Harder Than It Looks

• First Apple scales back plans to develop its own vehicle. Then Intel creates its own automotive chip unit.

This kind of two-step movement in the automotive electronics industry is becoming more common. NXP buys Freescale, then Qualcomm buys NXP. Harman buys Symphony Teleca and Red Bend Software, then Samsung buys Harman.

All of these moves are proof points that innovation and fleet-footed development are not enough in the automotive market. Those are just prerequisites. Also required are scale and staying power. That puts smaller companies at an extreme disadvantage, and it makes even big companies like Apple rethink their plans. According to multiple reports, Apple originally was looking to bring out its own branded vehicle. It has refocused its ambitions to focus on what goes inside an autonomous vehicle.

The automotive sector is not just another opportunity to sell electronics. Despite the race to put autonomous vehicles on the road, this is still a complex and cautious industry. Carmakers have to comply with a long list of regulations for safety reasons. A faulty car, or even a faulty part in a car, can result in massive liability problems for the carmaker. Recalls are expensive, and the tinge of a faulty design tends to stick around for a very long time.

Second, these are complex systems that are constantly being improved, and in some cases, invented. That creates design incompatibilities. Even the latest and most advanced SoC design today will be two or three nodes behind the market by the time it gets used in cars. This is like comparing a 40nm chip today with a 14nm finFET design. The communications network tied into those chips may be a different generation from one car to the next, or it may be a different generation than one a chip was designed to work with. Or it may be different in one vendor’s vehicle versus another, use a different voltage or wire harness scheme, or have to be redesigned in ways that make ECOs in the mobile market look simplistic.

The consolidation that is underway in the automotive electronics industry is an attempt to manage that kind of disparity and technology churn. Most executives involved in this market sector expect M&A activity to continue for some time. And the move by some of the biggest chipmakers into this space is another facet of the same trend. By controlling more of the pieces, and building architectures that can be extended with backward-compatible, more standardized chips, which can be updated with new software, the downside of these investments can be reduced.

But those investments are still significant, and ROI will take time. It also carries some risk. No one is quite sure how liability will shape up in the automotive sector when something goes wrong, and how far down the supply chain that liability will spread. Proving what went wrong and who caused it could require a massive investigation and years of litigation, which pushes the price of being first to market well beyond the economic limits of many companies—or at least their investors

This is one of the reasons the medical industry has been so slow to adopt technology, even though that technology presumably could have a huge impact on human health. Automotive is likewise a risky business with a long lead time for payback.

So why are big companies still involved? Because there is still the promise that with persistence and the proper approach, this market could yield dividends far beyond anything the semiconductor industry will see over the next decade. Automotive has emerged as the next big thing for the biggest players, as well as their partners. But it will be years before we know who got this right.

Related Stories
New Wave Of Consolidation
What’s driving it, what’s next, and who should be concerned?
Samsung To Buy Harman For $8B
Tier 1 acquisition boosts automotive portfolio.
What Can Go Wrong In Automotive
Experts at the Table, part 1: Security, diagnostics, standards and the future of autonomous vehicles.