If the past were any guide, semiconductor makers ought to be feeling chipper about 1991. Their industry's latest feast-to-famine cycle bottomed out last year, and chip sales have regularly defied the overall economic climate. During the recession of the early 1980s, for instance, the semiconductor business grew 20%. Then it turned sour in 1985, when the economy was strong.
This year was supposed to be countercyclical, too. In September, the Semiconductor Industry Assn. (SIA) predicted that the world would gobble up 12.5% more chips than in 1990. At yearend, however, expectations took a nosedive, although America's market leaders may still score double-digit growth. Silicon Valley, it seems, is falling victim to competition from Japan and to the past successes that have made chipmakers too big to escape general economic trends.
NO GROWTH? "Everyone is extremely cautious" about 1991, says Alfred J. Stein, chairman of VLSI Technology Inc. in San Jose, Calif. The prevailing wisdom now calls for overall growth in U. S. chip sales to be in single digits for the third year in a row. Even highly profitable Intel Corp., which is still struggling to meet heavy demand for its 386 microprocessor, "is as worried as everyone else," says its chief financial officer, Robert W. Reed.
In September, to cite one sign, the chip industry's closely watched book-to-bill ratio slid below 1.0, to 0.96. This signaled that new orders, or bookings, were lagging behind factory shipments, or billings. And sales didn't pick up in the fall as they normally do before a good year. Instead, weak orders hammered the November ratio to 0.90, its lowest point in 14 months.
That set market researchers to trimming their 1991 estimates. Dataquest Inc., one of the more optimistic forecasters, cut its worldwide growth estimate for 1991 chip revenues to 9%, six points below its October guess. For the U. S., a bearish In-Stat Inc. even predicts a second straight no-growth year.
A couple of trends work against U. S. chipmakers. In past downturns, sales of personal computers stayed relatively strong. Today, PCs are sitting on so many desks that increased sales are tightly coupled to the strength of the economy. Moreover, most domestic chipmakers have sharply curtailed production of commodity-type chips, where price pressure from offshore rivals is keenest. This helps protect their margins, but it also eliminates opportunities to participate in a wide variety of end uses, some of which may buck economic trends. For instance, laptop computers are still hot sellers. But most laptops are made in Japan--and contain few U. S. chips.
Among the markets that Silicon Valley now counts on to fuel demand for chips are engineering workstations and top-end PCs. Both could get pounded if the recession extends past spring. And consumer-electronics products, another leading market for chips, could end up in the same boat.
The telecommunications market looks like a safer bet. Such specialty items as pocket cellular phones and big electronic switching systems for phone companies should weather the slump fairly well. And auto makers, despite sagging sales, may order more semiconductors, as they build more electronics into cars.
But sharing in these growth markets will generally require a supplier to make CMOS chips, short for complementary metal-oxide semiconductor. Such chips consume less power and are thus ideal for portable computers and other miniature products. Market researcher Integrated Circuit Engineering Corp. figures that CMOS sales will grow about 20% this year, while some non-CMOS markets will plummet up to 48%.
Keeping up with CMOS technology takes a lot of capital, however, which many U. S. chipmakers won't generate as they retreat to ever smaller niche markets. The value of deep pockets shows in the growth prospects of the top 15 of America's 100-plus chipmakers, which will outgrow the overall industry by about 50% this year, says Daniel L. Klesken, an analyst at Prudential-Bache Securities Inc. These big producers, which bring in 80% of industry revenues, can afford to invest in new technology.
FLIP-FLOP. Even they may face tougher times, however, as the U. S. continues to lose market share: It now has 40.5% of worldwide chip sales, down from 56.7% in 1982, the SIA says. In CMOS chips, the situation is even worse: Dataquest pegs the U. S. share at only a third. One reason: The Japanese now buy 40% of the world's chips, compared with only 29% for U. S. electronics companies--a flip-flop from just five years ago.
U. S. chipmakers are scurrying to boost sales in Japan. Last October, Intel Chief Executive Andrew S. Grove unveiled his new low-power microprocessors for laptops in Tokyo. While the U. S. share of Japan's chip market has climbed to about 12%, from 8.6% four years ago, it's well short of the 20% Japan promised when the Semiconductor Trade Agreement was negotiated in 1986. That deal expires this summer.
The SIA is pressing Washington to negotiate a new chip pact. But skeptics doubt that it will be any more effective than the current one because of Japan's ingrained protectionism. "When a country says you can't sell them baseball bats because you have no experience with their baseball diamonds, it should tell you something," declares In-Stat President John H. Beedle. As a result, regardless of what happens in 1991, analyst Klesken foresees a bleak future for Silicon Valley's second-string players. The 1990s, he predicts, "will be the decade of consolidation."