For the last five years we’ve become accustomed to seeing the hottest tech hit our mobile phones, but that may be about to change. Sure, phones will continue to get smarter, and perhaps innovations rivaling capacitive touch or natural language personal assistants will still hit the mainstream on handsets, but our vehicles have a brighter future. The chip industry is betting on automotive in a big way: Firms are developing more and more speciality silicon for cars.
Nvidia, Texas Instruments and others are building special applications processors to run the consoles and dashboards of today’s in-vehicle entertainment and navigation systems, while smaller firms such as Freescale and Spansion are building specialty chips for handling co-processing or communications. Overall in the last three years the cost of chips inside a new car has increased from $299 to $355 per vehicle, according to IHS iSuppli. However, those numbers are skewed by the huge numbers of new cars with less silicon inside hitting highways in highly populated countries such as China and India.
When it comes to Western Europe and the U.S., the value of silicon and the new stuff coming online is far more expensive — and advanced says Egil Juliussen, principal analyst infotainment and ADAS at IHS iSuppli. Juliussen estimates that in the next decade we’ll see driverless cars, so to get there a lot will have to happen for the on board silicon and sensor networks.
There are several reasons that the car may be the new font of innovation for mobile applications. Automobiles are high-priced goods so they can absorb a few high-priced chips, and they have batteries that can power more silicon without being forced to shut down after lunch. This is a trend that has been building up for a while, but I think in the next year or two we’re going to see cars with services that redefine technology, much like the iPhone redefined touch screens or the Xbox 360 redefined video game consoles that are now full-featured entertainment machines.
A battle for the brains inside the car.
The automotive console on my 2006 Acura TSX is like a dumb toaster compared to the consoles of today’s new vehicles. But with all of this built in connectivity and navigation systems, cars need a smarter application processor to run everything and Texas Instruments and Nvidia both hope to provide it.
Nvidia scored a spot running the controls of the latest Tesla, while Texas Instruments is working with in-vehicle infotainment providers like Harmon to place its OMAP processors inside cars. Both firms have of applications processors built for automotive use that are related to their smartphone application processors, but which offer more performance. For example, a quad-core chip is better suited for a car than a phone where all the cores can process and consume power thanks to a huge battery.
Plus, as cars go driverless we’ll see a need for greater processing capabilities because they will have to track and model various possible collision scenarios based on the information sent from vehicles around them. Those cars will also have to communicate with the automotive components like brakes and tires as well as require more advanced communications chips for sharing info with other cars. They’ll also have more sensors on the vehicle.
Infotainment and connectivity drive today’s chips.
At the high end chipmakers are focused on entertainment and connectivity. As this chart below from IHS iSuppli shows, the amount of revenue derived from putting better navigation and Bluetooth into cars is steadily rising. And while the jury is still out on how cars will connect without incurring huge data bills and straining network capacity, it’s clearly a platform of interest to carriers.
What’s more interesting than the run of the mill infotainment and connectivity chips are the emergence of specialty processors such as a speech recognition chip from Spansion slated to come out in 2013. This co-processor will work with Nuance’s speech recognition libraries and contain databases of sounds on the chip so more of the processing occurs on the device as opposed to getting sent to a server somewhere.
The result is a faster response and more offline functionality. Sending less information to the cloud benefits the owner in the form of lowered data costs, but it also makes sense for someone who might be verbally requesting directions while driving at 60 miles per hour. Waiting for your spoken command to take a round trip on a 4G or 3G network might leave you a few hundred yards past your destination.
Another element of connected cars worth watching is how electric cars will connect with the smart grid and adapt to the changing electricity demand or how they interact with other cars to relay traffic information back to drivers and public safety officials.
So as Apple and others attempt a coup on the entertainment side, building a platform and walled garden for the apps, navigation and driver interface, under the surface are a host of other opportunities for innovation and chip vendors that want to make cars smarter and more responsive.
What will that mean for users? The average American holds onto a car for more than 10 years, according to Juliussen. So will new cars soon have a shorter useful life before their hardware malfunctions or just can’t support the latest software? I’m not sure I’m eager to see the future of mobility if it means I have to buy a new car every five years or less.