The field-programmable gate-array (FPGA) market currently finds itself in recovery mode from the dramatic decline it suffered in 2001. The recovery, while still small at the end of 2002, will start in earnest toward mid-2003 and last through 2004. However, a return to increasing revenues does not reveal the whole story for this market segment. To gain a better perspective, one needs to look at the larger picture that's driven by the underlying technology and the applications it empowers.
On the technology side of the equation, the addition of high-speed serial channels implies a certain level of sophistication in designs that use mixed-signal expertise. The advent of this mixed-signal capability is not new, as high-speed serial channels have been available for some time. But its increasing use underscores how the application landscape has shifted. No longer are pure digital designs and computing applications driving the industry. Now, mobile applications like wireless communications and portable computing that require increased mixed-signal functionality are becoming the market drivers. The product families themselves have evolved to encompass the capabilities needed to accomplish the new tasks dictated by the applications shift. This trend will continue for the foreseeable future.
Another evolution of the FPGA market is the potential inclusion of mid-level analog functionality into product families. This migration to the analog arena is similar to the recent move by FPGA vendors to offer DSP and CPU cores as different options to designers. It also is a reflex to the overall trend of silicon solutions moving toward higher levels of integration to provide more functionality and capability at lower prices.
A third trend could involve FPGAs using ferroelectric technology as their memory solution instead of flash or EEPROM as the preferred method of providing nonvolatile storage of chip configuration data. Recent advances in manufacturing processes for ferroelectric memory look to solve many of the problems that have prevented this approach from being used more widely in the industry. Only time and large production volumes of parts using ferroelectric memory will tell if these manufacturability issues have truly been solved. Yet it's safe to say that the performance of FPGAs will be greatly enhanced.
Finally, we believe the advent of product families with high performance and low price points is the trend that will most directly impact the FPGA market in the near term. This is simply because the lower the price point, the wider the market addressed by an FPGA-type of solution can become. FPGA manufacturers stand poised to capitalize by the extension of their products into markets never before considered as candidates for FPGA-type solutions. Furthermore, customers willing to use these new types of low-cost and highly-featured FPGAs as their silicon solution can finish their system designs in a fraction of the time needed to design and finish an ASIC.