The state of the car industry, or “What happens next?”
I’ve spent a lot of time reading about cars lately. It’s a fascinating industry right now, full of transformation. So here’s my take on the big challenges, how the industry will look in the future and what products and services would make the car experience better.
TL;DR
- Electric cars are here and will bring more suppliers inside the car companies’ value chain. This will likely change their unit economics, but even more so will transform their factories into assembly lines putting together bits and pieces from here and there. Design will be the only difference.
- Self-driving cars are not here quite yet but they’re coming. Companies like Google’s Waymo and Nvidia will provide car companies with off-the-shelf solutions, a bit like the Android OS for smartphone manufacturers.
- Autonomy will make the car industry switch to a platform model. It’s a winner-takes-all market, so companies like Uber are the best positioned. Tesla is building an Apple-esque position by mastering software, hardware and factories.
- Other kinds of mobilities will exist. These will be specific platforms (like Heetch in France, concentrated on the youthful nightlife market) or they will be based around hobbies: non-self-driving, vintage cars, etc. Many car companies will try to make the switch, some may become suppliers, many will die.
- Building a car will be easier, new kinds of mobility will emerge and entrepreneurs have many, many opportunities available.
A tale of old lads and fresh faces
When I was young and foolish, I thought my future would be at regular company, in some sort of management position — whatever “management” means. So when I had to do an internship, I thought about what I liked and decided it would be fun to do something related to cars. And that’s how I signed up for a year-long internship at Renault.
I love Renault. My father drove a V6 Espace MPV for years and during the annual breakdown (the gearbox would break every January, because apparently those 153hp were just too much for it to handle), I could enjoy our visit to the local retailer. In 2001 I had a shock when I set foot in the Avantime. Call me crazy, but for younger me that car was the future.
So going into my internship, I saw Renault as the brand that invented new kinds of cars (they invented the MPV in Europe, then did it again with the Twingo, a.k.a. the best city car of the 90s, etc.). That lasted for about an hour into my first day. There, in that grey building they called an office, my new boss told me: “Our job is to shape metal to make cars out of it.”
Let’s call that their motivational speech. But it does illustrate what traditional car companies are: shapers of metal. The end. Just a few weeks after that, I got another wake-up call, this time from Renault’s media system. See, back in 2014, Apple was trying to put their CarPlay system into all cars. When they approached Renault, they were sent home because Renault’s engineers found that their system was better. They felt it offered “more functionalities”.
Now, anyone who has ever tried this famous R-Link can tell you how horrible it is. The problem is that most traditional car companies build their cars from an engineering standpoint, going down the to-do list. Some premium brands still pay attention to the materials they use, but they are an exception. Most classic brands are no longer doing any comprehensive product design, instead just bringing together new bits here and there.
(Note) As I was writing this, Benedict Evans released a great post about how cars of today look like the features phones of 2007:
And that strategy makes sense due to their cost structures: they don’t build all of the car themselves, relying heavily on various suppliers. It’s not that there’s no product design; it’s that it tends to be incoherent.
As an example, let’s take the high-end Citroën DS brand. When the original DS was introduced in 1955, one distinctive feature was the pneumatic suspension, which offered the most comfortable ride on the market. Yet 50 years later, Citroën ditched that suspension for cost reasons instead of pushing forward its value in the eyes of the customer. Because Citroën failed, for decades, to market this innovation, they sent a valuable asset to the dustbin.
The car industry is full of peculiar little stories like that. And maybe they wouldn’t be a problem if the system wasn’t so crooked. But the truth is that fewer and fewer people are buying cars, much less buying new cars. Having a car has become a nightmare, both in terms of costs and logistics. Between gas prices, insurance, taxes, mechanics and parking, having a car is a pain in and of itself.
So if people don’t really care for metal-shapers anymore, how is the industry compensating for declining sales? First, by lowering the price. Competition has gotten so crazy that most conventional car companies (i.e., the ones that are not premium) have margins of 2-6%. Even though the average basket is rather high, that is a very low rate.
Second, most of these companies pay attention to vanity metrics, such as the number of cars registered. That’s convenient because those brands rarely sell the cars themselves. They sell them through a network of dealers. And at the end of each year, they force them to register cars that are yet to be sold, making the numbers look better.
Finally, there’s the regulation problem. As pollution rules get tougher and tougher — thankfully — car makers have resorted to micro-optimization. So engines are designed for regulatory tests instead of daily use, a practice that, as we saw with the Volkswagen scandal, yields catastrophic results. Sometimes this becomes totally ridiculous: in France for one, diesel engines have been encouraged for years because they had a “particulate filter”. But that filter didn’t take into account the smallest, deadliest particles!
