New Technologies: Disruptive or Evolutionary?
Tesla and other electric vehicle ventures are promising to displace the internal combustion engine. Then there is autonomy, self-driving cars that will eliminate accidents and make commuting painless. Finally, proponents of car-sharing services—Mobility 2.0 firms such as Uber, Lyft, ZipCar, and Autolib’—claim that they will make personal vehicle ownership a thing of the past. However, the rollout process will take decades. Furthermore, these extend existing technologies and business models. They are evolutionary, not revolutionary, and so will remain the province of today’s core firms.
Let us start with car sharing. By the 1910s taxis were omnipresent, while Tokyo imported Model T buses in 1923 to replace its streetcars. Sharing may have declined in the developed markets over the past 50 years, but is still the norm in poorer regions of the world. By offering convenience, reliability and lower prices Uber and others have expanded the market, but to date they have captured most of their business from existing sharing services such as taxis and rental car companies. Their impact is thus incremental. A large taxi company can buy a hundred cars a year; Uber may buy tens of thousands. That has the interest of OEMs. But Enterprise, the largest car rental company, buys a million cars a year. At one time car companies thought it important to be in that business—Ford owned Hertz, VW owned Europcar. They concluded that was a mistake, and exited. Investing in new mobility providers may likewise provide minimal benefits.
Car companies and suppliers are investing large amounts of money today in these new technologies. All are developing EVs and advanced driver assistance systems (ADAS) as part of their move to autonomy. Several are investing in new mobility, such as GM’s investment of $500 million in Lyft. However, there is no evidence to suggest that big winners or losers will emerge. Unfortunately, nor is there evidence that these technologies will expand the demand for vehicles—indeed, proponents claim both autonomous vehicles and sharing services will shrink the overall market. If so, then investing in these technologies is a game the car companies can only lose: spend big today, but reap no benefits tomorrow.
The New Technologies
As noted in Chapter 2, a little over a century ago electric vehicles were the largest segment, and electric taxis remained on the road in Paris into the 1920s. Electric motors are simpler mechanically, and are intrinsically more efficient, for example they have no need for a radiator to dispose of waste heat. The main challenge is that despite an intervening century of R&D, the energy density of batteries remains low and the cost is high. There is steady progress, but contrary to expectations of two decades ago, that has also been true of ICEs, which are now smaller and more efficient, and continue to improve. As a result, EVs do not present a good value proposition for the ordinary driver. Furthermore, whether they are environmentally sound depends on how electricity is generated. When the source is coal, electric cars may more polluting than gasoline vehicles, though EVs can move emissions from inside urban areas such as Beijing to areas remote from big cities.
Autonomy is a natural evolution of safety technologies. Virtually all accidents are the result of driver error; technology can today assist drivers, but the biggest benefits will come from eliminating drivers. In the 1990s firms such as Delphi were already working on elements of a passenger compartment “cocoon” with sensors monitoring the outside of a vehicle and airbags surrounding the people within. By 2005, radar, lidar, and vision systems had all been commercialized, as had electronically controlled acceleration (for cruise control) and braking (for stability control). E-steering was also common in small cars, eliminating the weight and parasitic losses of “always-on” hydraulic systems. The physical components needed for a computer to drive thus are already on the road in one or another context. Similarly, car navigation has long been built into every smart phone, if not every vehicle. Pulling these elements together requires cheaper computing power—already available from multiple suppliers—better maps, and a lot of work on software. Autonomy is now technically feasible, at least in good weather, but outside of limited markets such as truck convoys, the path to commercialization is not yet clear.
Mobility 2.0 rests upon a seemingly compelling business case. We have perhaps $3.5 trillion in assets in the NAFTA “park” of 300 + million registered vehicles. Few are driven more than an hour per day. If firms can monetize 1 percent of those assets, then $35 billion is in play, with cell phones, credit cards, and cloud computing as enablers. The sharing space is crowded, from the taxi-like services of Uber to Autolib’ in Paris, whose electric cars are available for short-term rental at charging stations on streets all over the city. As noted above, shared mobility has been around for over a century; these players are operating in existing markets. They will not affect new car sales in the near term, and if their cost and convenience increases the overall market, their own demand for vehicles will offset the impact on demand by consumers who forego owning a car.
Over time mobility and autonomy may converge; experiments are already underway. Autonomy will be easiest to implement on short, heavily-traveled routes. That is also a context in which ride-hailing services work well. Furthermore, EVs are a good fit for the distances involved. But this represents only a portion of the demand for vehicles, and there is no particular reason to think that Toyota or GM will not be major players.
The Challenge of Rolling Out New Technologies
We believe that the long run will be a world of all-electric vehicles, but that will not happen quickly. Government policy can accelerate that transition, through the provision of charging infrastructure and subsidies to initial users. But such policies have distinct limits. Rebates on 50,000 vehicles a year are one thing, but governments are likely to balk at providing subsidies on millions of vehicles a year. Similarly, showcase charging projects can fit within government budgets, but legislators will hesitate when they see the price tag for building a nationwide system while beefing up residential power grids.
