Between highlights and revitalization

As with many other industries in these dire days, this commerce is also in a deep slump. Practically all large semiconductor manufacturers worldwide have reduced their capital spending in favor of new wafer fabs, and in turn for new equipment. At the moment, today ´s available production resources are satisfactory to fulfill the market´s demands. During the last quarter of the year 2002, about ten fabs were closed worldwide, most of them running 100, 125 or 150mm lines, and built in the 70s/80s of the last century. Obviously, there is no 300mm hype to see this year (and there was none in the past), but a thoughtful improvement of production capacity by about 20 more wafer fabs exists, whereas 200mm technology still counts for the lion’s share. However, for the year 2004 there is a prognosis that 300mm lines will then make for about 60% (12 installations) of the overall newly erected fabs (150mm 3 fabs; 200mm 6 fabs).

According to latest information, the actual sales of semiconductors in the year 2002 were $140.7bn, a slight growth of 1.26% from 138.9bn in 2001. For this year, according to the entire group of market research firms, the prognosis lies between a plus of 12 to 26%, and that’s a lot of variance. Too huge a gap to bridge it easily on all sides involved, be they material, equipment and semiconductor manufacturers. Therefore, all business partners have to be prepared for the unexpected, unfortunately in all possible directions.

In the year 2002, as Semi’s European director Walter Roessger pointed out, the worldwide material consumption has, with $23bn, surpassed the equipment sales with its 22.7bn. We can expect that equipment sales will be generally flat this year, except in some areas for machines enabling advanced-packaging technologies and some other facilitating methods. As long as there is overcapacity and strong price pressure (a situation where semiconductor manufacturers will barely make profits), there is no hot machine-ordering activity to notice.

Otto Kosgalwies, European VP sales and marketing of STMicroelectronics, conservatively expects only a plus of 11% for this year’s semiconductor market, which is a year of transition for him: steepest recovery from the deepest slump ever. The market pie by region is divided into four segments: Europe 19%, Japan 21%, North America 23%, Asia/Pacific with China 37%. By product families, the market volume is splitted into these main solid-state supplies: discretes, standard and opto-electronics with 22%, memories 18%, processors/DSPs 17%, and differentiated ICs with 43%. According to him, flash memories are increasing, differentiated ICs and power management; decreasing are DRAMs and microprocessors.

For larger capital expenditure, he sees a fresh beginning again in the year 2004 in the developed countries, but the Asian market expansion has continued in the past two years and will proceed with a higher rate than in the nearly saturated markets. However, the costs of manufacturing have increased very much elsewhere in the world. Typical labor cost is now only a small fraction of the entire cost complex. For example, at the beginning of the 90s, a fab could be erected with an investment of about $500m; today $3bn is the coinage for this game. Obviously, one erroneous decision and a company could be almost dead (chapter 11 in the U.S.) or in the red for a very long time. Industry in developed countries can expect an encounter with some new competitors from the Asian region, acting unconventionally (and sometimes even improperly?). But their advantage is a “hungry” business mood and the time of new departures, not the cheaper labor. This isn’t an issue in a highly automated wafer fab. So what can we perhaps expect in semiconductor manufacturing? Besides of the 248 deals in acquisitions and mergers in the last seven years, much more of these will probably have to come.

www.semi.org

Gerhard B. Wolski

Indium Corporation is fielding five specialists at SMTA International 2026 to present new findings on solder paste rheology, BGA ball alloy selection for lower-temperature reflow, and the effect of print gap on paste…