Corrigan still sounds "cautiously optimistic" in this article. I wonder which way reality will break?
May 27, 1998, TechWeb News
------------------------------------------------------------------------
Downbeat On Chip Industry
By
WILFRED J. CORRIGAN
Chairman and
Chief Executive Officer
LSI Logic Corp.
Milpitas, Calif.
Two years ago the global semiconductor industry experienced a modest downturn, primarily as a result of the collapse of DRAM prices coupled with a worldwide inventory correction. Then, last year industry fortunes turned upward by low single digits, but still far from the historical compounded average growth rate of 17 percent.
Many observers were anxiously awaiting 1998, predicting that this year would serve as the transition from the moderate decline of 1996-97 to the resumption of the industry's double-digit growth rate of the past 50 years. But these hopes will obviously not be realized in 1998. Even the normally upbeat Dataquest Inc. is forecasting only 8.2 percent growth, while other market analysts are even more bearish.
The reasons are threefold: DRAM prices are expected to fall another 40 percent this year, adding to 1997's decline of 75 percent; microprocessors and micro-peripheral prices are under severe pressure as a result of the emergence of the sub-$1,000 PC, and the Asian financial crisis has not only resulted in depressed currencies in east Asia, but it has also constrained the expansion plans of Japanese and South Korean semiconductor manufacturers.
Is there any blue sky showing beyond this litany of preannounced quarterly reports, red ink and layoffs? Certainly there are.
Average selling prices are under pressure, but unit volumes have actually increased. If you view the global semiconductor market by means of unit volumes rather than actual sales, a different picture emerges.
In 1996 worldwide semiconductor sales fell 8.6 percent but IC unit sales increased 6.7 percent, from 45 billion to 48 billion units. Last year the industry grew only 4 percent in sales, but unit sales surged 21.8 percent to 58 billion units.
But that comparison tells only half the story because merely charting IC unit growth does not take into account the tremendous complexity and functionality increases for the average unit.
What would happen if you view the world from Tokyo as opposed to Silicon Valley? If we measured the global semiconductor industry in terms of the falling yen as opposed to the rising dollar, a totally different picture emerges.
Denominating the world in yen, the industry rose in 1996 by a modest 4.2 percent instead of falling by 8.6 percent. In 1997 the industry would grow by the historical average of 18.2 percent by using yen, instead of barely increasing by 4 percent in dollars. And if the industry grows only 5 percent in 1998 by employing dollars, it would reach 12.4 percent in yen.
Another measurement that runs counter to the sales figures is the 55-percent-per-year increase in transistors per capita, which is expected to reach 1 billion transistors for every man, woman and child by 2008.
One of the main reasons for this expected surge in the number of transistors used per person is the growing acceptance of the Internet. In fact, the Semiconductor Industry Association reports that Internet usage is doubling every 100 days. The key driver for this marked upswing in Internet usage is e-commerce, which totaled $3 billion in 1996, $11 billion in 1997, and which is expected to jump to $123 billion in 2000.
Another factor is the global movement toward system-on-a-chip integration. There are several reasons for the widespread acceptance of single-chip systems: lower system costs, expanded functionality, improved reliability, reduced power consumption, revolutionary new products and improved access to end products.
Many throughout the high-tech community are talking about single-chip systems. But actually serving major customers with system-on-a-chip products is a much more difficult proposition. A system-on-a-chip company must have intellectual property, engineering expertise, design-automation tools, process technology and methodology, global manufacturing and a proven track record.
This year will be remembered for the widespread public acceptance of single-chip DVD players, digital cameras, GSM and CDMA cell phones, and other breakthrough applications. It will also be the year in which process technology made possible multiple systems on a chip, combining two or more separate products on one highly integrated silicon chip.
Once again, system-level integration-in this case systems-level integration-will open the way for futuristic products. An example of the convergence of multiple systems on a chip could be a universal set-top box that incorporates the technology for a cable receiver, digital terrestrial television, DVD and a satellite receiver all on a single chip. A communication system for the automobile of the future could include any combination of global positioning, digital audio broadcast, cellular telephone, Internet access, speech recognition and even a printer, copier and fax-all enabled by a single chip.
So, there are reasons for optimism. The industry will eventually resume its double-digit growth rate, sooner than many people think.
Copyright r 1998 CMP Media Inc |
