Chip-Makers on Rebound : Focus Is on Stability Rather Than Growth
It was so rough last year for U.S. computer chip-makers that the telling of it begins like a Johnny Carson joke. “1985 was a bad year for the semiconductor industry” is the standard opening line.
A securities analyst’s report offers the straight line, heading a set of figures with: “How bad was it?”
Of the dozens of possible rejoinders, not one is funny.
It was so bad that the five leading U.S. sellers of chips posted losses totaling $343 million. (They had had profits of $1.3 billion in 1984.)
It was so bad that a chip that sold for $25 the year before cost a measly $3 in mid-1985--an 88% plunge.
It was so bad that thousands of workers lost their jobs; thousands more took pay cuts, went on forced vacations while plants shut down, or had raises cancelled.
It was so bad that the median price of housing in some communities of Northern California’s Silicon Valley dropped for the first time in a decade, as well as the number of qualified home-buyers.
It was so bad that for the first time, the United States lost (to Japanese giant NEC) the market share lead in an industry once hailed as a tonic for U.S. economic growth.
It was so bad that 1986 finds the semiconductor industry looking in every conceivable corner for a bright spot.
Most of the industry’s analysts and observers, and the companies themselves, are predicting an upturn in 1986. In part the optimism is spurred by similar predictions for the computer industry, a major buyer of the electronic circuits.
However, those who predict such things have been forced to take a more modest view than in years past. Now they have to factor in ever-stiffer competition from Japanese companies and large factories still echoing with excess capacity.
Conservative estimates place the semiconductor market growth as low as 5%, with others saying it will be 18% to 20%. Last month, a gloomy Jack Beedle, president of In-Stat, a market research firm in Scottsdale, Ariz., told a Newport Beach conference that this year’s sales will fall 4% below the dismal 1984 sales.
Even the optimists admit that it will be a long while yet before the semiconductor business rebounds to the dizzying profit levels of 1984. “The way to think about it,” said Howard Bogert, vice president of Dataquest, a market research firm based in San Jose, “is that all year long in 1985 the industry was sliding down a slippery hill and all year long in 1986, it’s got to climb back up that same hill.
“Our (9.8% growth) forecast doesn’t sound spectacular,” Bogert continued, “but it really is spectacular when you think that we’re saying it can climb back up.”
The industry, just a few years ago prized as America’s new business frontier, is sobering up and settling down. U.S. companies still are the leaders in the semiconductor technology, but must confront the prospect of becoming, as one executive said, “research and development houses” for Asian companies, already masters of consumer electronics.
Increased Pressure
As a greater percentage of the chips go into Japanese- and Korean-made consumer appliances, the U.S. sales portion of the market (it was $8 billion last year) will slip, putting even more pressure on the U.S. companies to develop new markets.
Hard times and the challenges of protecting its markets and technologies have forced many domestic chip makers to trim back and tone up business practices. And that has tested Silicon Valley’s commitment to its new-fashioned, people-centered style of running a business. For the most part, it held up, even though some of the more fanciful aspects were engulfed by the sea of red ink.
For instance, Advanced Micro Devices, though caught in the downward spin, maintained the no-layoffs policy it instituted during 1981’s recession. Instead, it cut some work days, froze salaries and slashed non-hourly salaries by 10%--with the top 100 executives taking 15% pay cuts that were restored only Jan. 20. And, AMD cancelled its ostentatious Christmas parties--on which it had spent about $1 million in 1984.
“It’s certainly a different ballgame,” said Daniel Klesken, semiconductor industry analyst for Montgomery Securities. “The Japanese dominate the memory market. The Koreans are coming . . . American companies are hurting financially; many face another quarter or half-year of losses.” The San Francisco-based investment firm predicts a 7.5% growth in worldwide sales of semiconductors this year, to $22.8 billion from $21.2 billion last year.
George Haloulakos, who tracks semiconductors for the Seattle-based brokerage firm of Cable Howse & Ragen, said, “Certainly expectations are remarkably different that what they were in the halcyon days of 1982-1984.” Haloulakos looks for a 5% to 10% growth this year.
In lieu of obvious hot new products to spark another boom--as the personal computer did in 1983 and 1984--the industry will settle for expanding already existing markets, such as automotive, military and aerospace, telecommunications and computer-aided design and engineering.
Area for Future Growth
Many say the industry’s future growth will be in ASICs--an acronym for application-specific integrated circuits. On the other hand, there’s not much profit in the standard memory chips, now referred to as commodity chips because of their high-volume production and usage.
Semiconductors are small, thin pieces of silicon--a natural element so abundant the word “ubiquitous” could have been invented just for it--onto which patterns are etched to create electronic circuitry. The devices are used to handle information and control things--such as a microwave oven, a calculator or even a personal computer. Customizing a chip, by integrating more and more tasks on each one, reduces the number of chips required for a product or process.
Companies such as Sony have been able to make ever-smaller radios and televisions by using ASICs that combine bunches of electronic components into fewer and smaller parts.
Each customized, or ASIC chip, is more expensive to make than a standard, all-purpose kind of chip, and has less uses. In 1985, ASICs accounted for about 12% of all sales worldwide, according to Dataquest. Bogert, who is director of Dataquest’s semiconductor industry group, said the ASICs segment will grow faster than the rest of the industry, and by 1990 will generate 15% to 20% of all semiconductor sales.
