commercialize it. But with the addition of hundreds of highly paid engineers and a large pool of potential customers, Urmson is under more pressure than ever to bring a product to market.
Since co-founding the company in January 2017—with former Tesla engineer Sterling Anderson and Drew Bagnell, who came from ATG—Urmson has been lining up deals to ensure that buyers will be waiting when his robot drivers are ready. The plan is to begin with long-haul trucking. Earlier this year, Paccar Inc.and Volvo Group signed agreements to install Aurora’s automated driving system in their trucks. The two companies would then offer these trucks, capable of operating themselves for long stretches, to their shipping customers, who would pay Aurora for the hours of automated driving.
After establishing itself in trucking, Aurora would begin cherry-picking the easiest, most lucrative trips from Uber’s ride-hailing network. A customer looking to go 25 miles, mostly by highway in light traffic, might be greeted by a driverless car. Aurora already has a deal with Toyota Motor Corp. to build robo-taxi fleets.
A 2019 investment from Amazon.com Inc. sets the company up for a similar strategy in delivery, allowing it to service the easiest routes for e-commerce customers. If all goes right, its robot drivers would take over entire fleets of cars, trucks, and vans.
Urmson says Aurora’s driving system can become better than the average trucker in a matter of years, not decades. He has a target in mind for when the first trucking product will be ready, though he’s not yet willing to share it. He knows as well as anybody how the work of building autonomous vehicles can expand endlessly. At 44, Urmson has been working on self-driving cars for most of his adult life, first as a 27-year-old graduate student at Carnegie Mellon University, where he led teams that competed in three Defense Advanced Research Projects Agency (Darpa) challenges, winning the final race in 2007. He then spent eight years with Google’s self-driving-car project, now known as Waymo.
Urmson left in 2016, shortly after Google passed him over for the CEO job and instead hired former Hyundai Motor Co. executive John Krafcik. “I’d been leading and building that team and, for all intents and purposes, general managing it for years,” he says. “Of course I wanted to run the program.” (Krafcik announced in April that he was leaving Waymo.)
Aurora was born from meetings Urmson had with Reid Hoffman, the LinkedIn co-founder and venture investor, and Mike Volpi of Index Ventures. The three thought that the automotive industry wouldn’t tolerate a Google monopoly in self-driving technology and that Urmson was the one to develop a viable alternative. “Chris had already gone through and solved the problem once,” Hoffman says, “and now knew which things he would rebuild entirely differently.”
In 2018, Hoffman’s firm Greylock Partners and Volpi’s Index co-led a $90 million fundraising round in Aurora, which has now raised more than $1 billion. It’s likely to take even more before Aurora will be able to support itself. “I don’t understand how Aurora is going to be able to survive without raising more capital,” says Grayson Brulte, co-founder of Brulte & Co., a consulting firm focused on transportation. The acquisition of ATG and its hundreds of engineers adds to that burden. “It’s like taking two money pits and making a bigger one,” says one former Aurora manager who asked to remain anonymous when talking about his past employer.
Urmson doesn’t shy away from the possibility the company may need to raise more money, and he’s confident it would be able to do so. He says the ATG merger is already showing tangible benefits. Before the deal, Aurora spent hours calibrating the sensors on each of its vehicles, with workers walking around holding placards at set distances to be sure the sensors measured accurately. ATG devised a faster, automated process using large turntables to rotate the vehicles. On the February call, a former ATG engineer said the turntables would soon be in place at all of Aurora’s vehicle depots.Electric vehicles, self-driving cars, and whatever else Elon Musk is up to this weekGet Bloomberg.com’s Hyperdrive newsletter.
Aurora will need every possible advantage in the race to bring autonomous vehicles to market. The leading competitor is Urmson’s old shop, Waymo, which, in addition to launching the first fully driverless ride-hailing service in suburban Phoenix last year, has a passenger-sharing agreement with Uber’s top rival, Lyft Inc. A handful of startups that have joined forces with auto manufacturers, including Cruise with General Motors Co. and Argo AI with Ford Motor Co., also promise to be formidable.
Urmson sees Aurora’s singular focus as its advantage. He doesn’t have to worry about quarterly earnings in other parts of the company or trying to keep the Detroit office at arm’s length. “This is the thing we do,” he says. “This is what we’re going to be excellent at.”
During the February meeting an employee asked if TuSimple, a San Diego-based autonomous-trucking startup that raised more than $1 billion in an initial public officer this week, might beat Aurora to that market. “We do think that we are going to be first or at least meaningfully first,” Urmson replied. He defines “meaningfully first” as a commercial product that can be scaled safely.
Urmson also believes Aurora’s lidar—the laser sensors that most autonomous vehicles use to perceive the world around them—is superior, giving it an edge. In 2019 the company bought the Montana startup Blackmore Sensors & Analytics, which makes a type of lidar called frequency-modulated, continuous-wave, or FMCW, that allows the sensors to simultaneously measure range and velocity, which vastly simplifies the problem of predicting the movements of faraway objects. “When you’re driving a truck at 65 miles an hour on the freeway and you want to react to the rare events, you need to see 300 meters down the road,” says Urmson. He says no one else in the industry can match that performance.
TuSimple and Waymo—which started its trucking division, Via, in 2017—also promote the superiority of their sensors. A wild card is Tesla Inc., whose CEO, Elon Musk, sees the lidar arms race as a never-ending fool’s errand. For the past six years the automaker has been using its millions of electric-vehicle customers as test subjects in a large-scale public road experiment in camera-based autonomous systems, offering an “autopilot” that allows drivers to relinquish control to the car. Musk has said Tesla owners will eventually be able to use their vehicles as robo-taxis during downtime. Urmson is skeptical. “It’s just not going to happen,” he says. “It’s technically very impressive what they’ve done, but we were doing better in 2010.”
In 2015, Urmson gave a TED Talk in Vancouver called “How a Driverless Car Sees the Road,” explaining how Google’s software was learning to recognize cars, buses, cyclists, and pedestrians and anticipate their movements. At the end he displayed a photo of his two sons. “My oldest son is 11,” he said, “and that means in four and a half years he’s going to be able to get his driver’s license. My team and I are committed to making sure that doesn’t happen.” His son, now 17, still doesn’t have his license, but he does have a learner’s permit. Over the past few months, while trying to teach trucks to drive themselves, Urmson has also spent a few hours teaching him to drive. “He’s working towards it,” he says. “I’m not actively sabotaging him.”