Over the past few years the semis industry has become somewhat obsessed with autos. Every major chip company now dedicates a fair amount of coverage to the topic of cars in all their investor presentation. Or at least it seems that way. In part that reflects a genuine growth in auto semis content, and in part the tapering of growth in may other categories (mobile, PCs, etc). Often left unsaid is the fact that semis growth in autos is still fairly distant. Aside from companies who have been selling to auto makers for years, cars are still less than 10% revenue for most chip companies. And so it is difficult to draw up reliable forecasts for the segment. In fact, one of the most important questions about auto semis remains a big mystery – who is going to be the ultimate decision maker for auto semis decisions?
There are really two pieces to this question. First, how will cars be manufactured in the future and secondly, who will control the software in those cars.
Put simply, from our point of view, the auto industry looks like it is poised for significant change. Traditionally, the auto industry has a high degree of vertical integration, and while that has changed, for the most part today’s big auto makers still organize themselves around manufacture, or at least final assembly, of the cars they sell. This stands in marked contrast to the electronics industry where the ODM/OEM model has long separated manufacture from design, sales, and marketing. Auto companies still spend a lot, by electronics standards, on the capital equipment required to build auto assembly lines. But there are many companies looking to shift this model.
The leading example of this is Foxconn, or HonHai, the group that among many other activities, does final assembly for the iPhone. Hon Hai has been in the news a lot lately, they are now offering five models of cars ranging from sedans to pick-up trucks. The CEO of Foxconn has been saying publicly this week that he hopes someday to build cars for Tesla. We have lost track of how many pieces of the auto supply chain that Foxconn has built, invested in or acquired. One of these days we will take a deeper look at what they are hoping to achieve with this investment. For our purposes here the main point is that the company which epitomizes the abstraction between electronics design and electronics manufacture is now aggressively promoting its ability to replicate that model for cars.
It is by no means a foregone conclusion that Foxconn will succeed with its attempt or that this model will transfer to other auto makers. On Twitter, someone recently pointed out that Foxconn’s capex budget is much smaller than that of the typical auto OEM – in 2021 Ford spent almost $7 billion on capex, Foxconn spent less than $3 billion. And while Foxconn’s capex has increased significantly in recent years, the fact their investment spending is less than half of the maintenance capex budget of a single auto maker draws into question their ability to provide a manufacturing platform for multiple OEMs. Electronics assembly also tends to be highly seasonal, with labor comprising a much more significant share of the bill of materials for phones than cars, making much easier to stand workforces up and down. It is also important to recognize that Foxconn will not even start production until 2023. So file all of this under “Too Soon to Tell”. There are multiple forces working both in Foxconn’s favor and against it.
What is clear is that as much of the world moves towards electric vehicles (EV) the supply chain is changing significantly. The design of EV cars is much simpler than those of cars using internal combustion engines. The skills required for this work very much favor the electronics supply chain and without a doubt this opens up the door to significant disruption of traditional methods.
And then we get to the question of software. Buyers today, especially the crucial demographic of young buyers, care much more about electronics and the interior experience of their cars than previous, older groups who prioritized performance and speed. This falls neatly into the realm of User Experience which the electronics industry now does very well. Ultimately, this is a software question.
We have very little faith that the incumbent car makers can gain this expertise. They have had 20+ years to adapt to this reality but Apple’s Car Kit continues to spread across the industry. Automakers recognize that they need to control the software experience of their cars and do their best to keep Apple, and others, at arm’s length. We regularly hear mention of the Motorola ROKR as a cautionary tale for auto makers. Nonetheless, consumers want Apple-like experiences and the auto makers struggle to emulate that, and probably always will.
If we put these trends together into a worst-case scenario, Carmageddon, there is a real possibility that the auto makers fail to differentiate on their own software, and new supply chains emerge which greatly reduce the cost of assembly. This is what happened in PCs and mobile phones, where the software owners hollowed out the value chain with the help of low-cost, flexible Asian supply chains. Theoretically that could happen to cars as well.
To be clear, we do not think the industry is destined to go down this path. There are still significant differences between building autos and building electronics – including far higher upfront capital expenses, a highly-tuned manufacturing model, the current scarcity of battery manufacturing capacity, and a very different regulatory environment. But this does not mean that cars will be entirely spared from the changes coming.
As always, it is important to watch what is happening in China. There are roughly 50 EV makers in China today. Some of them are outsourcing manufacture, others are doing it in-house. Some are using stock software, some are designing their own software, and even their own chips. As much as China’s government tries to plan out much of its economy, it often leaves sectors initially open to Darwinian levels of competition. Obviously, the success of Chinese auto makers is going to depend on many factors, geopolitics not least among them, but all this experimentation is very likely to lead to entirely new models of production which could destabilize the global industry in important ways.
Photo Credit: The Simpsons