The $800M Robo Taxi That Could Beat Uber

Published on July 18, 2018 by

Your robot taxi has arrived, kind of.
By Ashlee Vance

The mystery box sits inside an all-white room in an office building in San Francisco. It’s a large, wooden crate with no features other than the word “ZOOX” in big, black block letters and a sturdy-looking padlock. For about $100 million, you can get a key and have a look inside.

Few have had the pleasure. What they saw is a black, carlike robot about the size and shape of a Mini Cooper. Or actually, like the rear halves of two Mini Coopers welded together. The interior has no steering wheel or dashboard, just an open space with two bench seats facing each other. The whole mock-up looks like someone could punch a hole through it. But because you’ve just invested $100 million in the thing, you’ve earned the right to have a seat and enjoy a simulated city tour while you pray that this vision of a driverless future will come to pass.

Of the many self-driving car hopefuls, Zoox Inc. may be the most daring. The company’s robot taxi could be amazing or terrible. It might change the world—not in the contemporary Silicon Valley sense, but in a meaningful sense—or it might be an epic flop. At this point, it’s hard to tell how much of the sales pitch is real. Luckily for the company’s founders, there have been plenty of rich people excited to, as Hunter S. Thompson once put it, buy the ticket and take the ride.

Zoox founders Tim Kentley-Klay and Jesse Levinson say everyone else involved in the race to build a self-driving car is doing it wrong. Instead of retro-fitting existing cars with fancy sensors and smart software, they want to make an autonomous vehicle from the ground up.

The one they’ve built is all-electric. It’s bidirectional so it can cruise into a parking spot traveling one way and cruise out the other. It makes noises to communicate with pedestrians. It has screens on the windows to issue custom welcome messages to passengers. If the founders prove correct, it will be the safest vehicle on the road, having replaced decades of conventions built around drivers with a type of protective cocoon for riders. And, of course, Zoox wants to run its own ride-hailing service.

Both founders sound quite serious as they argue that Zoox is obvious, almost inevitable. The world will eventually move to perfectly engineered robotic vehicles, so why waste time trying to incorporate self-driving technology into yesteryear’s cars? “We are a startup pitted against the biggest companies on the planet,” Kentley-Klay says. “But we believe deeply that what we’re building is the right thing. Creativity and technical elegance will win here.”

Kentley-Klay, it should be clear, is a salesman. “We want to transform our cities in the way that we live, breathe, and work with our families and communities that’s really profound,” he says, by way of explaining the company’s name. (It’s an abbreviation of zooxanthellae, the algae that helps fuel coral reef growth, not a nod to some colorful hallucination from Dr. Seuss.) Levinson, whose father, Arthur, ran Genentech Inc., chairs Apple Inc., and mentored Steve Jobs, comes from Silicon Valley royalty. Together, they’ve raised an impressive pile of venture capital: about $800 million to date, including $500 million in early July at a valuation of $3.2 billion.

Even with all that cash, Zoox will be lucky to make it to 2020, when it expects to put its first vehicles on the road. “It’s a huge bet,” Kentley-Klay concedes. In the next breath, though, he predicts the future for all of his competitors—Alphabet, General Motors, Tesla, Apple, Daimler, et al.—if the bet pays off: “They’re f—ed.”

Kentley-Klay is a 43-year-old native Australian with a linebacker’s physique, a mischievous manner, and a family history of gimme-the-damn-wheel adventurousness. His great-grandmother was the first Australian woman to get a driver’s license. His grandmother, the second Australian woman to get a pilot’s license, taught Kentley-Klay’s father, Peter, to fly during endurance air races between Sydney and London.

Young Tim was a tinkerer. Growing up in Melbourne, he tried to build a space shuttle out of spare parts from washing machines and lawnmowers, crafted a giant fiberglass whale to compete in a soapbox derby, and, until his parents found out, produced and sold fake IDs to schoolmates. In his 20s, he bought a decrepit 1958 Land Rover and turned it into a surfboard carrier he called the General. “It’s still his pride and joy,” says his mother, Robin.

After getting a degree in communication design, Kentley-Klay went into the ad business and became an industry-leading animator and video producer. He made ads for companies including Visa, McDonald’s, and Honda Motor, and his salesmanship improved with his design skills. “Every eight weeks, there was a new script,” he says. “You had to invent a new world with new characters and go through the really tough process of pitching an agency.”

