Big changes are coming for the automobile industry, and everyone in the industry knows it. This week, Ford and the Chinese technology company Baidu announced a $150 million investment in Velodyne, makers of powerful LIDAR sensors that are widely used in self-driving cars.
It's only the latest in a long sequence of deals linking car and tech companies together over the past year:
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/6958073/TECH-deals2-1.0.jpg)
The proliferation of deals represents a growing realization among car companies that they are going to need help navigating the major changes of the next decade — Ford says it wants to start offering fully self-driving cars by 2021.
The auto industry is facing three big innovations — car sharing, battery-powered electric vehicles, and autonomy. By itself, any one of these shifts would represent a significant but manageable challenge. But the real problem is that all three trends are converging, and they jointly represent an existential threat to today's dominant car companies.
Threat #1: car sharing
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/6553227/495300544.jpg)
So far, car sharing is the innovation that has had the biggest practical impact on the way people get around. Ride-hailing services are still limited to a narrow elite — just 15 percent of Americans had ever used Uber or Lyft in 2015 — but investors are betting that they will grow quickly.
The big question for the auto industry is whether ride-hailing services will start to displace car ownership as the primary way people get around. Even before Uber and Lyft came along, there were a few areas in America — like Manhattan — where it was common for people to forgo car ownership and get around using mass transit and (for the more affluent) taxis.
Uber and Lyft are making this lifestyle both cheaper and more convenient, which could cause more families to give up their cars. Short-term car rental services like Zipcar are also helping people get by without owning a car, knowing that one will be available to them in a pinch.
By itself, the shift to greater car rental isn't necessarily a problem for the car industry. Someone still has to manufacture and sell the cars that Uber, Lyft, and Zipcar are renting out, and there's no reason Ford, GM, and Toyota couldn't be the companies that do it.
But selling cars to ride-sharing drivers or the corporate fleets of companies like Zipcar is a different business than selling cars directly to consumers. And Detroit benefits from an incestuous relationship with conventional car dealers.
For example, one of the big barriers that has held back Tesla's growth is the network of dealers conventional car companies use to sell their vehicles. Not only do conventional car companies have a lot of experience selling cars through these dealerships, but state laws often require cars to be sold this way, creating a barrier to entry for startups like Tesla.
In a future where on-demand car rental is the default way people get around, the key to success in the auto industry may be winning big orders from companies like Zipcar, Uber, and Google. And that could create openings for companies focused on creating different kinds of vehicles and selling them in new ways.
Still, car sharing alone doesn't pose a big threat to the auto industry for the simple reason that today's ride-hailing services are too expensive for mass adoption. If you live in a suburban area with plentiful parking, it's cheaper and more convenient to buy a car and drive it around yourself. And even many people who live in high-density areas and don't have a car rely on mass transit (and walking and biking), because these options are more affordable than hailing a ride.
Threat #2: electric vehicles
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/6553237/490597614.jpg)
Internal combustion engines are marvels of engineering, channeling rapid-fire explosions to power a vehicle as it moves down the road, and modern car designs are deeply influenced by the strengths and weaknesses of internal combustion engines. Conventional car companies have expertise not only in engines but also in transmissions, radiators, gas tanks, and more — and a lot of that would be rendered obsolete if cars were powered by batteries and electric motors.
This is the basic premise behind the creation of Tesla. CEO Elon Musk and his financial backers are betting that a company specifically created to build electric vehicles will be better at it than a conventional car company trying to make the switch from gasoline to electricity.
Yet so far, electric vehicles haven't seemed like much of a threat to conventional vehicles. In 2015, there were 17 million new cars sold in the United States. Of these, just 115,000 — fewer than 1 percent — of them were battery powered or plug-in hybrids.
The basic problem is that electric vehicles don't provide a very compelling value proposition for ordinary consumers. They might save money on fuel, but they tend to cost significantly more than gas-powered equivalents — even after you factor in generous government subsidies. With long charging times and relatively few charging stations in most metropolitan areas, they're significantly less convenient than conventional gasoline-powered cars.
Threat #3: automation
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/6553241/462675492.jpg)
Self-driving capabilities would represent the most radical change in car technology in decades — and the biggest threat to conventional car companies. Making cars drive themselves is mostly a software problem, and software has never been Detroit's strong suit. Google is investing heavily in its self-driving car program, betting that its expertise in software development will allow it to claim a major role in the car industry of the 21st century.
Yet even here, conventional car companies have managed to keep their heads above water by outsourcing the hard technical problems to third parties.
A bunch of car companies — including Tesla and several conventional firms — have begun offering cars with "advanced cruise control" or "autopilot" capabilities. And several of them have purchased their technology from Mobileye, an Israeli startup that builds car sensors and software that allows cars to stay in their lane and avoid hitting the car in front of them.
Detroit's vision is that this self-driving technology will gradually get more and more sophisticated. Cars will continue to look more or less like they do today, but over time you'll have to grab the steering wheel less and less frequently. Eventually, grabbing the steering wheel will become so rare that it will make sense to sell cars with no steering wheel at all.
