The Ultimate Driving Machine Prepares for a Driverless World

Hybrids, including the I8, will account for most of BMW’s sales within a decade.
Photograph by Adam Levy for Fortune Magazine

How 100-year-old BMW is racing toward automated driving, electric cars, and ride sharing.

 

This year BMW turns 100, and it’s still roaring like one of its deep-throated V-8 engines. In 2015 it outsold rivals Mercedes-Benz and Audi for the 11th straight year, keeping its crown as the largest luxury-car maker in the world. The newly launched 7 Series, which has reclining rear seats, has been wowing the auto press. What could go wrong?

Well, plenty. If Munich-based BMW wants to stay ahead of the pack, it must figure out how to maneuver through the biggest disruption in transportation since the car displaced the horse-drawn carriage. And that disruption is coming faster than most think. “The next 10 years will bring more change than the last 30,” says Christoph Grote, head of BMW’s advanced technologies group.

We’re not talking about more horsepower or improved handling. The very essence of the automobile is up for grabs. BMW’s honchos believe that new modes of transportation are needed to reduce traffic and boost safety while lowering greenhouse gas emissions and air pollution. It’s a major transformation that can be divided into three elements: autonomous driving, electrification, and car sharing. “There is an exciting alternative to the auto industry as we’ve known it over the last 100 years,” says Larry Burns, who once ran R&D for GM gm and now advises Google googl . “A lot of people will no longer be driving themselves in heavy gas-powered vehicles that they own personally, and the industry is going to have to adapt to that.”

That means BMW is now competing with some of the tech world’s most powerful names. Google, Apple aapl , Tesla tsla , and others are spending billions to build or plan electric, self-driving cars made of lightweight composites. Tesla has become the leading seller of electric cars in the U.S.  Google could have self-driving vehicles on the road within five years, a decade before the major automakers say they can do it. A Chinese startup named Faraday Future, backed by an Internet billionaire, Jia Yueting, announced that it is building a $1 billion electric-car factory in Nevada and unveiled its 200 mph concept car in January. “We’re going back to the future where we have the same kind of auto brand proliferation like we saw at the beginning of the 20th century,” says Tesla’s vice president of business development, Diarmuid O’Connell.

To see BMW’s profile on our World’s Most Admired Companies list, click here.

BMW’s new competitors have audacious ambitions. Uber, for example, is partnering with robotics experts at Carnegie Mellon and spending heavily to design a self-driving vehicle. The car-hailing service says its core vision is to “end private car ownership.” CEO Travis Kalanick reportedly told Tesla CEO Elon Musk that if Tesla could build 500,000 autonomous cars in 2020, Uber would buy every one of them. Kalanick’s notion: a massive fleet of self-driving Uber vehicles ready to pick you up whenever and wherever you’d like.

That’s a bit of hyperbole, perhaps, but there’s plenty to worry traditional automakers, starting with the fact that, increasingly, what differentiates cars is their software. A typical premium model has 100 million lines of code, according to Boston Consulting, which estimates that a vehicle’s electronics as a percentage of its total cost has doubled from 20% in 2004 to 40% last year. Software, obviously, is where Silicon Valley excels.

BMW_1

Still, BMW maintains it will lead this revolution. “Electro-mobility and digitalization are the clear drivers of the right strategy for the future,” says BMW CEO Harald Krueger. Adds Ulrich Quay, who runs the company’s venture investing unit: “We will not succeed if we just keep doing business in the same way.” BMW’s new ways start with launching cutting-edge electric and plug-in hybrid cars, including the popular i3 city car and the head-snapping i8 sports car. Its 7 Series can drive itself on the highway. BMW has invested in companies that enable you to rent or share a car, find parking, and get traffic information. This fall BMW joined Audi and Daimler to acquire Nokia’s digital mapping and location services business for roughly $2.8 billion—an essential component of self-driving technology. And BMW has been working on mobility technologies similar to those of its Silicon Valley rivals.

Yet being aware of these trends is not the same as capitalizing on them. BMW faces the classic innovator’s dilemma: Successful established companies get pushed aside unless their managers know when to abandon the very products that built their prosperity. Auto manufacturers like BMW have a huge legacy to overcome, including the billions invested in factories that make gasoline-powered vehicles. Add to that unionized workforces, hidebound dealers, and conservative boards of directors.

