California officials Tuesday released annual statistics showing the amount of driverless car testing in the state between December 2016 and November 2017. Just like last year, the numbers show Waymo, Alphabet’s driverless car division, comfortably in the lead. But unlike last year, Waymo now has a serious challenger: Cruise, the driverless car division of General Motors.
Just like last year, Waymo’s driverless cars logged hundreds of thousands of miles on California roads. And both this year and last, its cars were able to go more than 5,000 miles between “disengagements”—situations where a safety driver has to take the wheel.
But last year, Waymo didn’t have any serious competition. Cruise only logged 10,000 miles in 2016, and a safety driver had to take the wheel every 35 miles, on average. Other companies did even worse. This year, things are different. Cruise vehicles drove a respectable 131,000 miles. And safety drivers only had to intervene once every 1,250 miles.
Cruise clearly still has ground it needs to make up. But the company is improving fast, and it’s easy to imagine Cruise closing the gap entirely in the next year or two. Meanwhile, if anyone else is going to pose a serious threat to these two leaders, it’s not apparent in the California data. No one else logged more than 10,000 miles on the state’s roads in 2017, nor did any other company report a disengagement rate better than once every 160 miles.
The new data reinforces the view of other industry analysts who see GM and Waymo as the industry leaders. But it also suggests that GM is rapidly narrowing the technology lead Waymo has enjoyed for years. And it suggests that these two companies could leave everyone else in the dust.
Cruise is catching up to Waymo in California
It’s important not to read too much into the California data. The number of disengagements is a self-reported number, so there’s always a danger of companies under-reporting to make themselves look better. And these numbers only cover California. We know that lots of companies—including Waymo—are testing heavily outside the state.
With all that said, the reports are one of the few places where we can get direct head-to-head comparisons between various driverless car companies. And the gaps these reports show are so stark that they almost certainly reflect real differences in company accomplishments.
In its new report, Waymo said its cars go around 5,600 miles between disengagements. To put that into perspective, if Waymo gave one of its cars to someone with a 10-mile commute, that person would only need to take over the wheel about once a year, on average. That’s impressive, but it’s not much of an improvement from last year’s report, when Waymo averaged around 5,100 miles per disengagement.
By contrast, Cruise appears to be making rapid progress. This is apparent in its annual numbers, where it went from 35 miles per disengagement in last year’s report to 1,250 miles per disengagement this year. And it’s also clear if you look at the monthly statistics provided in this month’s report.
At the start of the year, Cruise was way behind Waymo—both in terms of miles tested and its disengagement rate. But Cruise has been gaining fast. For the last three months of the reporting period—September, October, and November—Cruise vehicles were going 5,200 miles between disengagements, while Waymo vehicles were traveling 8,000 miles per disengagement. Waymo was still ahead—but not by very much.
And Cruise CEO Kyle Vogt believes that Cruise is well-positioned to gain more ground on Waymo in the coming months.
For one thing, Vogt has argued that Cruise gets a lot more value per mile of testing than Waymo. Cruise has focused on testing its vehicles in San Francisco, a dense urban area full of construction cones, emergency vehicles, and unpredictable pedestrians and cyclists. Waymo, by contrast, has done much of its testing in suburban Silicon Valley and the even more suburban Phoenix metropolitan area.
Vogt argues that cars on suburban streets just don’t encounter interesting situations nearly as often as they would in a big city—which means they learn less from each mile driven. Which could mean that Waymo’s 4 million miles of testing might not give it much of a real-world advantage over Cruise—which has logged fewer miles but focused on a particularly challenging environment.
Being owned by GM could be a big advantage for Cruise
Cruise’s other potential secret weapon is its relationship with GM. Once a company like Waymo or Cruise has designed the software and sensors necessary for a driverless car, the next step is to bake that technology into hundreds, thousands, and eventually millions of vehicles.
Waymo recognized early on that it wasn’t going to be able to build its own cars, so it sought partnerships with automakers. After a possible deal with Ford fell through in 2016, Waymo signed a series of deals with Fiat Chrysler to modify Chrysler Pacifica minivans for Waymo’s use. Waymo ordered 100 vehicles from Chrysler in 2016, 500 vehicles last year, and “thousands” of vehicles earlier this week.
In a September blog post, Vogt argued—without ever mentioning Waymo by name—that this kind of partnership was unlikely to work very well.
“Hand-building a few hundred complex cars is tortuous and expensive, but it’s technically possible,” Vogt wrote. “But things start to fall apart beyond that.”
“The wiring harness in our new car has 4,085 wires and 1,066 connectors, meaning that without rigor and process, vehicles will spend most of their time in the shop while technicians chase ‘ghosts’ (chafed wires, loose connectors, or anything else that causes things to suddenly stop working).”
“This kind of development requires full alignment, teamwork, rapid communication, and compromise between dozens of teams,” Vogt argued. “Since Cruise is a fully owned subsidiary of GM, we’ve avoided the seemingly impenetrable wall of politics, competitive fears, NDAs, liability concerns, data ownership, asset financing, long-term alignment, and a host of other issues that plague other relationships and partnership models.”
Vogt doesn’t mention Waymo in his post. But it’s not hard to guess who he had in mind as he wrote those words.
To be sure, Waymo still seems to be the industry leader. The company spent 2017 working out the details of a commercial driverless car service in the Phoenix area—a service it has hinted will be launched in the coming months. It has also begun expanding testing to new cities—including Detroit and Atlanta—where Waymo will presumably offer commercial services in the next few years. Cruise still has a lot of work to do to catch up and surpass Waymo.
But this week’s disengagement report suggests that Cruise has been working hard—and that hard work is starting to produce measurable results.
If Waymo and Cruise have serious rivals outside California, they aren’t tipping their hands
An important caveat here is that not every driverless car company is competing in California. Tesla and Ford both filed reports stating that they’re not testing driverless cars in the state, but we know both companies are working on fully driverless technology. Ford’s efforts are likely centered in Detroit as well as Pittsburgh—where Ford has invested in a startup called Argo.ai.
Uber has invested heavily in driverless cars and has done extensive testing in Pittsburgh and the Phoenix area. It announced a contract to buy 24,000 cars from Volvo in 2019—a signal it was serious about outfitting the vehicles with driverless sensors and software. But Uber’s efforts have been bogged down by managerial chaos and an ongoing legal battle with Waymo. Experts we’ve talked to believe they’re far behind Waymo.
German car companies—Daimler, Volkswagen, and BMW—are working on their own driverless car technology and are presumably doing a lot of their testing outside the United States. So the fact that they aren’t testing cars in California doesn’t mean all that much.
Still, if any of these companies are making rapid progress toward driverless cars, they’re being very quiet about it. When I talked to people in Phoenix, they said most of the driverless cars they saw belonged to Waymo or Uber. When I got a test drive from Volkswagen’s Audi last summer, the company mostly talked about its plans to add advanced driver-assistance technology to its vehicles—not on building the kind of fully self-driving vehicles that Waymo and Cruise are working on.
So right now, it looks like Waymo and GM’s Cruise are the industry leaders. We’ll have to wait to see if a surprise challenger emerges over the next year or two.