Those are not the biggest issues. As suggested above, one problem common to all new vehicle technologies is the “park,” the number of registered vehicles. In the U.S., about 12 million vehicles are scrapped a year, while 17 million are added. Put those numbers into a spreadsheet, against the “park” of 300 million vehicles. Even if in 2020 a full 100 percent of new vehicles are electric and autonomous, it will take at least 10 years before such vehicles account for half the cars on the road. But new technologies cannot be rolled out that quickly. First, they have to be designed into vehicles. There is a long lead time. By early 2017, the drivetrains and sensor systems for model year 2020 will already locked into place. In addition, making major changes such as a shift to an electric drivetrain can be done only when a platform is redesigned, and that takes place on a rolling basis over the course of 8-10 years. Ford’s F-150 pickup is fairly new, so may not be changed until 2022 or later, but it is the biggest selling vehicle in North America. In addition, it will take many years for the supply chain to go from the ability to supply batteries for the sale of 20,000 EVs to turning out 20 million, a thousand-fold increase.
The industry understands this well. So while pistons will not be needed in an electric vehicle, the suppliers that make them, such as Federal Mogul and Mahle, are working on technologies that they do not expect to launch until the mid-2020s. Indeed, given a growing global market, they continue to invest in new capacity because they expect to be making more pistons in 2030, not fewer. The rollout of EVs, autonomy and mobility technologies will take decades. Their impact will be gradual, not disruptive.
Increasing (Marginal) Costs, Decreasing (Marginal) Benefits: Is There a Route to Commercialization?
Adding additional safety features adds costs. Adaptive cruise control with automatic braking can be attractive, because it makes driving in congestion easier, and lessens the chance of a rear-end collision. How much will it be worth paying to add automatic lane keeping? To add full autonomy? The benefits of new features surely diminish. As a commercial proposition there may be no natural progression from advanced safety features to partial and then full autonomy. Falling costs may help—once software is developed, adding it to an additional vehicle may cost nothing. New technologies can also first be launched on high-end vehicles, for which demand is less price-sensitive. Indeed, carmakers are already bundling elements of autonomy in high-end vehicles as a measure of product differentiation. Similarly, “light” electrification—mild hybrids, e-steer—improves the performance of traditional ICE vehicles. The business case for incremental electrification is clear, while that for full EVs remains problematic.
More generally, vehicle technology choices are deeply embedded in interconnected social structures, from where we live, work, and shop, to how roads and fueling and legal liability are organized. Changing one piece of the puzzle by itself can be hard. Electric vehicles are more attractive if charging facilities are widespread, but investing in charging facilities makes sense only if electric vehicles are numerous. Habits must adapt, with drivers realizing that it can be practical to have an EV for daily use, and use rental vehicles or other alternatives for the infrequent longer trip. Similarly, autonomy and mobility can reduce congestion, but that faces the hurdle of workdays that push commuters into a narrow timeframe, the segregation of land into residential use and business use that pushes people onto the same handful of roads, and a historic unwillingness to share a ride. In fact 10-hour, 4-day weeks and early- and late-starting times are already widespread; the easy adaptations have been made. The basic point is that there are multiple interdependencies in complex systems. Shifting those adds to the timeframe required for commercialization, and lessens the real-world benefits of conceptually attractive individual technologies.
Conclusion
From the OEM perspective, the key question is whether new technologies will increase total revenue. EVs merely replace an existing power source, and will not in themselves increase the demand for vehicles. In contrast, the new mobility providers claim that their services will lower the demand for cars, though some of this would be at the expense of public transport, not car ownership. But these same firms also point to underserved populations, such as the elderly and the poor. The net impact is thus likely to be small in volume terms, and we believe it unlikely that consumers on the whole will be willing (or able) to pay more for autonomous vehicles.
The industry is thus trapped in a prisoner’s dilemma: no individual car firm wants to fall behind, so all end up investing. All end up worse off: OEM investor presentations already note that costs are higher today. Meanwhile with all in the game, any improvement in revenue remains modest, and lies far in the future.
Individual software and component suppliers may do better. They are better positioned to link R&D to discrete projects that improve individual components, step by step. But the hurdles are great for would-be new players. On the hardware side components must operate from the –40°C [–40°F] of Quebec to the 60°C [140°F] temperature of a car in the Nevada sun, all while withstanding vibration comparable to dropping a cell phone on the floor continuously for a week. They have to keep working for 15 years, with defect rates of at most single-digit parts per million. The auto industry has protocols for verifying such performance, and incumbents know how to navigate the technology adoption and production approval process. In addition, for firms such as Delphi, Denso, and Bosch, the technologies of EVs and autonomy are extensions of existing product lines. There will be new players, but there has already been a steady ebb and flow of suppliers over the last three decades, so in relative terms the next decade will not be exceptional. Unlike for car companies, for the supply chain as a whole these technologies may generate profits. However, the slow pace of the “revolution” means that for most firms the near-term contribution to the bottom line will remain modest.
What of new entrants as OEMs? Cars remain complex, assembled goods, with challenging distribution channels and service needs. It is with good reason that in 2015 the top six firms each sold over 6 million vehicles, with production centered in the Auto Alley of the United States and the Auto Corridor of Europe, with a comparable geography likely to arise in China. Tesla, the most visible entrant, has yet to demonstrate a path to profitability, or a way to quickly expand the scale of its operations. In contrast, the successful new entrants in China fit the mold of existing car companies in structure and technology. Any new entrant must compete with the major global OEMs, all of which have active programs for autonomy, electric vehicles and the new mobility, paired with the ability to bring these into volume production and feed them through distribution channels to end consumers.
Even in geography, technology will not be disruptive. Car companies are indeed setting up outposts in Silicon Valley, and enlarging their in-house venture capital funds. But actually incorporating new systems into vehicles requires close cooperation between OEM engineers and those at multiple other suppliers. While it captures fewer headlines, Silicon Valley is thus also setting up technical centers in an arc stretching from Toronto through Ann Arbor. That is where most of the R&D jobs will end up.