In 2012, Kentley-Klay stumbled on a blog post about Google’s self-driving car project, then pretty much the only one in the field. He saw the company’s prototypes as unsightly half-measures, with their bulbous sensors mounted on some other company’s car like robot taxidermy. He started designing concepts, researching artificial intelligence, and, per the custom of would-be tech visionaries, wrote a manifesto. He also made videos depicting robo-taxied cities of tomorrow. Then, one day, he walked into his Melbourne office and announced he was off to America to fulfill his driverless dreams.

In a move that some will call devious and others will call ingenious, Kentley-Klay reached out to some of the biggest names in the field and told them he was making a documentary on the rise of self-driving cars. The plan was to mine these people for information and feel out potential partners. His first “interviewee” was Sterling Anderson, then a robotics researcher at MIT and later Tesla Inc.’s self-driving car chief. “I played the oldest trick in the director’s book: the vanity card,” Kentley-Klay says. “I showed up at MIT with a Canon and a bullshit microphone and interviewed Sterling for two hours in a grassy field. In my defense, I might have been making a documentary. The jury is still out on whether I am full of shit.”

Eventually, Kentley-Klay ended up in California and in front of Anthony Levandowski, then one of Google’s lead autonomous-vehicle engineers. They hit it off, and Levandowski was impressed enough to invite Kentley-Klay to give a presentation at the Googleplex in June 2013. On the appointed day, Kentley-Klay popped some dextroamphetamine and gave a talk to 20 people. “I said, ‘I’m Tim, and I will be the first to bring autonomous mobility to the world,’ ” he recalls. “It was a stupid f—ing thing to say, and I did not think it went very well.”

But the Google team was taken with Kentley-Klay’s chemistry-aided passion. They didn’t agree with all of his ideas, especially the bit about needing to build a whole new vehicle from scratch, but they were impressed by how much he’d thought things out. Rather remarkably, Google offered this oddball Australian nonengineer a job with the world’s leading autonomous vehicle team. “He had skills, and it’s good to have bright minds with opposing views around,” Levandowski says. Just as remarkably, Kentley-Klay said no. He didn’t think Google was radical enough.

Kentley-Klay returned to Australia. Several months went by. Contacts at Google and elsewhere stopped returning his emails. He began to think he’d made a terrible mistake. He says he saw a psychiatrist. Then he flew back to the U.S. in April 2014 and waited outside Levandowski’s house until the Google engineer got home one evening. They talked. Levandowski mentioned one guy Google really wanted to hire but could never get, a Stanford engineering grad student named Jesse Levinson.

Levinson, 35, is Kentley-Klay’s polar opposite. He’s thin, quiet, and picks his words carefully. He does his absolute best never to bring up his fabled Silicon Valley parentage. At Stanford, Levinson became the protégé of Sebastian Thrun, the professor who went on to lead Google’s self-driving car project. “Jesse has been one of my smartest students ever,” Thrun says.

While at Stanford, Levinson invented a new way to calibrate the sensors on self-driving cars. These types of vehicles typically rely on cameras and lasers to build a picture of the world around them. To fine-tune the imaging systems, engineers often hold up posters with checkerboard and target patterns as a baseline. In the field, though, the sensors can be difficult to reconfigure when out of whack. Levinson wrote software that made it possible to configure the sensors while driving, using objects in the real world to provide feedback instead of the test patterns. “The vehicle can figure out where its sensors are with superhuman levels of accuracy, down to 2 millimeters and 1/100th of a degree,” he says.

After Kentley-Klay tracked down Levinson, they agreed that the salesman’s vision and design skills could pair well with the engineer’s technical acumen. Both liked the idea of battling conventional thinking and building something they could call their own. “I could never tell what Google’s end goal with the technology really was,” Levinson says. “And I have a hard time motivating myself to work hard on something if I can’t see in my mind where it was going.”

Levinson didn’t buy in right away, however. First, he hired a private investigator to run a background check on his would-be partner. “I didn’t think he was a crazy person,” Levinson says. “I just didn’t know who he was, and he had an unusual background for someone starting a self-driving car company in Silicon Valley.” The only things that turned up were a couple of speeding tickets. “I didn’t know whether to take it as an insult or a compliment that he was taking me seriously,” Kentley-Klay says. “But they never found the body, so I passed the test.” Zoox incorporated on July 29, 2014.