BMW CEO Harald KruegerPhotograph by Myrzik and Jarisch

“It’s going to be the agile that will end up the winner,” says BMW’s Grote. “We are a great company with a lot of engineering talent. We never got stuck on what we have but always have invested in future technology to build something our customers love. And if you do that, you shouldn’t be afraid of the future.” But if self-driving cars become the norm, how will BMW market itself as the ultimate driving machine?

BMW made its reputation by building fast, great-handling cars with high-performance gasoline engines. That formula is speeding toward a brick wall. According to the International Energy Agency, the number of cars and other light-duty vehicles on the road will double to 2 billion by 2050. Most of that growth will occur in the developing world. The air quality and traffic jams in Beijing and Mumbai are already horrendous. Imagine what will happen when you boost the numbers of cars on those roads dramatically. Something has to give.

Governments are already putting increasingly strict limits on emissions. In the U.S., automakers’ fleets will have to average 54.5 miles per gallon by 2025. In the EU, regulations will force carmakers to reduce their average CO2 output per car to about 95 grams per kilometer traveled by 2021, a 27% reduction from today. China is adopting standards based on the EU rules.

Perhaps the biggest challenge comes from California. By the 2025 model year, 15.4% of vehicles sold in the Golden State will have to be what regulators define (with some latitude) as having “zero” emissions—electric, hydrogen, or a plug-in hybrid. The auto research firm Edmunds estimates that in California alone companies would have to sell 270,000 such vehicles annually, up from a few thousand today, to meet the standards. Nine other states, including New York, New Jersey, and Connecticut, have said they will follow California’s lead.

It’s hard to see how tinkering with powertrains will get carmakers where they need to be on emissions. That’s why BMW is pursuing a two-prong strategy: plug-in hybrids and hydrogen cars. Already it offers five plug-in hybrids: the 330e and 740e xDrive, the X5 xDrive 40e SUV, the i8, and the i3. In 2015, U.S. sales of the i3 jumped 81% from the previous year even as gas prices fell.

BMW insists it can create the same driving experience with a hybrid as it can with its powerful gasoline engines. After driving the i8 on the streets of Austin, I concluded BMW might have a point. With its gull-wing doors, this $140,700 two-seater sports car employs a small, turbocharged three-cylinder engine and two electric motors for a combined 357 horsepower. The i8 hits 60 mph in 4.2 seconds, faster than the company’s high-performance M3. It handles as well as if not better than any BMW I have driven. On top of that, the car emits only 49 grams of CO2 per kilometer, well within the 2021 European standards.

In the next five to 10 years, BMW’s line will probably consist mostly of plug-in hybrids. Keeping a combustion engine onboard avoids the dreaded range anxiety of electric cars but it also adds weight, complexity, and cost compared with a pure electric. BMW makes an all-electric city car—the i3—but believes that drivers don’t want the hassle of having to charge up every 80 miles. (Indeed, an i3 model with a small gas engine that extends the car’s range to 150 miles has handily outsold the purely electric version.)

BMW's I8 plug-in hybrid can go from 0 to 60 in 4.2 seconds. BMW’s I8 plug-in hybrid can go from 0 to 60 in 4.2 seconds.Photograph by Adam Levy for Fortune Magazine

One way to eliminate range anxiety is to build a national network of superfast chargers that can refill an electric in 20 minutes—precisely what Tesla is doing. BMW, however, thinks that’s too expensive. It’s more enthusiastic about hydrogen cars for long-distance travel. A hydrogen car is very similar to an electric but instead of being powered by a battery, it uses a hydrogen fuel cell to run the car’s electric motors. Just fill up with hydrogen and you can drive 300 miles or so. BMW, in collaboration with Toyota, has a hydrogen prototype car about which BMW’s Grote says, “You’d be surprised how convincing it is.”

As with electrics, hydrogen cars don’t yet have the fueling infrastructure—a huge shortcoming. Germany is already building such a network, and the state of California has earmarked $200 million to support anyone who wants to build hydrogen stations, but so far the U.S. has only a handful of them. It’s also not clear whether using natural gas—the most common process to make hydrogen—significantly reduces carbon emissions. Hydrogen’s shortcomings have led Tesla’s Musk to refer to the technology as “hydrogen fool.”

Another argument against hydrogen is that batteries are getting cheaper and better. GM recently announced it will start selling the Chevrolet Bolt this year, an all-electric that gets 200 miles per charge and sells for $30,000 after federal subsidies. In 2017, Tesla plans to have an all-electric, the Model 3, for $27,500 (after subsidies), with a range of at least 200 miles.

