Who Will Build the Next Great Car Company?

Silicon Valley and Detroit are in a race to create our driverless future.
June 23, 2016, 10:30 AM UTC
Illustration by Gianmarco Magnani for Fortune Magazine

Who Will Build the Next Great Car Company?

Silicon Valley and Detroit are in a race to create our driverless future. And for the first time ever, the car may take a backseat.

My brain knows that this demonstration has been carefully staged and will work exactly as planned, but the rest of my body tenses up as I step on the gas of a Ford Fusion sedan and accelerate directly toward the “parked car” in front of me.

Rear-ending cars on purpose is not a natural act. It takes a metric ton of concentration to resist slamming on the brakes, so I play the role of distracted driver. To the right, a parking lot packed with brightly painted Ford vehicles sparkles in the afternoon sun. In the distance, an American flag waves from its pole. Ford’s iconic blue logo is painted into the side of a grassy hill.

I hear a warning beep from the car’s dashboard. A red light flashes on the display panel. I imagine that I’m a character from the movie Mad Max: Fury Road, chanting in my head, “I live, I die, I live again.” Inches from the decoy car, I let out a shriek as the automatic brake kicks in, jerking the Fusion to a sudden halt.

From the passenger seat, Scott Lindstrom, manager of driver-assist technologies at Ford Motor Co. (F), assures me I’m not the first person to scream during this demo. I ask how he simulates accidents day after day. “It is very nerve-racking,” he says. Also, he’s hit the decoy plenty of times. In 2012 he even did it in front of Ford’s board of directors.

Back then the idea of self-driving cars looked, to Ford’s leadership, like a frivolous Silicon Valley moonshot. Four years later things have dramatically changed. Today Ford’s vehicle lineup features more than 30 options for semiautonomous features, including the automatic brakes I tested, and the company is aggressively working on cars that fully drive themselves. By year-end the company expects to have the largest fleet of autonomous test vehicles of any automaker.

Ford is not alone. The entire automotive industry is in the midst of a radical transformation that is reshaping the very definition of what it means to be a car company. There is hype, hope, fear, and insecurity—and at the center of it all is the self-driving car. Thanks to cheap sensors, powerful machine-learning technology, and a kick in the butt from the likes of Google (GOOGL) and Tesla Motors (TSLA), driverless vehicles are becoming a sooner-than-you-think reality. General Motors, Toyota, Nissan, Volkswagen, Fiat-Chrysler, BMW, and just about every other auto company are wading—some cautiously and some with big, headline-grabbing moves—into territory that executives in Detroit and elsewhere not long ago considered a science-fiction fantasy.

A Ford Fusion Hybrid autonomous research vehicle.Photo: Courtesy of Ford Motors

Everything changed in March, when GM spent $1 billion on a tiny self-driving startup called Cruise Automation. For months tech companies, startups, and automakers circled one another, unsure of whether and how to partner. In an industry moving this quickly it’s hard to know if a friendly partner will turn into a competitor. It happened when Google’s venture capital arm invested $258 million in Uber in 2013; now the ride-hailing company is poised to directly compete against its investor with its self-driving project.

But the hesitancy ended after GM’s deal. In May, Toyota struck a partnership with Uber, Volkswagen invested $300 million in ride-hailing company Gett, Apple poured $1 billion into China’s Didi Chuxing, and Google partnered with Fiat Chrysler to outfit 100 Pacifica minivans with self-driving technology. All in the span of one month.

“Any good company has a point of view of the future,” Ford CEO Mark Fields says.Photo:Qilai Shen—Bloomberg via Getty Images

No one is sure how our driverless future will look. Will we be able to scroll through our Instagram feeds from the backseat while our cars drive us to work? Will we even need to own private cars anymore? Some people expect that regulation will limit self-driving cars to closed-off areas in major cities. Others believe that the first driverless fatality will set the whole effort back by a decade. Predictions as to when fully autonomous vehicles will actually hit the road range from 2050 to as soon as next year. (The answer for Tesla owners is last fall—they have been posting “Look, Ma, no hands!” videos online since October, when Tesla updated its software with an “autopilot” feature that steers the vehicle on highways—not fully autonomous, but a strong start.)