Smack dab in the middle of Silicon Valley sits the SLAC National Accelerator Laboratory. The most distinguishing thing about the 426-acre compound is a 2-mile-long particle accelerator that cuts through the grassy hills of Menlo Park and onto the Stanford campus. It’s a high-security jewel of American nuclear physics. It also contains a series of winding, out-of-sight roads that are perfect for quietly testing autonomous vehicles. Somehow, Kentley-Klay persuaded someone there to let him use an old firehouse at the complex as Zoox’s first proper headquarters.

In early 2015, Zoox began hiring staff and retrofitting the firehouse into a prototyping facility. Engineers created skeletal versions of the robots while a software team worked on the contraption’s brain. In a distressing early sign for Zoox’s investors, Kentley-Klay also spent $16,000 on a Sub-Zero office refrigerator because he thought it looked cool.

From those first days, Zoox and its founders had a clear picture of the vehicle they wanted. It would have an identical front and rear, and would be easy to service on the rare occasions when it wore out its built-in redundant parts. Each wheel would have its own motor, so the vehicle could make precise maneuvers in tight spaces and park just about anywhere. And its array of sensors and cameras would be seamlessly integrated, not jammed onto an existing vehicle.

Three generations of self-driving vehicles (VH1, VH4, and VH5) at Zoox manufacturing headquarters in Foster City, Calif.PHOTOGRAPHER: CHRISTIE HEMM KLOK FOR BLOOMBERG BUSINESSWEEK
Lines of LEDs on the front and rear of the vehicle would send signals to other drivers, such as alerts that the robot taxi had spotted an obstruction up ahead on the road. Similarly, its directional sound system would let out a bleep or a blurp to tell a pedestrian in a crosswalk that the vehicle saw him, or to sound an alarm to the driver of a fast-approaching vehicle that he needed to get off his smartphone and hit the brakes to avoid a wreck. Early on the Zoox engineers considered a giant airbag that would envelop the vehicle before an accident; they ultimately went with more conventional airbags for the cabin. Zoox cars will come with high-end audio, plush seats, and some sort of conversational app for interacting with the riders.

The company has six prototypes, or mules, in auto industry lingo. They’re named VH1, VH2, and so on—the VH being short for “vaporware horseshit,” which is how a car blog once described the company’s technology. During a recent visit to SLAC, the mules were put into action with a series of demos. In one, the mule parked with extreme precision in a spot outside the firehouse. In another, it came to a controlled stop for a pedestrian making his way through a crosswalk and issued a bleep-bloop in greeting. In a separate demonstration at an abandoned airfield, the mules really showed off, tearing autonomously through an obstacle course at 50 mph. Your reporter, in crash helmet and safety harness, had the privilege to be the first person experiencing this test in the backward-facing seat.

A passerby photographs a Zoox self-driving car as it pulls out of the back lot at the San Francisco office and hits the road for a test drive.PHOTOGRAPHER: CHRISTIE HEMM KLOK FOR BLOOMBERG BUSINESSWEEK
The real proving ground for any self-driving car, though, is actual streets and highways where texters, road ragers, and the generally erratic roam. On a weekday in May, Kentley-Klay greets me in a parking lot behind the firehouse. A Toyota Highlander is parked about 100 feet away. The Zoox prototypes aren’t yet street legal, which means the company must rely on a fleet of Highlanders to train and test its sensors and software. The vehicles have cameras and lasers dangling off their sides and huge, humming computers in the rear storage area.

Kentley-Klay hands me an iPhone. I open the Zoox app and summon one of the Highlanders. We hop in and tell the car to head north toward Zoox’s new headquarters in Foster City, which is about 20 miles away, to drop off Kentley-Klay and pick up Levinson. From there, Levinson and I ride another 20 miles to San Francisco. Bay Area traffic being what it is, the whole journey takes about 90 minutes and is, well, pretty amazing.

Highways are easier for self-driving cars. Many makes and models now on the road use adaptive cruise control and other features that can follow another car on the freeway and maintain a safe distance. Prototypes from Waymo, Alphabet Inc.’s self-driving spin-out, can handle city streets, albeit in less densely populated areas such as Arizona. It’s really only Zoox and GM Cruise that have been willing to take outsiders on autonomous drives through a place as busy as San Francisco.

