It was pick-your-poison time at SEMI's annual January forecasting symposium in Pebble Beach, CA. With the world-famous golf courses of the Monterey Peninsula resort as a backdrop, equipment and materials company executives learned that the chances of selling lots of products this year are as likely as holing out a 90-ft chip shot from the rough—merely to remain 10 over par.

SOURCE: VLSI

VLSI Research forecast a 5% decline in the process equipment market in 2002, down from sales in 2001 that were already close to 38% lower than their record mark in 2000. Worse, Dataquest predicted equipment revenues most likely would decline 19% in 2002 from last year, when revenues reached $25 billion. Even worse, the "optimistic" alternative scenario from Dataquest saw a decline of 12%, while its pessimistic view predicted a 26% drop in 2002.

SOURCE: DATAQUEST/GARTNER GROUP

Sluggish sales of PCs, equipment order delays, a capacity glut, and a precipitous drop in capital spending worldwide all contributed to the gloomy forecasts. China may be the only bright spot. Some good news came in the form of forecasts predicting 20% growth in the global IC market in 2002 over 2001, when sales plummeted 31% from record-setting 2000. The industry analysts disagreed over whether fab utilization rates would improve, however.

Materials suppliers fare better than their equipment brethren. Dan Tracy, a SEMI senior analyst, forecast overall growth of 11.3% in 2002, 17.2% in 2003, and 11% in 2004. Displaying a SEMI slide from 1985, however, Tracy told the assembled executives "we've been here before" as he pointed to similarities in the figures on semiconductors shipped, silicon shipments, and photoresist revenues.

SOURCE: SEMI

As in 1985, materials vendors still carry too much of the financial burden, Tracy pointed out. "There's a cost burden for some of these materials suppliers to develop advanced technologies and integrate them in the fab. Materials suppliers tend to be a lot smaller than their customers."

The burden could cause a shakeout in that market segment, Tracy insists. Photoresist suppliers, in particular, are in a shaky position, "asked to make resists" and meet demands for 248-, 193-, and 157-nm applications.

As chipmakers move on to the next technology node, the same scenario threatens to play out for providers of both low-k dielectric materials and photoresists, Tracy believes. Low-k dielectric precursor revenues will reach somewhere in the range of $80 million in 2002 before hitting close to $200 million in 2003 and close to $300 million in 2004. A slight dip is foreseen for 2005, to approximately $275 million. The bulk of those sales will come from interconnect low-k applications at the 130-nm node, according to figures from Kline and Company cited by Tracy.

"Looking ahead, when the next-generation devices ramp up with low-k technologies," Tracy says, "I think there are only going to be a couple of winners at the next-generation node. There may be two or three players left at that point. It's consolidation. There are just too many suppliers."

The photoresist market may see a similar fate. "Potentially, you can argue that each resist supplier needs to spend $32 million in equipment alone. If you're a resist supplier with a 10% market share and a $6 million market, will you see a return on your investment? I see that same scenario playing out in the slurries markets, too."