Automakers, technology companies, and ambitious startups all agree that this transformation isn’t just headlines and hype, but inevitable. Every person I interviewed for this article deliberately pointed out that, no, really—this is happening. Even regulators are onboard. “We’ve got a clock ticking,” U.S. Transportation Secretary Anthony Foxx told Reuters in April. “This technology is coming. Ready or not, it’s coming.”

Sure, there are hacking fears and privacy issues and ethical questions and infrastructure challenges and concerns about lost jobs. But those will all be worked out, the industry’s thinking goes, once everyone realizes how many lives could be saved with this technology. Car accidents cause more than 1 million deaths and 15 million injuries globally each year. Over 90% of them are caused by human error. For some executives, the argument is so obvious it’s worth blowing a gasket over. “How could you accept that? Why is this not a national crisis?” bemoaned Eric Schmidt, executive chairman of Google’s parent company, Alphabet, at a shareholder meeting in June. “I am, frankly, beside myself over the lack of consensus on how important this problem is.”

Beyond saving lives, consider the money saved (traffic accidents cause $500 billion in economic damage worldwide each year), the benefits to cities (parking lots become green spaces), the increases in productivity (commuters can work in transit; truckers can sleep without pulling over), and improvements in accessibility (the elderly, blind, and disabled get affordable robot chauffeurs!).

But most important from a business perspective, driverless vehicles are poised to threaten the $570 billion that Americans spend each year on new cars. For 125 years U.S. auto companies made their money on the manufacture of motor vehicles. Now they must be in the business of ride-hailing apps, shuttle buses, 3D maps, and computers on wheels that drive themselves. They’re no longer automotive companies either—they’re now calling themselves “mobility” companies, just in case all those predictions about the end of car ownership come true. At stake is a transportation services market that Ford believes is worth $5.4 trillion, a sum that makes you wonder why it took the auto industry so long to go after it.

It’s telling that ride-hailing leader Uber got its start while the auto industry was facing bankruptcy in 2009. Now Uber is worth more than GM and Ford by tens of billions of dollars, despite generating approximately $145 billion less in revenue. Alphabet has enough cash lying around to buy Ford or GM outright if it wanted to. Apple (AAPL), which is widely rumored to be working on its own car, could buy them both plus Fiat-Chrysler in a Big Three value pack.

The Silicon Valley invasion has not gone unnoticed. “There are enough people working at legacy companies in Detroit who have read The Innovator’s Dilemma. They know all the business school case studies about how incumbent organizations can’t sit around and ignore innovators,” says Brian Johnson, a Barclays analyst who has argued that the economy will reach “peak car” ownership in the next year. “The incumbents are willing to move more quickly than I would have expected. There is almost an eagerness to show they’re not dinosaurs.”

Ford president and CEO Mark Fields, for example, has taken to declaring that his company needs to “disrupt itself” before Silicon Valley players get the chance. The sentiment raises several questions: Does he really mean it? And does he understand what that takes? Even if the answer is yes to both, a much bigger question looms: Can he?

For six years Anthony Levandowski worked on Google’s self-driving car project, developing sensors, creating policies, and testing technology. He’s credited as one of the architects of the company’s famous driverless vehicles. In turn Google, the online advertising giant that’s increasingly known for its ambitious research projects, is credited with showing the world how fast this loopy curiosity could become a business reality. Despite that progress, Levandowski and Lior Ron, a product lead on Google Maps, couldn’t stand the idea of waiting any longer to bring autonomous driving technology to market. In January they left to launch Otto, a startup that retrofits commercial big rigs with fully autonomous driving systems.

Google’s first self-driving concept.Photo: Courtesy of Google

After only five months of development, the founders made their first public demonstration, sending an 18-wheeler barreling down California’s 101 highway at 55 mph with no human interference. There’s a sense of urgency at the company. “We’ve been moving as fast as we can,” says Ron. “It’s not going to be a 10-year or even a five-year horizon. We intend to bring the technology to market faster.”