Given all of that, where is the innovation? Apart from small, specialized and very high-end car companies (Koenigsegg, Pagani, etc.), no big player has emerged since the second half of the 20th century — with two main exceptions: Tesla and Chinese car makers. I won’t talk a lot about the latter because a) my knowledge of the Chinese market is very limited and b) they have a low-cost approach, not really a paradigm-shifting one.
The big surprise (some even call it a black swan) of the last few years has to be Tesla. They were the first to make a good electric car. As often happens, they didn’t make any huge breakthroughs; instead, like Apple did in 2007 with the iPhone, they made the best use of available technologies. Tesla was the first cool electric car, the first with a big battery, a powerful motor and a decent media system.
But if it was so simple, why didn’t somebody else do it earlier? There’s a myriad of reasons, of course, but Tesla’s strength is essentially a) its supply network and b) its distribution network. Tesla is the first vertically integrated car company in ages. They design, build and assemble most of the car themselves, they’ve developed their own top-notch media system and they will soon start producing batteries in their gigafactory.
And in place of the traditional network of retailers, they have their own distribution channel. Even though it’s brought them into court in the US and means that they have no buffer to absorb over-production, that channel enables Tesla to sell with a higher margin (no retailer cut) and to have much more control over how they sell cars.
Hence two questions arise: 1) again, why didn’t anyone else do it before, if none of this involves serious technological breakthroughs and 2) why do traditional car companies seem unable to react?
Regarding the first question, things are simple to conceptualize but incredibly difficult to implement. Fitting a huge battery into a car that looks good and setting up electric charging stations powerful enough to support that battery requires lots of cash and lots of attention to the thousand little things that can go wrong. (And that is exactly why Barclays called Elon Musk a black swan ;-)
As for the second question, we need to go a little further into the challenges of the electric car.
Challenges of the electric car
Regardless of how successful Tesla is, it’s still far from being used by the masses. Obviously part of the problem is the cost. Not everybody can afford an $80k vehicle, even if gas is no longer an expense. This will change somewhat with the arrival of the Model 3 — already a success if we look at the 300k preorders. And although supply may be a problem at first, there is no doubt that Tesla can overcome production troubles.
The real problem is the technology, which still poses its own limits. The first challenge comes from the electric motor. While it has many advantages over the classic combustion engine, its major drawback is its energy consumption at steady speeds. Unlike a traditional engine, which sucks up less fuel when you’re driving on a highway than when you’re in the city, an electric motor doesn’t like to maintain a steady speed for a long time. So at exactly the time when you’re wanting to use highways and go further, the autonomy of an electric car plummets.
And when I say plummet, I mean “catastrophically plummet”. From a range of 250 miles you can jump down to a mere 100 miles. Add air conditioning or heating and things get worse. As a result, an electric car is at its best in an urban or suburban area where it changes pace frequently. Every time the car brakes, energy is generated and converted back into electricity that is then stored in the battery.
So the only fully electric solutions are to limit the car to urban areas or put in a very large battery. Interestingly enough, in the first category we find a lot of traditional companies (GM with the Chevy Bolt, Opel’s Ampera-e, Renault with the Zoé and BMW’s i3, for instance). In the latter, there’s Tesla. Of course, there’s also a myriad of smaller players but here I’m focusing on manufacturers that can handle high demand.
By the way, this is also why plug-in hybrids are so appealing today: you can use them in a purely electric mode in the city and switch to a conventional engine for long trips. The only problem is weight, since there’s a conventional engine and a battery. As a result, the battery stays rather small, to keep from making the car too heavy.
But this is just a temporary solution. Batteries will get more and more efficient at a given weight and size and we’ll eventually get rid of the traditional engine. But that doesn’t mean it’s simple. Higher battery capacities need more powerful chargers. Battery capacities are rated in kWh. To give you an idea, the best plug-in hybrid today gets around 10kWh; a Renault Zoé gets 40kWh; a base model Tesla 60kWh; and a high-end Model S 100kWh.
Tesla’s superchargers are rated at 120 kW. Which means a 90kWh Model S gets fully charged in 75 minutes and 80% charged in just 40 minutes (charging gets less and less efficient as time passes — see here to learn more). The problem is that Tesla’s solution is currently the fastest on the market, although a consortium of European brands recently announced a network of 350KW stations in Europe, with Tesla doing the same a few days later.
