The Automotive Electric Vehicle Transition

The only really interesting part of the automotive industry is the electric vehicle (EV) segment. These vehicles are also called NEVs in China for “new energy vehicles”. The reason that I say this is the only interesting segment is because it is clear that the whole world is moving fast to EVs and internal combustion engines (ICEs) will decline. One challenge for traditional automotive manufacturers, OEMs in the terminology of the industry, is that they all have a strategy of milking their ICE portfolio as a cash cow and using the money generated to fund the huge cost of developing an EV portfolio. The reason that this has a huge cost is that, just like Tesla went through a decade ago, they will have to build new factories and run them way below scale as they work out the difficulties of manufacturing electric vehicles. After all, the key skill that traditional OEMs possess is building ICE engines, which, of course, is a useless skill with EVs.

Other skills needed for EVs are software development, the use of advanced electronics, and perhaps even designing their own integrated circuits (ICs). The main useful skill they have is stamping metal and painting it. Since EVs have regenerative braking (charges the battery when braking), even seemingly similar elements of the car are different in ICE and EV. Surprisingly, last year, Toyota had to recall its bZ4X EV because the wheels were liable to fall off. That is one area of the car I would have guessed would be unchanged, but perhaps the much higher torque of an EV affected the hub bolts.

There is also a long laundry list of other capabilities that the traditional OEMs will need to acquire or develop, such as over-the-air software updates (OTA), since traditionally a vehicle’s code (in the ECUs) was never updated after start-of-production (SOP). If an update was really necessary, then it required taking the vehicle to the dealership. As an example of how OEMs struggle with this sort of thing, on one YouTube channel I subscribe to (E for Electric), the creator owns a Volkswagen ID.4 EV. It has a slew of software problems and supports OTA…but in over a year, VW has yet to make a single OTA update.

I see two major challenges to traditional OEMs. First, they have a lot of capabilities that are now redundant and lack a lot of capabilities that are now a necessity. I think one of the biggest issues is hiring talented software developers in competition with big tech companies. As someone put it at last year’s AEK conference, “Would you rather work for General Motors or Google?”

But an even bigger problem is the strategy of milking the cash cow of ICE vehicles to fund the investment in EVs. There are two issues with this. One is that most OEMs are very late to the EV party. Toyota famously resisted the basic idea, and maintained that hybrids were all that was needed, and only after a change of CEO has that changed. Stellantis drove a very nice-looking truck onto the stage at CES (see my post CES 2023: AMD, Stellantis, Cadence, and More) but it is just a concept vehicle and the actual vehicle will not be available until 2026. Since their main competition is Ford, who already has an EV version, the F-150 Lightning, of the most popular truck (and second most popular vehicle), the F-150, 2026 seems very late to start to compete. The second aspect of the cash-cow strategy is that ICE vehicle sales could fall dramatically faster than the OEMs seem to anticipate, leaving them with no revenue stream and massive investment needs. For example, Ford currently is profitable since it sells so many ICE vehicles, but it forecasts that it will lose $3B on EVs this year (I heard someone say that it currently loses $20,000 per vehicle sold). In fact, only Tesla (and BYD in China) make money on EVs.

General Motors announced that it is going to discontinue the Chevvy Bolt. That will actually improve its profitability since it loses money on every Bolt sold. It is also an old and somewhat obsolete design. Nevertheless, GM sold 38,000 of them last year. But surely the most money-losing vehicle of all must be the Hummer EV. If it is not a misprint in the tables reporting the numbers, GM sold two of these last quarter. Another money sink is its self-driving subsidiary Cruise. Cruise just announced that it now provides robotaxi service across all of San Francisco 24 hours a day. But that comes at a cost: Cruise loses about $500M per quarter.