A robotic arm sculpts an entryway for the Foster City location.PHOTOGRAPHER: CHRISTIE HEMM KLOK FOR BLOOMBERG BUSINESSWEEK
The Zoox vehicle makes its way through the suburbs with ease, politely waiting its turn at four-way stops and giving cyclists plenty of room. When a black delivery truck unexpectedly whips across two lanes of traffic, the Highlander stops and avoids a collision. A few minutes later, we get on the freeway, and the Toyota merges in a manner that could be described as ultra-safe mode. Instead of jamming the accelerator to outrun oncoming traffic, it hugs the edge of the on-ramp as it waits for an opening.

It’s in the city, though, where Zoox really shines. The screens inside the vehicle show an overwhelming amount of information, as the computer vision software keeps tracks of cars, people, stoplights, and road markers all at the same time. Unlike many self-driving cars, it glides to stops. At an intersection with a left turn, it allows oncoming traffic to pass and then waits for some slow pedestrians. Overall, the vehicle performs so well that you forget no one is driving.

In May, Zoox moved most of its 500-plus employees into a new 130,000-square foot headquarters in Foster City that Kentley-Klay helped design. The facility is sleek and spare, with lots of glass and polished concrete. At its center is an all-white manufacturing hub, where workers will soon begin hand-building the first fleet of vehicles. A machine runs a Zoox drivetrain in place, simulating thousands of hours of driving to keep training the autonomous system. “It’s like a virtual reality suit for a car,” Kentley-Klay says. Nearby, a custom computer rig contains 1,000 ultra-high-end graphics chips that can each perform 40 trillion calculations per second on driving and artificial intelligence problems. “It might be the largest supercomputer of any startup,” Levinson says. In another part of the facility, there’s an enclosed area that only a handful of specially authorized people can enter. This is the top-secret zone where Zoox will put the finishing touches on things such as its vehicle’s industrial design, sound system, and branding. (Because he chairs Apple, Arthur Levinson hasn’t seen any of this. “It’s not that I don’t want to,” he says. “It’s just better to keep a little distance and read about Jesse in the newspaper.”)

Current and former employees say Kentley-Klay and Levinson have handled their first gigs as company chiefs well, for the most part. Zoox has managed to hire away hundreds of engineers from Tesla, Apple, Google, Ferrari, and, in large part because it offers a harder engineering challenge than anywhere else. The co-founders do have their controlling sides. Kentley-Klay is health-conscious, forbidding sodas in the office—even diet—and publicly shaming employees who send out “doughnuts in the kitchen” messages. Levinson takes pride in correcting grammar, to the point that employees proofread one another’s e-mails.

A self-driving vehicle in Zoox’s garage at SLAC in Palo Alto.PHOTOGRAPHER: CHRISTIE HEMM KLOK FOR BLOOMBERG BUSINESSWEEK
The pair have mastered the hyperbolic vernacular of the Silicon Valley startup scene. Text running around the wheel wells of the Zoox vehicles reads, “Infinity is enough,” a phrase the company has trademarked. Kentley-Klay’s own name is another invention. He was born Tim Kentley and adopted the Klay. “I have added Klay to my surname, as I find I just love making things and [sic] is a big part of me,” he wrote to the Zoox employees in 2013. “So, clay, or mud, is the primal aspect of this spirit, and ‘K’lay is a word play on that to keep the K in the mix of the evolving family name.”

There are plenty of people in the self-driving car business who regard Zoox as pure VH. Levandowski is one of them, even though he considers Levinson and Kentley-Klay friends and first-class minds. Levandowski still drives an old Lexus he bought from Art Levinson, and he designed a stockpicking system called FutureGame with Jesse.

“I don’t think making the car is the gangsta differentiation,” Levandowski says. “It’s overly complicated, and the wrong way to go.” Zoox, in essence, wants to beat Waymo at self-driving car technology, Tesla at electric vehicles, and Uber at ride-sharing. “If anything goes bad with one piece of that, you’re f—ed,” Levandowski says. “But they’ve heard that a billion times.”

Kentley-Klay and Levinson both readily admit that Zoox could end badly. Neither, however, appears to have any intention of altering course, and they vow to stun competitors and consumers with surprises yet to come. “For me, it’s not complicated,” Kentley-Klay says. “You have to think about what will give you the best result and then walk down that road, even if it means it’s the harder road to go down.”

Category Tag


Add your comment