This should dramatically improve the fate of electric cars: we can see 200kWh batteries offering a 300-mile range on the highway with 30 minutes for an 80% charge arriving a few years from now. But there are two big letdowns here. First, how are we going to produce all that electricity? Nuclear seems to be the best option so far, but we hope that green energies catch up quickly (and solar roads are possibly a solution, if they prove cheap enough).
The second big issue is how these stations are deployed. Even though some private companies seem to be aware of the power issue, most chargers installed today are low-power chargers. For instance, Bolloré seriously invested in a 7kW (seven, as in, 50 times less than the project mentioned above!) to cover all of France. And let’s not even mention the state’s encouragement of that sort of thing.
Essentially, all infrastructures are not going to be equal. This is a big problem, especially when such scams are financed by public money. Even if there is more awareness of the need for powerful charging stations, there is still a lot of progress to be made, first of all by governments themselves (it’s the same story in the US, apparently).
And are these issues too big for traditional companies to overcome? Not really, but it’s clear that not everyone has taken the same steps in electric car development. The premium German brands should be able to fight Tesla with really good cars as soon as 2018 (see the Porsche Mission-E, for instance).
That will in turn benefit their more general brands (if Porsche has such a car in 2018, Audi should get it in 2019 along with Volkswagen, all at a more affordable price point). Renault has advanced as well with the Zoé. Even though they chose a tougher, non-premium market (where price matters more), they have the technology and it’s improving (they recently announced a 41kWh battery offering a 200-mile range).
Will these brands benefit from a first mover advantage? Maybe, maybe not. What we should note is that a) the technology is here, b) traditional players have started using it (and more competition will probably end up lowering prices) and c) infrastructure is still lagging behind, with mixed quality even as solutions begin to appear and to be deployed.
One last point is that electric car technology might reshuffle the game because it lowers the barriers to entry on the car manufacturer market. An electric engine + a battery is much more readily available and easy to get working than a traditional combustion engine. Car companies will change their factories into an assembly lines of various bits sourced here and there. Only the design and interior decoration will make the difference.
Besides, electric cars are more and more built around software. Making the elements work together is about integrating software. There may at some point be an Android for cars, or at least some sort of SDK (in fact, Renault has already launched something in this direction). This might eventually lead to more players both in car manufacturing and associated services (see how BMW views an AppStore today - there is still progress to be made). Some companies will go bankrupt, others will survive and new leaders will emerge.
The autonomous vehicle
The electric car is a big change but it’s almost nothing compared to the next big thing: how cars will drive themselves. There’s currently not much argument around the idea that autonomous vehicles will be driving in cities and on highways in 5 years, maybe 10 if something unexpected happens (or if governments are slow to pick up on the change).
John Zimmer wrote a very detailed article about how autonomous cars will change our lives: cities will get rid of parking, streets will be public spaces just like they were at the beginning of the 20th century and car ownership will be something of the past (why bother when you can order a car when you need it and send it away when you don’t?).
So what are the challenges and opportunities? Self-driving cars are sorted into 5 levels of autonomy, with level 5 being full autonomy without human supervision. It is clear that the company closest to a level 5 self-driving car is Google, with their autonomous pods. Even though their speed is limited and they run on a homogenous set of roads, they are working rather well.
To be clear, we’re not there yet. But we’re getting there. On highways, autonomous cars work and will probably be safer than human drivers by the end of the year. Even with accidents (like the recent Tesla crashes), the self-driving car remains statistically safer than human driving and things keep improving, quickly and all the time, thanks to machine learning.
Outside of highways, things might be a bit more difficult because the risk of hazardous situations is higher. And those situations are precisely the problem. Having a car drive itself on a track is something fairly easy to engineer. I mean, it’s been years since Audi sent a self-driving RS7 around the Hockenheimring.
Car companies have mastered a car driving itself around a racetrack at 150mph. Basically, for any situation that gives the human brain less than 1 second to react, a computer can be better. So all of our reflexes behind the wheel are covered; what’s left are the hazards on the road.
Mastering the road is mainly about software engineering. We have the hardware and it’s cheap: cameras, radars, processors… Even if the retail price of Nvidia’s latest chip is $15,000, there is no doubt it will eventually get cheaper, to the point of becoming a marginal cost for automakers.
The chances are high that at first we’ll have a hybrid self-driving car, with the human brain involved in only 20% of situations. Then that number will drop to 10%, then 5%, then 3%, 2%, 1%,... Just like when self-parking abilities arrived, little bits will keep appearing until we basically don’t need to touch anything. Some cars will provide more control, others less, depending on whether they are sports cars or family sedans.
