Now Is The Time For Chipmakers To Lock In Their Gains

Why, exactly, Intel selected New Albany, Ohio, for the location of a new $20 billion semiconductor plant remains unknown. But there’s no doubt it was at least partly connected to tax and other financial incentives, showing once again the growing clout of the industry as the proliferation of electronic devices requires a huge supply of semiconductors.

Additionally, U.S. Commerce Secretary Gina Raimondo has renewed calls this week for Congress to approve a $52 billion plan to boost American chip production, now deemed essential for economic growth.

While many sectors rely on semiconductors, the newly acquired power of the industry is especially significant when it comes to cars. As severe shortages continue, chip companies will increasingly have the upper hand over car-makers, which, along with competition from newcomers like Tesla, may ultimately force auto manufacturers to adopt a more innovative mindset about onboard computing.

As cars become more connected and software based, chipmakers are poised to play the role of those who sold shovels and other gear during the Gold Rush: whoever is making chips will profit no matter which kinds of cars–those of traditional manufacturers, newcomers like Tesla, or future players like Apple– will dominate the roads of the future. Semiconductor companies can - and should - take advantage of this moment not only for growth, but for innovation; as can traditional or future carmakers, who must focus more on the role of onboard software, and could, ultimately, be tempted to enter the chip space themselves, adding even more momentum to the industry.

For a long time, carmakers have failed to take an innovative enough approach to onboard computing. In fact, most carmakers today still rely on old-fashioned chips, which are among the types experiencing the sharpest shortages. These chips lack the flexibility of multitasking; with each component of a car, from the brakes to power steering, powered by its own chip processor that then communicates with all the other processors onboard, which means that hardware is not upgradable via software.

The reality is that as cars become more advanced, connected, and innovative, they should, ideally, use smaller and more modern chips, capable of multitasking. The electronic guts of cars should be marketed the way airbags, parking assistance or all-leather interiors are, as defining features of quality. This, in fact, is what Tesla, which is the most visible example of a car company that embraces the viewpoint that all hardware should be upgradable at a software level, is doing. The insides of a Tesla resemble those of a laptop – easily upgraded through software updates, capable of evolving and multitasking – while the inside of most other cars can be compared to “dumb” thermostats, which have one job to do and cannot be easily upgraded and used for multiple tasks. This is mainly because Tesla uses more advanced chips, some of which they make themselves.

While chip companies have been wishing for years that more car makers would take a Tesla-like attitude to chips, most automotive companies have resisted, not eager to have to prove and test new chips for safety when the old ones still work. The chipmakers have capitulated to these demands, continuing to produce outdated chips. But, now, chipmakers’ approaches are changing, and fewer chip makers are likely to want to continue making these outdated chips, which have no use outside cars. This is in contrast to other more modern chips, which can be used interchangeably in cars, video game consoles, and mobile phones.

If semiconductor companies can free themselves from having to devote resources to old-fashioned car chips, they, too, will be in a position to be more innovative. In fact, chipmakers are already showing signs of this as they see demand growing for increasingly connected and innovative vehicles, especially electric cars. For example, some have added carbide to silicon-based chips, creating chips that can operate more efficiently at higher voltages and temperatures, and result in less power lost when battery current is converted to energy that runs the motor and turns the wheels.

But there is also a risk for the chipmakers – the car manufacturers themselves. While the car industry has been one of the leaders of subcontracting parts manufacturing to third parties, could they wake up one day and decide to pull more ownership inside their four walls? Twenty years ago, Amazon was a bookseller in a garage. Now, it builds its own chips for machine learning. The trust gap that emerged during the pandemic between the car manufacturers and their chip suppliers due to canceled and late orders could widen even further, giving manufacturers even more reason to pull production in-house. This is starting to happen, albeit slowly, with Hyundai Motors recently saying it would start making chips.

To be clear though, we don’t see a future where someday a Sony or Amazon or Facebook or Apple car uses nothing but their own chips. If Apple could build an iPhone today using only Apple components, it would, but TI, Samsung, USI, and Broadcom are still better positioned to crank out the commodity components that aren’t Apple’s marquee A15-series processor. Cars in the future are most likely to have all the above: some Qualcomm, some Nvidia, some ARM, some TSMC, some designer processors and some (relatively) primitive, commodity chips.

This is an industry with a tide set to raise all boats, and the winners will be chipmakers of all stripes. The news of Nvidia dropping plans to acquire ARM, potentially forming a giant company likely to be cut down by regulators, is another hint that the future of semiconductors will not belong to a few giants but will instead be a diverse bunch of long-time manufacturers plus some possible newcomers.

(Raj Shah is the North America Lead for Technology, Media & Entertainment, and Telecom at digital consultancy Publicis Sapient)