However soon you believe that driverless vehicles are coming, Silicon Valley innovators want them even sooner. Big and small tech companies alike have decided that transportation is the next frontier in need of disruption. They’re not adopting Detroit’s historical approach of incremental improvements, deploying ever more semiautonomous features until the robots finally take over. Rather, they’re plotting revolution. Google’s self-driving pod-cars don’t even have a steering wheel or pedals. (The National Highway Traffic Safety Administration defines this as “Level 4” autonomy—the agency’s highest level.)

Self-driving technology could reduce spending on roadways, parking, and public transit, says Chris Urmson, director of Google’s self-driving car project.Photo: Drew Angerer—Bloomberg via Getty Images

Google argues that revolution is the safest option. Requiring a licensed driver be able to take over from the computer actually increases the likelihood of an accident because people aren’t that reliable, the company says. (Plus, keeping a human element in the driving mix destroys the economic and accessibility benefits that autonomous technology promises.) Chris Urmson, Google’s director of self-driving cars, made the case for full autonomy—in vain—to California regulators earlier this year. GM, meanwhile, is pro–steering wheel and pedals, according to recent comments made by CEO Mary Barra. The U.S. Department of Transportation plans to announce a new set of guidelines in July.

Zoox, a Menlo Park, Calif., startup, is firmly on the revolutionary side of things. According to a snazzy computer rendering from the company, it’s building perfectly symmetrical, bidirectional, neon-green-and-black robo-taxis that look like giant remote-control toys. The vehicle is not really a car, founder Tim Kentley-Klay likes to say. “It’s what comes after the car.” That’s because Zoox is not bothering with incremental, semiautonomous features like the automatic brake I tested with Ford. That approach is “not going to give us the changes we want to really solve mobility issues in our cities going forward,” Kentley-Klay explained at a 2014 conference in Berlin. Zoox has not yet shared its progress beyond the early rendering, but it has big plans. The company is currently raising $252 million in funding at a reported valuation of $1 billion.

Many people assume Zoox, Google, and Uber (which revealed its own driverless test car in May) will eventually launch their own urban fleets of on-demand autonomous vehicles. Basically robo-taxis. For Uber, reducing or eliminating its army of more than 1 million contract drivers would slash the unprofitable startup’s costs. Uber currently pays about 75% of its income to drivers. Eliminating that outlay would likely make up for the new expense of buying, maintaining, and storing a fleet of driverless cars.

For Google, a ride-hailing service allows the company to turn one of its most expensive moon shot projects into a real moneymaking business. Google will reportedly make its driverless car project a standalone subsidiary this year. Executive chairman Schmidt told investors the company believes the cars will be commercially available in “some years, not decades.” But first Google needs to reduce the cost of its sensor-covered vehicles from around $170,000 today to less than $30,000, according to a person familiar with the company’s plans.

Otto self-driving truck.Photo: Courtesy of Otto

And that’s a key problem in Silicon Valley’s robo-taxi plan: The nerds have little experience bending metal. Tesla, the one Silicon Valley company that’s managed to actually make a car, has been late to deliver every new model it has released since the launch of its very first vehicle (the Roadster, thrice delayed) in 2008. Tesla operates at a much smaller scale than its established peers in Detroit, Japan, and Germany, but its manufacturing problems aren’t just normal growing pains. For its latest Model X sport-utility vehicle, Tesla blamed a components shortage in part on its own hubris. It tried to add “far too much new technology” to the vehicle, the electric-auto maker said in a statement.

Auto revolutionaries, take note: If Elon Musk, the real life Iron Man, struggled with manufacturing, then everyone else will too.

When Google unveiled its fleet of zippy little pod-cars in 2014, executives in the automotive industry could barely hide their smirks.

