For those of you who may be feeling pessimistic regarding semiconductors in the next year, you must not have teenagers. They are absolutely addicted to e-mail, instant messaging, cell phones, MP3 downloads, digital photography, Web cams, DVDs and anything else that can be converted into digital bits. These demographics do not bode well for TV networks and books, but they sure do give me optimism for a future of relentless consumption of bandwidth, electronic consumer gadgets and, of course, systems-on-chip.

Back in 1990, the ASIC industry was abuzz about SoCs. In those days, when I was managing the ASIC division of another chip company, cutting-edge process technology was at 0.8 micron, and SoCs were really just a technologist's dream. Certainly there were many white papers around this topic, but very few systems-on-a-chip were ever produced.

Today many devices, from iPods to Wi-Fi-based routers, contain SoCs. What allowed this dream to finally come to fruition? Basically, it was Moore's Law and what I refer to as the 90-nanometer wall. Moore's Law enabled SoCs to hit the sweet spot of performance, low cost and miniaturization. Meanwhile, the 90-nm wall has caused a sharp increase in tooling and engineering costs that make all-layer, from-the-ground-up customized chips economically prohibitive.

No other market is driving this trend harder than the consumer market. In Asia, the combination of low cost, speed and integration will make or break a consumer project. With the advent of intellectual-property vendors coupled with SoC platforms, new integrated solutions are being developed at an unprecedented rate. It is not uncommon for highly integrated SoCs with two or three bus interfaces, a wireless function, security, an internal switch or high-speed bus, a multiport memory and a microprocessor with cache to be developed in less than six months. As a result, the SoC is now eclipsing the microprocessor as the center of gravity for most electronic systems from 130 nm and beyond.

It feels as if supply and demand for wafers are pretty much in balance, which is as common an occurrence as a lunar eclipse. Regardless, the majority of wafer suppliers appear to be visibly worried about a "glut" in 2005 because of all the 300-mm wafer fabs that are scheduled to come online.

I don't think an oversupply will happen, because next year is likely going to be an even stronger one for semiconductor consumption than 2004. There are many indicators currently moving in a positive direction, specifically: low inventories, increased employment, accelerating corporate and service provider earnings, a generally healthier economy and a confluence of enabling technologies that will inspire more consumption and increased capital investment.

Bob Bailey, President and Chief Executive Officer, PMC-Sierra Inc., Santa Clara, Calif.