Lane Departure Warnings For The Auto Industry

The automotive chip market is undergoing a series of subtle but significant shifts behind the scenes that could have major implications for the global automotive supply chain.

After a few years of racing toward autonomous vehicles and setting in motion a frenzy of activity, some of the big auto makers have begun taking the design of key functions such as centralized logic in-house.

There are a number of reasons for this move, and there are a variety of possible outcomes.

The starting point for all of this electrification of vehicles and a push toward increasing levels of autonomy, which have completely up-ended the automotive market. What differentiates one vehicle from another is no longer what set it apart in the past. With a couple hundred parts in an electric motor, compared with a couple thousand in an internal combustion engine, the differentiator between one vehicle and another is no longer how fast it accelerates or how fast it can go around curves. In an autonomous world, all vehicles will move in sync. They will accelerate at the same speed as other cars, and they will drive at the same speed – probably down to the millisecond or less.

This has left many carmakers wondering whether the real added value is in the chassis and drive train, which is where it has been for the past 140 years, or whether it has shifted to the electronics that will control those vehicles. And this is what really has carmakers worried, because it opens them up to competition from companies such as Google and Apple, which are very good at developing complex electronics, as well as the Tier 1 suppliers, which are designing and in some cases building their own chips.

That has a direct bearing on cost, which is a key piece of the equation. Economies of scale require the use of a platform approach, particularly in automotive where a single car model will not generate enough volume to achieve that cost benefit. This is particularly important for the AI chips, which will be developed at the latest process nodes, and if carmakers allow their suppliers to develop those chips, then they will be required to use a platform over which they have little control. This is why automakers have been hiring chip engineers, and putting a heavy focus on developing software that will work with those chips.

Internal development has other benefits, as well. If something goes wrong, the time to fix a problem is potentially shorter because there is no finger pointing. The classic vertically integrated model has great benefits if cost isn’t an issue. For instance, engineers who developed an electronic system or the IP blocks within those systems will have a much better understanding of what to fix if something goes wrong, and they can patch it or replace it without any fingerpointing involving one or more third parties.

The big questions are how quickly these companies can come up to speed on 7/5/3nm logic design, and whether they can lure away or train enough hardware, firmware and AI engineers to make this work as well as companies such as Tesla. Tesla started out with an expertise in battery technology and electronics, and its initial vehicle used a chassis developed by Lotus.

For the Tier 1, 2 and 3 suppliers, this raises some serious questions about their role in the development of these electronic systems, as well as their financial model. If carmakers develop these electronic architectures, then it’s likely that none of them will be standard outside of their own subsidiaries. So General Motors, Volkswagen and Toyota, for example, may develop platforms that span all of their brands, but they are likely to be different than those developed by Hyundai, BMW and Honda. That means the carmakers will reap the benefits of scale, while everyone else will be left scrambling.

There are a lot of big “ifs” in this equation. But the main one is if the external suppliers are limited to developing components for platforms developed by the automakers, then they will have much more work and much lower profits than if they could develop a consistent platform themselves. It will be more like developing applications for Linux, MacOS and Windows than a single API, and with hardware that will be even tougher. And if they sink development dollars into developing their own platforms, will there be enough return on investment to make it worthwhile.

It will take most of the next decade to see how this ultimately unfolds. But at least for the time being, this industry segment has hit the reset button and it’s not clear who the winners and losers will be.