“No automotive manufacturer would ever have dared to show such an ugly potato in public,” said Ulrich Weinberg, a professor at the Hasso Plattner Institute in Germany, in an interview published in Audi’s 2014 shareholder letter. Auto critics had a similar reaction. (Google’s first self-driving concept looked “like a vintage Japanese cartoon character had an illegitimate child with a computer mouse,” wrote Andrew Krok.) “People were literally laughing at Google and Tesla,” says Lars Reger, chief technology officer of NXP Semiconductors, a supplier of self-driving vehicle technology.

Anthony Levandowski, one of the architects of Google’s self-driving car, left the company to start Otto, which retrofits tractor-trailers with autonomous technology. Anthony Levandowski, one of the architects of Google’s self-driving car, left the company to start Otto, which retrofits tractor-trailers with autonomous technology.Photo: Courtesy of Otto

The dismissal shows just how different the two industries’ worldviews are. In a vehicle, Google saw “a robot that by accident has four wheels,” Reger says. Traditional car manufacturers saw “a horse carriage with a combustion engine.” Tech companies launch ideas and iterate on them with continual improvements. “That doesn’t work in our business,” Ford’s Fields says. “You can’t hit control-alt-delete when you’re going 70 miles an hour.”

It’s easy to be dismissive when times are good, as they have been for Detroit lately. For the past three years the auto industry has delivered record sales, profits, and growth. GM and Ford are still Fortune 10 companies.

Yet Ford’s stock has fallen by 12% in the past year. GM’s stock has dropped by 19% despite an announced $9 billion of share buybacks. At Ford’s annual shareholder meeting in May, a frustrated investor even asked about bringing former CEO Alan Mulally out of retirement to boost the company’s share price. The existential threat facing automakers is compounded by investor fears that the automakers’ recent boom is as short-lived as today’s low interest rates and cheap gas prices.

Worse, analysts have long speculated about the coming of peak car, comparable to the peak in horse ownership in 1920. With cheap robo-taxis available to chauffeur people around cities, households that spend an average of $9,000 a year on transportation could lower that outlay to just $2,000. “Can the mobility sector capture all of that? I think no,” says Robin Chase, co-founder and former CEO of car-rental company Zipcar. “Car companies will still make money selling cars, but the whole market will shrink because we’ll use those vehicles more efficiently.”


Meanwhile, young people are increasingly uninterested in driving. The percentage of people between the ages of 16 and 44 who obtain a driver’s license has been steadily declining since 1983, according to a study by the University of Michigan Transportation Research Institute. Automakers bristle at the notion that millennials hate cars—GM president Dan Ammann calls it an “overly cliché story written by lazy journalists”—but it’s clear that automakers are no longer laughing at Google’s ugly potatoes.

The U.S. auto industry has been caught flat-footed before. In the 1980s, Japanese imports came to quickly dominate the low end of the North American market as European brands secured the high end. The specter of Google, Apple, Uber, Tesla, Lyft, or even Zoox cornering the future market of How Americans Get Around has created two kinds of paranoia in Detroit: a fear of taking on too much risk and a fear of not taking on enough. “Nobody wants to go too far and pull a Time Warner–AOL,” one exec told me. On the other hand, “nobody one wants to get iPod-ed like the music industry,” said another.

That fear is driving the great mobility transformation. In early 2015, Ford CEO Fields was preparing to deliver a “change the world”–style speech at CES, the annual consumer electronics trade show, spanning urbanization, the middle class, air quality, and millennials. It included lines like, “Mobility is about far more than motion—it’s really about progress. Human progress.”

Preparing for that event was the moment Ford’s executives realized they needed to shift, says Ken Washington, Ford’s vice president of research and advanced engineering. “It kind of jelled for us that, wow, this is really about being a mobility company,” he says. “Over the last year that messaging has gotten more and more crisp.”

That epiphany led to the creation, in March, of Ford’s Smart Mobility subsidiary, a separate LLC that will develop software, tech services, and business models related to transportation in the same way Apple makes software like iTunes and the App Store to complement its hardware businesses. “We’re viewing ourselves as the kind of company that gets that there’s more than one way to be mobile,” says Washington, adding that he doubts Ford’s Smart Mobility subsidiary “will have the famous Thursday business process review meetings that Ford is known for.”

