Cruise’s Kyle Vogt: GM Will Deploy Automated Rideshare Cars ‘Very Quickly’ – Forbes
When General Motors bought Cruise Automation for $581 million one year ago it was a head-turning deal that brought together a massive, traditional Midwestern carmaker with a tiny San Francisco startup. This quirky pairing may be working as Cruise CEO Kyle Vogt says automated GM vehicles controlled by Cruise technology are going to be deployed “very quickly.”
Vogt, who began dabbling in self-driving car technology as a student at MIT in the early 2000s and has been called a “robot guru,” has kept a low public profile the past few months amid nonstop announcements by others in the driverless car space. But in a recent conversation with Forbes he gave an update on Cruise’s relationship with GM, progress on their automated vehicle program and the rush of competitive activity.
“We started with about 46 people at the time that we announced the acquisition, a fairly small group. Now we’re up close to 150,” Vogt said. “We had one or two Nissan Leaf (test) vehicles and now we have 50 Chevrolet Bolts that are driving around three states, California, Arizona and Michigan.”
Those all-electric Bolts, equipped with a 360-degree vision system comprised of laser LiDAR, cameras and radar, software and an operating system developed by Cruise, and a cloud-connected communications system Vogt declined to discuss, seem to have gained remarkable autonomous driving skills.
Recent test drive videos show Cruise’s Bolts having little trouble navigating complex San Francisco road conditions, easily maneuvering around stopped vehicles, avoiding pedestrians and making turns at busy intersections without apparent human assistance.
Vogt wouldn’t confirm a Reuters report that “thousands” of self-driving Chevrolet Bolt hatchbacks will go into service for ride-hailing company Lyft (partially owned by GM) in 2018, but such a move wouldn’t be surprising.
“We’ve had a plan in place for a while and it’s going according to schedule. From what I can tell it’s much faster and going to happen much sooner than most people in the industry think,” Vogt said. “We’re planning to deploy in a rideshare environment, and very quickly.”
Still, he provided no specifics on the timing of that deployment and where it would start. GM in December announced plans to build an expanded fleet of self-driving Bolts at a plant in Michigan, where it’s also testing those vehicles on snowy roadways, along with Cruise’s urban assessments.
Numerous automakers have announced plans to begin different types of autonomous vehicle programs starting in 2020 and 2021. Along with the need to perfect artificial intelligence systems and drive down the cost of critical hardware, the timetable will be determined by federal and state regulations, which remain a work in progress.
Surprisingly, GM has taken something of a hands-off approach to Cruise, and that strategy may be working.
“Rather than integrating it fully into the bureaucracy they seem to be letting it do its thing, to operate without as much of the bureaucratic red tape that a big automaker might suggest,” Jeremy Carlson, a Los Angeles-based automotive technology analyst with IHS Markit, told Forbes. “That was a pretty smart move and one that I think we’re going to see from other automakers as well.”
Vogt spoke prior to Intel’s announcement of a plan to buy Mobileye, an Israeli computer vision company, for $15.3 billion, the biggest deal to date in automated car technology. The motivation for that move is massive revenue potential created by automated and connected cars, and related data services. Intel estimates that may be worth $70 billion a year by 2030.
In the wake of the Cruise acquisition, Uber bought Otto and its self-driving truck team for an estimated $680 million and Ford committed $1 billion over the next five years to Argo AI, a startup tasked with creating software and algorithms needed for its autonomous vehicle program. Google’s Self-Driving Car project, which since 2009 has been the best-funded autonomous tech research effort, in December spun off as Waymo, a standalone Alphabet Inc. company tasked with commercializing its technology.
Elon Musk, never one to be overshadowed, announced in October that Tesla would lead in automated driving technology by pre-loading its premium electric cars with the hardware needed to drive autonomously. Though the AI and algorithms needed to make that happen aren’t ready, they’re to trickle out over time via wireless updates.
Along with major auto parts suppliers such as Delphi and Bosch, seemingly every month a new company moves into the driverless car space, including China’s Baidu and Chuxing Didi, Zoox, nuTonomy, Nuro.ai, Drive.ai, Chinese-Silicon Valley startup NIO and Los Angeles-based Faraday Future.
Though he’s only been part of a global automaker for a year, Vogt understands an essential fact: At every level, the car business is very, very hard.
“We were caught of guard by the scale of complexity and the challenge of bringing out a vehicle,” he said. “Not just running it through a production assembly plant, but the extensive amount of safety and validation a company like General Motors goes through before a car is ever released out on the road to paying customers.”
As a result, he’s skeptical many new autonomous car startups that have popped up over the past year will make it to commercialization.
“Perhaps the pendulum has swung too far in terms of investment in this space right now,” he said. “Probably in the next few years we’ll see things shake out a little and see where the true capabilities lie and where the follow-on actions have been made that won’t be as successful.”
He wouldn’t discuss which individual companies might now survive, but not surprisingly says those with connections to major automotive resources are best positioned for long-term success.
“It’s become crowded. There are many people in this space but there are few or none that have what GM has managed to assemble here, which is both a highly capable technical team that’s been working on this and has deep industry experience, plus the actual ability to deploy this at scale in a safe and reliable fashion. I think we’re in a pretty good place.”