To chart a practical course for navigating uncertain waters, electronics companies need to set in motion strategic plans to develop personalized products specific to niche markets.

The technology industry needs consumers to start spending money again and stimulate the economy — and customized consumer products just may be the way to do it. After all, consumers are accustomed to having it their way, to paraphrase the popular slogan of a fast-food empire.

All the same, custom-made consumer products won't come out of the traditional ASIC-based development cycle, currently in practice in most electronics companies. Project teams have found that ASIC design demands high-volume products in order to cover a company's' costs. In 1997, for example, mask costs were at $100,000 for 0.5-micron fabs. Mask costs are now closer to a staggering $1 million for 0.13-micron fabs, and the number keeps growing.

Next-generation digital televisions, residential gateways, small-office/home-office routers and automotive entertainment systems will be powered not by ASICs but by FPGAs, since project teams favor them over ASICs. Complex, affordable and even high-speed, FPGAs are quickly closing the performance and gate-count gap. Technology breakthroughs, from 10 million-gate densities to 400-MHz clock speeds, enable FPGAs to handle many applications currently served by ASICs, while 300-mm wafer sizes have made them readily affordable for volume production.

Most important, FPGAs take far less time to design and implement, while providing the means to customize, update and upgrade. That makes them preferable to ASICs for an increasing number of commercial applications.

Dataquest Gartner research confirms the trend. The number of ASIC design starts went from 5,000 in 2001 to 4,300 in 2002 and is projected to drop to less than 2,000 in 2004, according to Gartner Dataquest. Comparatively, FPGA design starts went from 300,000 in 2001 to 400,000 in 2002 and are expected to reach 500,000 in 2004.

FPGAs have long had an advantage over ASICs for both turnaround time and implementation. Add nonrecurring expenses, prolonged time-to-manufacturing, the risk of respins and high inventory costs to the budget, and it's easy to see why project teams are reassessing their approaches in favor of FPGAs. Using an FPGA implementation, they can quickly create products customized for a niche market.