The 2015 Los Angeles Auto Show “The end of car ownership is real, serious, and going to change your world,” Lyft co-founder and president John Zimmer told auto-industry executives at the Los Angeles Auto Show in November.Photo: Patrick T. Fallon—Bloomberg via Getty Images

GM president Ammann had a similar revelation two years ago. Ammann is a car person; he likes to drive his 1961 baby-blue Cadillac Series 62 convertible. But he found himself increasingly relying on a company-provided driver to chauffeur him to work and back because it freed up an extra hour and a half each day to get work done. “There’s this significant opportunity cost of the time people spend driving,” he says. GM’s customers were also telling the company that they wanted to use cars without the cost and hassle of actually owning them—the same need that gave rise to on-demand ride-hailing services like Uber and Lyft. “We see our customers behaving in this way,” Ammann says. “We need to move with them.”

At the Los Angeles Auto Show in November, Ammann watched Lyft president and co-founder John Zimmer deliver an ultimatum to the automotive industry: “You can fight [the end of car ownership], and that will probably not turn out well. Or you can acknowledge that this is happening. This is real, serious, and going to change your world.” The room, full of insurers, regulators, auto dealers, and parts manufacturers, was silent for a moment before a “slow golf clap” began to build, Zimmer says.

After the speech Ammann pulled Zimmer into a hotel suite, where the two quickly hashed out a deal for GM to invest $500 million into Lyft, valuing the company at $5.5 billion, investment included. The meeting, which Zimmer kept secret from his entourage, went for three hours. Zimmer has had plenty of meetings with auto executives over the years, but it was strange to explain his vision of the future and have one respond with “I agree,” he says.

A few months later Ammann met with Daniel Kan and Kyle Vogt, co-founders of Cruise, a 2½-year-old startup building a software platform for autonomous vehicles. As with Lyft, they shared visions, Ammann said “I agree,” and boom—a big, splashy deal. The companies considered merely partnering but realized they could move faster with an acquisition. “Every time we visited, they’d moved along another nine steps,” Ammann says. At the time, Cruise didn’t even have a plan to bring its technology to market. It had explored everything from pizza delivery and package transport to semitrucks and ride sharing, Kan says. Ammann was still convinced, and GM plunked down $1 billion for the 40-person firm in March.

The three companies plan to work together—using Lyft’s ride-hailing service, Cruise’s self-driving technology, and GM’s vehicles—to launch an on-demand driverless ride service to compete with Google, Uber, Zoox, and anyone else that comes along. Says Ammann: “We feel we have the three key pieces we need at this point.”

It matters little that the motivations of automakers (fear of being iPod-ed) and tech players (desire to change the world) stand in opposition. More important is that both parties share a goal: to get driverless cars on the road. Indeed, a lobbying group called the Self-Driving Coalition for Safer Streets formed in April, linking Google, Uber, Lyft, Ford, and Volvo. Digital transformation makes for strange bedfellows.

In the 1930s a driverless Tin Lizzy called the Phantom Auto toured the country to dazzle crowds with its self-driving capabilities. Imagine, gushed the Free Lance Star of Fredericksburg, Va., an empty automobile traveling the streets: “No one touching it, no wires or strings attached to it, weaving in and out of traffic, climbing hills, turning corners, stopping for traffic lights, just as though there were an invisible driver at the wheel!

The glory lasted until 1932, when the Phantom Auto mowed down 10 people, hospitalizing two and fracturing numerous limbs (and a skull) at a demonstration in Hanover, Pa.  The car’s operators, who remotely controlled the vehicle from a trailing car, were arrested on charges of assault and battery, and the driverless vehicle was retired.