Hybrid, electric, and hydrogen cars are more expensive to buy than conventional cars, but the lifetime ownership costs can be cheaper. That’s because electricity costs less than gasoline (hydrogen has similar potential) and the mechanicals and maintenance are simpler (no transmission, complex gas engine, or cooling system). That said, sales of plug-ins, such as the Nissan Leaf and Chevy Volt, have suffered from today’s low gas prices.

Another big philosophical gap between Silicon Valley and BMW is the timetable for self-driving cars. Tesla and Google see no reason they won’t be in use by 2020. The benefits are obvious: reducing global traffic deaths (now 1.2 million each year) and congestion. If you had self-driving cars whirring along the interstates at high speeds, each five feet apart, a four-lane road could handle the same volume as a 32-lane highway.

Self-driving vehicles could provide transport for the elderly, the disabled, and the young. You could enjoy a few (or more) glasses of wine at your favorite restaurant and be safely chauffeured home. Musk, speaking last summer at a conference, argued, “We may outlaw driving cars because it’s too dangerous. You can’t have a person driving a two-ton death machine.”

Along with most other major carmakers, BMW has gotten on the self-driving bandwagon. It has a 5 Series prototype that can drive itself. Toyota, Nissan, and Daimler have all announced big investments or self-driving concept cars.

But most major automakers believe completely autonomous cars won’t be common until 2030 or 2035. First, the technology is expensive (though costs are sure to drop when produced at scale). That’s why the first place we’re likely to see self-driving vehicles is in shared fleets like those Uber is envisioning. A car driven 70,000 miles a year offers a better return on expensive laser and sensor systems than one that’s privately owned.

BMW_2

Another worry is that the technology isn’t ready. Can the car react properly if something unexpected suddenly appears in the road? Do you program it to hit a kid who runs in front of your car or direct it into a tree where you end up the casualty? “The human brain is an amazing device,” says Tesla’s O’Connell. “But if we have a learning software out there where every mile a car drives it learns something new, the software becomes safer and more useful.”

To understand why Silicon Valley is more optimistic about autonomous vehicles, take a drive with me on Highway 101 in Palo Alto in a Tesla S sedan equipped with its new Auto Pilot software. The owner of the car, Stefan Heck, is a Stanford professor and founder of Nauto, a startup that’s designing a system to retrofit and network today’s cars to make driving safer and more efficient. We are traveling about 75 mph, and I don’t have my hands on the steering wheel. I’m feeling very nervous, and I’m thinking Heck should look a lot more worried than he does. Still, I decide to try the automated lane-change function: I flick the directional signal to the left; the car accelerates until it finds an opening and then darts into the left lane, with my hands still off the wheel.

BMW.03.01.16.Apple.Sidebar Illustration by The Heads of State

As we continue down the 101, I ask Heck why the directional signal keeps blinking after we change lanes. “Good point,” he replies. He touches an icon on the dashboard screen and says, “Make the directional signal automatically turn off after changing lanes.” This bug alert is sent to Tesla’s engineers in Fremont, Calif., who presumably will work on a fix. The point is that Tesla is willing to take risks with self-driving technology and fix problems on the fly.

BMW_SB1

BMW knows how to make a car drive itself. But it’s more cautious—it wants its R&D crew, not its customers, to do the testing. This might be prudent, but it puts the company at a disadvantage. Silicon Valley is very good at quick iterations: Launch, get user feedback, fix, then relaunch. This approach helps explain why the Googles and Teslas of the world believe they can begin selling truly autonomous vehicles a decade ahead of the major automakers.

In Manhattan’s West Village, a neighborhood with ever more startups and hipster restaurants, BMW opened an office for its venture capital arm, iVentures. (The “i” stands for innovation.) The loft space is inhabited by a number of startups run by twentysomethings—many backed by BMW. Think of it as an incubator for the mobility of the future. Ulrich Quay, whose previous job at BMW was as an in-house lawyer, runs this $100 million fund. His mission is to invest in startups that will keep the company up to speed on the dramatic changes occurring in transportation.

Quay says he and his team have looked at about 1,500 different companies and invested in about a dozen. BMW owns part of Just Park, an app that allows people to rent out their driveways when they’re not being used; ChargePoint, the largest maker of electric-car charging stations in the U.S.; and the Israeli company Moovit, which has an app to help people get from A to B using public transportation.

