Computers and semiconductors continue to be growth engines of the global economy, but 1997 is not shaping up as a go-go year. After a 10.5% revenue decline in 1996--the first since 1985--the chip industry should grow a modest 7.4% in 1997, to about $139 billion, according to the Semiconductor Industry Assn. Meanwhile, personal-computer sales will grow about 18% in 1997, to 85 million units worldwide, on track with last year's estimated 19% growth, says market researcher Dataquest.
In many industries, numbers like these would be cause for celebration. But for the personal-computer and chip sectors, which two years ago enjoyed growth of 26% and 42%, respectively, moderate expansion is frustrating. For now, the industries seem to have picked all of the low-hanging fruit. International Data Corp. pegs PC growth in the U.S. at only 13% in 1997. So PC vendors are targeting the rest of the world. Sales should rise 22% in Latin America and 25% in Japan and other parts of Asia, according to IDC. "Asia is growing like wildfire, and it'll continue to," says Ted Waitt, CEO of mail-order PC vendor Gateway 2000 Inc.
Finding PC growth markets is paramount because the fortunes of the chip business are closely tied to personal computers: PCs consume 40% of total semiconductor output and two-thirds of all memory chips. To lessen their vulnerability to PC market swings, chipmakers are pinning hopes on emerging digital consumer-electronics products. Trouble is, these devices--digital video-disk players, handheld computers, personal communicators, digital cameras--won't hit the big time in 1997.
That's why the PC industry is revving up to convince customers that they need the latest wave of multimedia computers, built on Intel Corp.'s next-generation microprocessors. These chips incorporate Intel's MMX technology, which speeds up such multimedia features as audio, TV-quality video, and 3-D graphics. MMX will appear in 200-megahertz Pentium chips delivered in January and then migrate down to 166-MHz models. By midyear, it will also show up in a slimmed-down, 233-MHz version of Intel's Pentium Pro, code-named Klamath.
"LEAP FORWARD"? But these $2,000-plus machines may not catch on with the 60% of American homes that still lack PCs. "Consumers are looking for some kind of leap forward, but it's not clear it's there," says IDC analyst Richard Zwetch-kenbaum. That's why some analysts see a growing niche for boxes priced at $1,000 or less. Ronald Chwang, president of Acer America Corp., predicts that several major PC vendors, including perhaps his company, will target this segment. That's promising news for chipmakers Advanced Micro Devices Inc. and Cyrix Corp., whose inexpensive Intel-compatible chips could power many bargain-basement PCs.
The business segment is a different story. To spark upgrades in the saturated corporate market, Intel and
Microsoft Corp. are pushing Pentium Pro desktop PCs running the high-end Windows NT operating system. Sales of Pentium Pro systems are expected to climb by a factor of 10 in 1997, to around 25 million computers. But only half of those are expected to run NT. Windows 95 and Windows 3.1 will still command 80% of the 1997 market, estimates IDC.
Intel and Microsoft are also responding to the threat of so-called network computers, whose backers claim they will be cheaper for businesses to operate. Network computers fetch programs and data off central servers instead of storing them on a built-in hard drive, simplifying maintenance and giving computer managers greater control over what programs users run.
In answer, the twin giants of the PC world plan to co-develop "NetPCs" that are like conventional PCs, but with extra hardware and software to streamline their management across corporate networks.
Prospects vary in other parts of the computer business. Notebook computer sales growth should be in the high teens. Growth in servers--the powerful PCs that run networks--could top 30%. Driving up server sales is the rush by businesses of all sizes to put up Web sites on the Internet. Unit sales of mainframes are expected to climb in 1997, but Dataquest projects revenues will fall 9.5% because of declining prices. A hot spot in mainframes is Unix-based machines, with growth rates in the high teens to low 20s, says Salomon Brothers Inc. analyst John B. Jones Jr. But minicomputers and workstations should both post only modest 6.3% gains in 1997, says Dataquest. Apple Computer Inc.'s $400 million purchase of company co-founder Steve Jobs's NeXT Software won't keep Apple's sales from being flat this year, analysts say.
For PC makers, the holy grail is the convergence between computers, consumer electronics, and communications. To some extent, MMX PCs address this vision through their support for audio and video. But to really reach the mass market, says IDC's Zwetchkenbaum, "it's going to take new, less versatile gizmos designed specifically for certain tasks." In other words, low-cost information "appliances" such as Web-surfing TVs, pocket organizers, or smart wireless phones.
In the near term, chipmakers must continue relying on PCs to drive growth--and they have to keep closer tabs on demand. The semiconductor downturn of 1996 occurred because chipmakers overestimated demand and added too much production capacity. Prices of dynamic random-access memory chips (DRAMS) fell more than 75%. The glut won't go away overnight. "The major vulnerability for chipmakers in 1997 is overcapacity in DRAMs," says Jean Philippe Dauvin, chief economist for SGS-Thomson. On the bright side, he adds, non-DRAM business started upward again in October, 1996, after nine months of decline.
DRAM DREAM. The chip recovery will eventually help suppliers of semiconductor manufacturing equipment such as Applied Materials Inc. and Lam Research Corp.--but not this year. Their business lags behind the chip market by at least six months, so the chipmaking-equipment industry forecasts a 5.5% sales dip in 1997, to $25.4 billion.
Makers of DRAM chips are hoping that an increase in the amount of memory shipped in PCs will help their results in 1997. The average DRAM per new PC should climb from 20 megabytes to 30 megabytes or more in 1997, says Vladi Catto, chief economist for Texas Instruments Inc. Still, it'll take the PC industry a few years to chew up the memory surplus--especially since new government-backed ventures in Singapore and elsewhere in Southeast Asia will add to the supply. Because memory capacity per chip is also rising, the total number of "bits" produced in 1997 will exceed demand by 35%, keeping prices depressed.
Other parts of the chip industry will fare better. The microprocessor field, dominated by Intel, will see revenue growth of 32%, to $22.5 billion, says market watcher Micrologic Research. Of the roughly 95 million PC microprocessors sold in 1997, Dataquest predicts 60 million will be Pentiums and compatibles and 30 million will be Pentium Pros and compatibles. One 1997 star could be Advanced Micro Devices, whose Pentium-compatible K6 chip will likely ship in the first quarter. AMD says the K6 could even outperform Intel's Klamath. If AMD can undercut
Intel's price and sell the 5 million to 6 million units it plans to make, "it'll be in fabulous shape," says Dataquest analyst Nathan Brookwood.
Analysts remain bullish on long-term chip prospects, promising a return in 1998 to historic 15%-to-25% growth rates for the next decade. By 2000, then, expect sales of 132 million PCs a year. But in 1997, PC and chipmakers face less heady times: The name of the game this year is getting positioned for what everybody hopes will be the next great wave.