If that's how the market will evolve, it's not such a scary prospect for incumbent car companies. They have decades of experience integrating third-party components into their vehicles, and their existing manufacturing and distribution facilities will give them a big advantage over companies that try to start building self-driving cars from scratch.
The big threat to Detroit comes from a combination of trends
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/6553299/461479642.jpg)
So incumbent car companies would be able to cope with any one of these three trends taken individually. The problem for Detroit is that these changes are not going to come one at a time. They're happening all at once, and each of them is likely to accelerate and magnify the impact of the others.
For example, it's hard to make electric cars compelling to consumers, but it should be easier to make them attractive for car-sharing services. A ride-hailing car doesn't need the 300-mile range of a conventional gasoline-powered car. It just needs enough power to get through a morning of driving; the driver can then recharge it during his lunch break.
So electric cars purpose-built for sharing use might have smaller batteries, which can reduce vehicle weight and further improve energy efficiency. And of course ride-hailing drivers care a lot about the energy efficiency of the vehicles they drive, since a gas-guzzling car will cut into their earnings.
Self-driving technology will magnify this effect still further. As long as taxis need human drivers, they're going to mostly be sedans with room to carry three or more passengers. Nobody would want to drive a two-seat taxi and miss out on fares because he only has one passenger seat available.
But once you get rid of the driver, much smaller and lighter vehicles become viable. Lots of people take taxi rides alone, and they could take these rides in tiny one-passenger electric vehicles. These lighter vehicles would be more power-efficient and require smaller batteries, which in turn would further lower the cost of taxi service.
The optimal one- or two-seat electric car is likely to look significantly different from the best five-seat sedan — especially if the car doesn't need a driver's seat. So the shift to self-driving, electric ride-hailing cars will create a bigger opening for new companies to disrupt the car market.
And of course, with improved efficiency and no driver, these automated taxis will cost a lot less to run. That will bring them into financial reach for many more customers, dramatically expanding the market for ride-hailing services.
At the same time, the shift to ride hailing will help accelerate the shift to self-driving technology. One big reason for this relates to mapping. Right now it's widely believed that Google has the world's best self-driving car technology, and Google's technology depends on having extremely detailed maps of every street where a Google car drives. Uber has traditionally relied on Google's map data, but it recently announced plans to create its own independent mapping data.
If you're trying to sell cars to consumers, that's a huge problem, because it means you have to map the entire country before you sell your first vehicle — no customer is going to buy a car that refuses to drive in rural areas. On the other hand, if you're renting cars, it's not such a big problem. You can roll out your car rental service one metropolitan area at a time. Nobody minds renting a taxi that will only go to destinations in the same metropolitan area.
Renting self-driving cars will have other advantages too. Companies that write self-driving software will be concerned about safety (and the liability that comes with accidents). Maintaining ownership of the vehicles will allow them to guarantee that they get regular maintenance and that defective parts are replaced promptly.
So there's a good chance that fully self-driving cars will initially be available only for use as part of a ride-sharing service. If that happens, it will undermine the advantage provided by car companies' dealership networks and experience selling cars directly to consumers.
Electric, self-driving cars designed for a rental market are likely to look dramatically different from today's cars. With no driver and less space needed for an engine, we're likely to see companies experimenting with different shapes and configurations — rear-facing seats, built-in desks and mirrors, big touchscreen displays and televisions, perhaps even beds for long nighttime trips.
Why you shouldn't write off Detroit yet
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/6553309/506261184.jpg)
So conventional car companies are going to face some serious competitive threats over the next two decades. But it would be a mistake to write them off, because they have one big advantage over Silicon Valley upstarts: They know how to manufacture millions of cars.
Tesla is the only Silicon Valley company to even attempt to manufacture cars at a significant scale, and it's been learning the hard way that it's really difficult. Tesla CEO Elon Musk has become notorious for announcing production delays and cost overruns.
All three of Tesla's cars so far came to market later than originally announced, and Tesla recently announced it was raising $2 billion to fund further expansion of its manufacturing facilities. If all goes according to plan — and it probably won't — that will allow Tesla to produce 500,000 cars a year, about 3 percent of US car sales in 2015.
Manufacturing a big, complex object like a car is difficult, and conventional car companies have a 100-year head start on learning how to do it. That explains why Silicon Valley has been as eager to partner with Detroit as Detroit has been to partner with Silicon Valley.
The big question is how these partnerships will work. Will Google, Uber, and other Silicon Valley giants design the cars and rely on their Detroit partners to assemble them, much as Chinese companies like Foxconn assemble smartphones for Apple and Motorola? Will car companies treat Google as just another supplier and Uber as just another distribution channel? Or will a new generation of companies emerge that combines the strengths of both sides? Companies in Detroit and Silicon Valley are spending billions in the hope that it will help them come out on top.