You may have heard that Tesla made some significant price cuts to its product line, first in China, then in Europe and the U.S. I believe this was partly to get the price below the threshold to get a $7,500 (or $3,250 in some cases) tax break in the Inflation Reduction Act. But also for the usual supply-demand reasons that they want to sell more cars. Although Tesla says it doesn’t think about the competition much, it also puts a lot of pressure on all the traditional OEMs to lower their prices and lose even more money on every vehicle. Or keep their prices high and not sell many. From the beginning, Tesla has been very explicit that their strategy is to drive manufacturing costs down and pass those savings to the customer in the form of cheaper prices. This operates on two levels. One is the product mix. Tesla started with the Roadster, at a very high price point (around here you’d see them on Sand Hill Road where many of the venture capitalists have their offices). It used the profits from that to develop luxury vehicles. Then used the profits from them to develop more mainstream vehicles, the Model 3 and the Model Y (California’s #1 and #2 best-selling vehicles). The next step is to use current profits to manufacture what pundits call the Model 2, a $25,000ish smaller vehicle that will be manufactured in Monterrey, Mexico, in a new Tesla factory announced a few weeks ago.

The big challenge for any EV manufacturer, as it was for Tesla a decade ago, is to get to scale. Just like a semiconductor fab, a vehicle factory needs to run close to capacity to make money, otherwise the fixed capital costs are a killer. There is also the learning required. Tesla built its Shanghai factory in 10 months, and ramped production very fast since it already knew what it was doing based on its experience in Fremont. Ford or Toyota simply cannot do that. As Elon Musk has said repeatedly: “Building a prototype is easy. It’s not building the machine, it’s building the machine that builds the machine.”

If it seems like I’m a Tesla fanboy, I’m not. I own a Mini Cooper Convertible. The only time I’ve been in a Tesla was to drive it six feet to a charging station in the Cadence parking lot to make a Sunday Brunch video. I’ve never driven one, or even been in one, besides those few seconds. In fact, I’ve never been in an electric car at all, even as a passenger. However, I’ve attended Automobil Elektronik Kongress in Germany for half-a-dozen years, where senior people from all the usual suspects present (BMW, Mercedes, VW, Porsche, Bosch, etc). But increasingly there have been presentations about software.

Last year, the conference theme was “Toward the Software-Defined Car.” One of the presenters interviewed people in the street about whether they were interested in a software-defined car, and they had no idea what was being talked about. The presenter begged the industry to come up with a better term more like “smartphone.” Another presenter raised the possibility that, in ten years’ time, some of the companies presenting would not be there. You can read my more detailed coverage in my posts Automobil Elektronik Kongress 2022, AEK: Semiconductors: The Base for the Software-Defined Car, and AEK: Powerpoint Is Easy—Change is Hard.

Here’s one paragraph from the last of those three posts. This is a quote from Riclef Schmidt-Clausen of CARIAD, Volkswagen’s software subsidiary:

Will we be here again in 5 to 10 years? Will our companies be here? We don’t know. That’s what I’m going to talk about. It’s our jobs at stake here.

I have the same feeling, that most of the OEMs know where they need to be but are pretty clueless about how to get there from both a technical and business point of view. Here are a couple of quotes from the Associated Press about an interview with Mary Barra, the CEO of General Motors, from last year:

…Barra said she’s confident GM can unseat Tesla with higher-priced specialty vehicles, and it will beat Elon Musk to high-range EVs at prices that people can afford. Last year GM sold just 25,000 electric vehicles in the U.S., less than one-tenth of the estimated 352,000 sold by Tesla. …Already the company has pledged to cut the starting price of the Chevrolet Bolt small SUV to around $26,000 later this year.

Actually, GM canceled the Bolt instead.

Back in 2018, this is what she said:

And this is just the beginning. We are well on our way to bringing at least 20 new all-electric models to market by 2023—our next step toward a zero-emissions world.

I don’t think that will happen, although obviously, 2023 is not over yet.

As I titled one of my posts from AEK last year: PowerPoint is easy; change is hard.