The semis content in autos has been a big topic in the press lately. All those reports about the current supply chain crunch go back to the example of the $70,000 car that cannot be produced due to the lack of a $2 chip. The reality of this is more complex, but it makes the point – semis are important for cars. Increasingly, the big chip companies have been talking more and more about their autos business. In an upcoming post, we plan to look more closely at some of those efforts, but before we get there, we think some background in semis for autos might be useful.
First, the market for automotive semis is big, but with certain limits. Depending on who you ask and how you count it, the auto semis market is about $50 billion a year. That is a little bigger than the market for mobile semis and a little smaller than the data center market, and roughly half of the PC CPU market. The world produces about 60 million to 70 million cars a year.
Second, the autos semis market is forecast to grow significantly, but it has been forecast to do that for a long time. Indie Semi, a recent SPAC focused on automotive semis put the average semis content in cars today at $310 per car (slide 7), by our math it is a little higher than that, somewhere between $400 and $600. And if you look at that same Indie slide, they forecast the content will grow to $975 “Tomorrow” and $4,000 in the “Future”. For most people, that intuitively makes sense, everything seems to be getting more semis and cars are getting better electronics all the time. But there’s a problem. While Indie does not put dates on that estimate, we have seen this slide before. We have created this slide before. And the numbers were roughly the same – a few hundred today growing to a few thousand in just a few years. But we made those slides almost ten years ago, and semis dollar content hasn’t really grown nearly as much as everyone back then expected. In part this is due to price considerations – car prices do not move that much (in normal times) which means the auto companies are loathe to give too much value to chip companies absent their ability to recapture that from consumers/ This is a process that will take a change in consumer behavior, which means it will take a long time.
Which brings us to the third factor, Automotive semis move really slowly when compared to all other semis. Most semis people can tell some version of the same story about this. They win an auto design for their latest chip, but by the time that design actually goes into production, and the chip is put in the first car, the chip itself has reached its end of life, and the semis companies really wants to stop supporting it. The average semis for a consumer device goes from design to production to end-of-life in four or five years. By contrast, the average car takes five years just to go from design to production (very roughly speaking). To some degree, this has improved, the auto companies have gotten better at product lifecycle timelines, and the semis companies have gotten better at designing longer-lifed product lines, but there is still a big disconnect between the two industries in timing. (To say nothing of inventory management, which is a whole other problem.)
Finally, there is the fact that auto semis need to go through a rigorous certification process which can take considerable time and expense. Cars are built to last, and that means all theor components have to be built to last. Generally speaking, most semis can withstand the rigors of life in a car, but chip makers need to prove this first, and certification process around that proof creates a fair-sized barrier to entry. Chip companies will not get all their parts auto-grade certified, which further constrains the market.
As we will discuss in our next piece, there are still good reasons to believe that auto semis content will continue to grow. It will happen. And we want to be clear, we think very highly of the Indie team, they have been grinding away at this market for years, we only singled them out because they have the best slides. We believe their forecasts, but we think the whole industry needs to be seen in context.