Admittedly, we’re far from being at complete level 5 autonomy. Toyota’s CEO said it himself: “I need to make it perfectly clear, [full autonomy is] a wonderful, wonderful goal. But none of us in the automobile or IT industries are close to achieving true Level 5 autonomy. We are not even close.”. But even without arriving at that point, self-driving cars will soon be much safer than humans behind the wheel and road fatalities will surely decrease as a result.
While there are some difficult problems to overcome, self-driving cars are likely to prefer larger, faster, safer roads anyway. Autonomous vehicles don’t lag around after a red light. They don’t hesitate. Their speed doesn’t keep going up and down. In the end, their rise may increase the bandwidth of existing roads. Secondary roads may become even less popular, or even fall out of use.
What’s more, all autonomous cars are improving very quickly but some companies seem more advanced than others. Google is clearly ahead (they started early, too), followed by Tesla and the Volkswagen group. The necessary hardware is readily available, with Nvidia selling the PX Drive 2, used by Tesla for their Autopilot feature, for $15,000.
The real problem is software engineering. Overall, we lack talent to build software. Given that what talent is available mainly gravitates around Silicon Valley, Google and Tesla have an obvious advantage over other car companies (see recent, significant employees moving from Apple to Tesla). And some European brands (mostly German, like Audi and Mercedes) are demonstrating commitment to this need. Perhaps in the future, as technology progresses, software will become a commodity, too.
Google recently announced their Waymo kit will be available to any car maker, a bit like Android for phone makers. The byproduct of this situation is that if we don’t need humans to operate cars, they can manage themselves. They can drop us somewhere, then pick somebody else up, then come back and get us, etc.
That’s why owning a car won’t make sense anymore. It will be cheaper to pay a subscription or a per-ride fee. Plus, people are already used to signing up to use a car for a certain period of time. Leasing now accounts for a majority of sales in the B2C market (for new cars) and people are used to paying a fixed amount of money every month for their car.
Remember that besides using the car, there’s maintenance, insurance, gas, etc. What’s more, because a subscription car service would be constantly in use, with the vehicle being in use 95% of the time rather than parked, it would probably be way cheaper.
As of today, the cost of an Uber ride is roughly split into 30% for the car, 30% for gas, maintenance, insurance, etc. and 40% for the driver. Eliminate the driver and gas, lower the maintenance costs (an electric car being simpler, as we saw above) and insurance (a self-driving car doesn’t crash) and you end up reducing the cost of an Uber ride by almost 70%. At this point, it would be essentially as cheap as a subway or bus ride. Sounds far-fetched? Just read Tesla’s master plan:
In an nutshell, self-driving cars are coming — fast. At first, they will take the form of a partially autonomous vehicle (like Tesla today) and pretty soon there will be no human supervision needed in most cases. There might still be some tricky roads, but generally we’ll go from one point to another without driving.
What does this mean for established companies? It should likely be seen as a real threat. Indeed, the tech might be readily (and rather cheaply) available, but what matters is going to be the platform. Since people won’t own the car, they’ll be using it as a service.
But they’ll only use one. This is a quintessential winner-takes-all market. There won’t be multiple winning platforms because, unlike two cars from two different companies, platforms aren’t inter-operable. And the bigger the platform, the bigger the network effects and the economies of scale.
Uber seems to be in the best position to dominate, with Tesla and Lyft not far behind. Because of how well-funded these companies are and given their user communities, it may be as hard to push them away from their goal as it would be to go into direct battle with Facebook.
Obviously, anything can happen. Uber might need its Snapchat. If you feel like taking on such a mission, you should probably get started ;-) But remember that car manufacturers will simply be suppliers for the platforms. Given the intense competition, margins are going to get crushed and not many will survive.
A tale of two mobilities
Electricity & autonomy. It’s no wonder the car industry is living its biggest revolution since the Model T. Electric, self-driving cars mean manufacturers will be mere assemblers: batteries will be made in a giant factory (such as Tesla’s), autonomous hardware and software will be provided by Google’s Waymo (see this article: carmakers are still confused about autonomous vehicles), etc.
The only one that sets itself apart is Tesla, because they’re massively investing in the full-stack car company of the future. They’ll build the batteries, the self-driving functionalities and, of course, the cars themselves. Here again, there are echoes of the smartphone industry and the battle of Android vs. Apple.
Even more than that, these innovations redefine mobility itself. The most evident change will be the rise of platforms. This is actually a bigger threat to the car makers than technical challenges such as electricity or autonomy. It’s what will lead to reduced margins and more pressure on prices. If the Android example shows how it can be profitable for phone makers, it’s only because they’ve mastered distribution.