“ASICs are a way of making semiconductors cheaper,” Bogert said. “They are more expensive, on a per-unit basis, but if they replace enough cheaper units, they become cheaper.”
Saves Assembly Costs
For instance, he said, one $30 ASIC might replace several hundred 20-cent chips. “So you might break even on the actual chip cost, but . . . you don’t have to connect those hundreds of pieces, and so you save costs there.
“It’s the difference,” Bogert added, “between buying the lumber for a house and buying the house.”
To many analysts, the bright spots in the U.S. industry are those firms that have been specializing in the customized chips, where there is little competition, as yet, from foreign producers.
For those companies, jokes are not so hard to come by. Michael J. Callahan, chief operating officer of Monolithic Memories, a custom chip house, drew chuckles from his audience at a recent Montgomery Securities-sponsored conference on the industry when he said “We call them C-SICs . . . for customer specific.”
The five large semiconductors makers, Texas Instruments, Motorola, Intel, National Semiconductor and Advanced Micro Devices, say they are either making some ASICs already or are beginning efforts in that segment.
Some of their smaller competitors declare with bravado that “the gorillas,” as the five companies are called, are having difficulty switching gears rapidly enough, and thus are missing the lead in the ASICs market. But Jerry Crowley, president and chief executive of OKI Semiconductor, sounds a more temperate note. “The gorillas,” he said, “are important and they’re going to be successful.”
Wide Swings in Orders
One thing the larger and smaller companies share is a desire to shave back the wide up-and-down swings of the industry. Nearly everyone concerned now admits that much of the U.S. market’s troubles in 1985 can be traced to over-zealous ordering by the computer industry during 1983 and 1984. When growth of computer sales didn’t keep pace with previous years, the companies just slammed the brakes on orders.
Orders have been a nagging problem for the industry since August, 1984, the last time orders and shipments were keeping pace. But the drought may be ending.
In December, the book-to-bill ratio--the measure of the balance between orders and shipments--was almost even, according to the Semiconductor Industry Assn., a trade group based in San Jose. Many believe that January’s measure (scheduled to be reported this week) will be greater than even for the first time in 17 months.
“What were really looking for now is a rise in the order rate,” said Sheila Sandow of the SIA. Sandow said orders were up 9% in December from November, according to a complicated formula that she said tends to understate the actual results during an upturn.
Many analysts say that computer companies, who also suffered losses and implemented cutbacks last year, have learned to better control inventories and are being more restrained in placing new orders.
Blessing in Disguise
Also, not having a hot product to fuel such over-zealous ordering of semiconductor chips may be a blessing in disguise, said analyst Haloulakos. “This may moderate the boom-bust cycle” and provide for a more stable industry.
Semiconductor makers say the industry doesn’t need another personal computer boom to generate growth. Rather, they say, they’re fashioning leaner, more efficient companies that can produce moderate growth.
Texas Instruments, which lost $118.7 million last year, has had several rounds of cutbacks. In the latest, last October, it cut another 2,200 employees--bringing the year’s total to more than 7,000--closed two factories and froze wages.
Now, said Bill Sick, executive vice president, TI is focusing its market strategy in four areas, including ASICs and semiconductors for the military. He said, “TI is the largest supplier to the military, by far, and the military market will grow.”
Analysts agree. “The military and aerospace markets have provided a steady, stable source of demand and will continue to do so,” said analyst Haloulakos. For instance, he said, in 1984 cars, the semiconductor content was about $40. He expects that will grow to $100 in five years.
And despite the industry’s intentions to become more stable, this is not a dull time for semiconductor makers, say the analysts. “I think it’s going to be even more exciting, because new developments will occur more often; (the chips) will have wider and better applications than before” as they reach more consumer appliances, said Haloulakos.
Revels at Advances
And Dataquest’s Bogert revels in the advances in miniaturization and power performance. “It means you can put something in your pocket instead of the closet. . . . I joined the semiconductor industry in 1960. And you can take everything the whole industry produced that year--everything--and fit it in your pocket today. It’s just miraculous.”
Still, the quest for relief from the reckless swings of the past has not diminished the fond hope that another gee-whiz product will come along.
While it is true that one isn’t on the horizon, the hot products “were not ever predicted,” said analyst Andrew Neff of Montgomery Securities. “No one predicted video games, no one predicted personal computers. . . . This is an innovative industry and by definition it’s driven by unpredictable elements.”
SEMICONDUCTOR MARKETS: U.S. Hit Hardest in 1985 in billions
% 1984 1985 Change U.S. $11.6 $8.0 -30.9% Japan 8.0 7.3 -9.1% Europe 4.7 1.5 -7.1% Rest of 1.7 1.5 -8.6% world 1.7 1.5 -8.6% Total $26.0 $21.2 -18.5%
Source:Montgomery Securities
SEMICONDUCTOR SUPPLIERS: How They Ranked in 1985 Ranked by annual sales, in millions (1984 rank in parentheses)
1. NEC, Japan (3) $1,984 2. Motorola, U.S. (2) $1,850 3. Texas Instruments, U.S. (1) $1,766 4. Hitachi, Japan (4) $1,671 5. Toshiba, Japan (5) $1,459 6. Philips, Netherlands (6) $1,068 7. Fujitsu, Japan (9) $1,020 8. Intel, U.S. (8) $1,020 9. National Semi, U.S. (7) $940 10. Matsushita, Japan (12) $906
Source:Dataquest