A gesture can control a phone or the sound system on BMW’s 7 Series cars.Courtesy of BMW

BMW_SB2

One of BMW’s investment themes is car sharing. Consider that we use our private vehicles on average only 4% of the time; they sit idle 23 hours a day. The value of the billion cars on the road globally is roughly $20 trillion—does it make sense to use such valuable assets only 4% of the time? Says Brook Porter of venture capital firm Kleiner Perkins, which is an investor in Uber and car-sharing startup Turo: “If OEM’s are just making widgets, they will obviously become commoditized. Increasingly today customers are asking for transportation as a service instead of a product —which is dramatically increasing the utilization rate of cars beyond 4%. Autonomous driving will no doubt increase that utilization even further.”

The appeal of these services is obvious. Americans spend about $1,500 a month on their cars, and in an urban environment a vehicle can be expensive and a hassle to park. Why not just borrow one when you need it? It’s easier and cheaper. Susan Shaheen, a professor at the Transportation Sustainability Research Center at the University of California at Berkeley, estimates that today there are 4.8 million car-sharing members in 33 countries, with about a third of them in North America. According to Navigant Consulting, the global revenue of car-sharing services will grow from $1 billion in 2013 to $6.2 billion by 2020, with over 12 million members worldwide.

BMW is betting on DriveNow. It owns 50% of the sharing service, which operates in nine cities, most of them in Europe. The DriveNow app offers access to a range of BMW vehicles, which it rents by the minute. Unlike a Zipcar, which needs to be returned where you started, a DriveNow vehicle can be dropped off anywhere you’d like within certain zones. (Zipcar is now testing its own one-way service.)

One challenge is finding parking spaces where renters can drop off the cars. In October, DriveNow was forced to pull out of San Francisco—its only U.S. market—when the city refused to grant it permits to park on the street. Undaunted, BMW says it will open new locations in the U.S.  Another obstacle is that Americans, with the exception of city dwellers, are unlikely to give up the convenience of having their own car.

BMW.03.01.16.chart

So far the car-sharing market is crowded, and profits are hard to come by. Besides BMW, other carmakers such as Daimler, Audi, Ford f , and GM have car-sharing programs, as do most large rent-a-car companies. (Avis acquired Zipcar in 2013.) In January, GM announced a $500 million investment in Uber rival Lyft. Add to these a slew of independents, such as City Car Share and Getaround. One of the reasons it’s hard to make money is that the giant car-rental chains and major automakers have a lot of overhead, such as offices and drop-off centers, that cut into margins. (Hertz last year shut down its U.S. sharing business.)

BMW is hedging its bet on DriveNow by testing peer-to-peer sharing. Any owner of a Mini (a division of BMW) can rent out her car while she’s not using it. The business model is appealing because the owner who rents her car has no extra overhead (think Airbnb for automobiles). BMW provides the app to connect the Mini owner with the renter and the insurance and collects a fee on the rental. It’s too early to tell whether this concept has wheels.

BMW doesn’t view car sharing as a big threat to its sales. In fact, sharing could expose young drivers to brands they otherwise couldn’t afford. BMW says its DriveNow members in Germany are on average about 30 years old, while buyers of its Mini brand are about 40 and BMW customers about 50. The hope is that when those thirtysomethings become old enough to afford a BMW, they will already know and like the brand.

Because it takes four or five years to design and build a new car, BMW’s next generation of smart vehicles won’t start hitting the showrooms until the beginning of the next decade. Until then the company will do just fine selling its traditional luxury vehicles. The real question is, Will the decisions it makes today be radical enough to compete in the long run with aggressive Silicon Valley players like Google, Apple, and Tesla? BMW’s surviving another 100 years depends on it.

A version of this article appears in the March 1, 2016 issue of Fortune.

Sign In

Get

Thank you for your interest in licensing Fortune content. Please find information on various licensing contacts below and choose the one that best suits your needs:

  • 1. To license Fortune articles, excerpts, or headlines for republication in various media (including books, eBooks, film, web, newsletters, newspapers, magazines and others), please email syndication@timeinc.com.
  • 2. To license a Fortune cover, order reprint or e-print copies of an article or cover, or license an accolade, please contact PARS International at www.timeincreprints.com.
  • 3. To license text only photocopies of Fortunearticles as print or digital handouts in academic settings, or in academic coursepacks, please contact the Copyright Clearance Center at www.copyright.com