Microprocessor vendors have convinced automakers to cram dozens of 8-bit controllers into every modern car. Now these chip vendors are seeking the mother lode: an in-dash computer with a high-performance 32-bit CPU. Jokes Mike Iannitti, director of Intel's In-Car Computing Division, the car will soon be just a "$30,000 wireless Internet device."
The in-dash computer will offer a variety of functions that are compelling to both car buyers and car makers. Drivers will appreciate downloadable music and digital radio, navigation assistance, hands-free cellular calls, and Internet access, all controlled through a voice interface. Car makers seek monthly service fees and continued contact with their customers.
This confluence of desires will accelerate the adoption of in-dash computers. A few add-in systems, such as Clarion's AutoPC, are available today, and General Motors expects to have 1 million users of its OnStar system by the end of this year. Widespread deployment of in-dash computers will begin in model-year 2002 (just over a year from now). I expect in-dash computers will be standard or optional equipment in all but the cheapest cars by 2005.
These functions require a fairly powerful processor. A good voice interface consumes 100 to 200 Mips. MP3 decoding takes another 40 Mips, and emerging audio formats may need more. Basic GPS functions are not too demanding, but calculating and analyzing multiple routes can consume arbitrary processing power, as the driver often needs new directions as soon as possible. Adding DSP functions to the processor will be helpful unless the system includes a separate DSP chip. Thus, the entry level for this market is likely to be a 200-MHz CPU, with faster processors appearing as they become affordable.
About 54 million cars were sold last year, making the potential market for in-dash computers about half the size of the desktop PC market. At this early point, the market is open to any vendor that can deliver the combination of performance, integration and price desired by the automakers. But as the market grows, standards will emerge to ease the burdens of software development and portability. Ultimately, one or two processors will dominate. With tens of millions of high-end units at stake, now is the time for CPU makers to chase this opportunity.
Linley Gwennap is the founder and principal analyst of the Linley Group (www.linleygroup.com), a technology analysis firm in Mountain View, Calif.