Just as the thrill of self-driving cars is not new, neither is the fear of them. Eighty years later, in Google’s hometown of Mountain View, one of its pod-cars ran into a bus. In May, Google posted job listings for test drivers in Arizona, which tech bloggers painted as a dream job. Who wouldn’t want to make $20 an hour sitting in a car doing nothing for eight hours a day? But the social media reaction from nontechies was a glimpse into the public’s fears of robot cars. “You’re gonna have to pay more to get me in that tin can with a mind of its own,” wrote one Facebook commenter. Fueling the fear, Google filed to patent a grim-looking “human fly paper” material for the outsides of its vehicles in case one of its cars hits a pedestrian, complete with images of bodies stuck to cars.

Silicon Valley is great at building hype. Trust is a different story. Yet just about everyone working on self-driving vehicles believes that people’s fears will dissipate once they experience the technology in person. Regardless of whether they’re skeptics or enthusiasts, riders always react in the same way, according to Ziv Aviram, whose company, Mobileye, provides collision-avoidance systems to 25 automakers. During that first nervous minute, riders position their hands above the steering wheel and their foot above the brake, poised to take over in the event of a glitch. By the second minute they relax, and by the fifth minute they take their eyes off the road and chat as if they weren’t in the driver’s seat. After that, “there’s no way back,” Aviram says.

Auto execs understand this phenomenon well, rattling off examples of vehicle features that the public initially resisted: seatbelts, airbags, antilock brakes, cruise control, even automatic transmission. “I don’t want any computer stomping on the brake for me,” says Ford technical leader Jim McBride, remembering pushback from drivers of the past. “I don’t want an airbag blowing up in my face and I can’t see out the front window.” Even today, Ford’s self-driving test vehicles contain giant red fail-safe buttons next to the gearshift that are essentially for looks. (No one has ever used them, the automaker told me.)

Despite all this—the technological advancement, the sudden rash of dealmaking, the this-is-happening clarity stretching from Motor City to the City by the Bay—self-driving cars today remain a Phantom Auto curiosity. Our self-driving future requires interstate cooperation, ubiquitous connectivity, regulatory approvals, and buy-in from ancillary players like parking garages, insurance providers, and dealerships. “It’s not enough to throw some sensors and software on a two-ton projectile and say ‘Look, Ma, we made it smart,’ ” says Jamyn Edis, CEO of Dash, an automotive software startup based in New York City.

Many experts don’t expect that the world will adopt Level 4 autonomy—the ability to summon a car that goes from point A to point B without human interference—for 25 years, or a full generational change. But partial autonomy arrived much sooner than anyone expected. Google already has demonstrated “A to B” technology that works. By 2021, one of Mobileye’s automotive clients (Aviram won’t say which) will have it too. Five years may seem like forever to the hoodie-wearing tech crowd, but “in our industry,” Aviram says, “2021 is like tomorrow.”

Until then GM, Ford, and other auto giants will endure investor skepticism over whether their investments in apps and robo-taxis can ever become profitable businesses. Disruption, by definition, requires making one’s existing business smaller. It’s why so few Fortune 500 companies have managed to pull off what Mark Fields is attempting with Ford: disrupting themselves. That challenge is not lost on Big Auto executives. “Will Ford as a company be the size and scale it is today in 15 years?” asks Zipcar founder Chase. “Probably not.”

That doesn’t mean they won’t try like hell. Next year GM will launch a semiautonomous Super Cruise feature that allows for hands-free driving in its Cadillac CT6 at the same time that it develops “straight to Level 4” technology within its subsidiary Cruise. Ford is also trying to strike a balance between evolution, with its vehicle lineup’s semiautonomous features, and revolution, with its test fleet of driverless cars. Ford executives insist it is not totally contradictory to protect the company’s existing business while also building technology that will make it obsolete.

I think about that conundrum as I accelerate down a ramp along Interstate 94 on my way out of Detroit in my non-self-driving rental car. I squint and stare at the tractor-trailer in front of me while a reckless driver swerves between lanes at 80 miles per hour to my left. I realize that everything around me—from the radio ads for auto dealerships and cheap car insurance to the billboards for accident settlement attorneys—will be completely transformed by the self-driving car. And then I nearly rear-end the truck.

A version of this article appears in the July 1, 2016 issue of Fortune with the headline “Some Assembly Required.”