And that’s where the analogy leads for carmakers: they will become suppliers for platforms. The only exception may, again, be Tesla, which intends to develop its own platform (albeit one that is more high-end; but then they’ll be the Apple of car platforms rather than just being the Apple of cars). This makes the future a bit gloomy for them, but one has to consider that there are lots of different mobilities.
The first distinction is commuting vs. traveling. Most of the frustration in a car probably happens while commuting for obvious reasons: congestion, weariness, time… Of course, in a world where cheap electric cars drive themselves, many of these problems disappear.
Taking a car will be simple and affordable. No time will be lost commuting. People will work, play, and sleep in their autonomous car. And even if level 5 autonomy takes a while to arrive, driving will be a lot less tiring. Traffic jams will no longer exist, reduced first by better ride-sharing options and then by autonomy. And since we won’t own the car, we can take a small car to work and then have a big family van for going on holiday.
And that brings us to the second form of mobility: traveling. Some people do like to drive. Going somewhere on vacation or for the weekend is a whole different experience from city driving. And there are two ways of thinking about mobility for a trip: one is separate from the trip (mobility solutions are present when you arrive, and are available until you depart) and one is as part of the trip (mobility itself is part of the experience).
For the first, nothing can beat electric self-driving cars. Just like taking a train or a plane, you could open your platform’s app, call a car and be driven wherever you want. It will be cheap, comfortable and flexible. The only question is will there be one platform for urban and extra-urban mobility or will there be two separate platforms?
The second is seeing the trip as a journey in and of itself. This can be supported by entertainment in the self-driving car or… by getting back to the sheer pleasure of actually driving. That’s why high-end performance car companies will probably survive, alongside prestige/vintage brands.
In a sense, with the rise of the autonomous vehicle, secondary roads will be the new Disneyland. Driving, getting lost, looking at a map will become hobbies, not necessities. This is probably why we won’t see self-driving Ferraris (and they may not even have to transition to electricity). Just like Amtrak offers train trips, some car companies will specialize in offering driving experiences. In a nutshell, we could sum up future mobility like this:
Essentially, the automobile industry is undergoing a complete, tripartite revolution:
- The electric revolution: electric cars will be simpler to build and maintain. The technology is here, it’s getting better and cheaper.
- The autonomous revolution: the technology isn’t there yet but it’s getting better quickly, too. And it’s mostly a software issue, since the hardware is here and easy to install.
- The mobility revolution: people are already renting cars and taking cabs; the end of ownership will come when cabs will us anywhere.
For automakers, it means more and more of their business will be supplying platforms. Electric self-driving cars mean hardware will be similar across the industry. Batteries, cameras and motors will become like the processors, screens and cellular chips of our smartphones: readily available through a few market suppliers.
But the biggest change might be that automakers won’t even have a chance to sell their cars to the consumer anymore. If they become mere suppliers for the platforms, their margins will suffer and the pressure on prices will become unbearable, allowing only a few to survive.
So now they have a choice: either they become a full-stack platform (building the cars and the service — the Tesla path), accepting more competition and more risk; or they become a regular supplier that deals with having a low-margin business.
It’s likely that these changes are a bigger opportunity for car rental companies. But these would need to build out the needed IT to operate large fleets of vehicles and seamlessly serve the consumer. At least they do know how to manage a lot of cars… (by the way, that article gives an example of how low margins can go).
This revolution also means a lot of opportunities for new suppliers in the auto industry: Nvidia might be the Qualcomm of chips for self-driving cars, LG and Panasonic may provide the batteries, etc… Just like microprocessor factories are very profitable thanks to the smartphone era, that may be a great position for building a sustainable business for cars.
Finally, it also means new opportunities for Entrepreneurs. First, there are likely many other platforms, aiming at specific needs, that are waiting to be invented. Maybe there will be a new secret sauce that lets someone serve customers even better (again, look at how Heetch serves the youth market better than anyone in France). At the same time, difficulties remain with a winner-takes-all economy: there will probably be one big company that eventually leaves only bits and pieces to the others.
But as I said above, one way to envision mobility is to make it an experience in itself. There might be a lot of businesses here: building performance cars, tours in transparent, self-driving cars on secondary roads, renting vintage cars… Just like Disneyland offers entertainment parks all over the world, one startup might emerge as the leader in mobility entertainment.
Did you like this article? Click the green heart!
Special thanks to both Oliviers from O2 Quant, Gilles Barbier and Oussama Ammar for their discussions about autonomous vehicles and AI, Kyle Hall for the English-proofing, Julie Pathé, Quentin Audiffren and Alexis Feertchak for their help in grasping industry specifics and all of TheFamily’s team for our discussions and readings on the subject.
