US leads new revolution fought with tiny 'chips'
Essex Junction, Vt. — The Green Mountain Boys marched from these rumpled hills to force the British from Fort Ticonderoga, seize Lake Champlain, and generally stir things up for the sake of independence.
Now, in the shadow of Smuggler's Notch, Americans are launching another revolution -- one fought with "chips" instead of guns.
Nestled on an oxbow along the Winooski River, an IBM plant is making slivers of silicon that are the raw material behind today's revolutionary electronic products -- semiconductor integrated circuits (ICs), the heart of everything from space station computers to digital watches and disco lighting systems.
American's first computer, ENIAC, was a tangle of tubes and switches the size of a three-bedroom condominium. Now, a calculator built around ICs packs the same data-processing power into a space no bigger than a credit card. And these chips have reduced costs as well as size. If the cost of a new car followed the trend in the cost of integrated circuits, an average sedan today would cost less than 1 1/2 cents. A RollsRoyce would cost a quarter.
In 1950, the United States could claim barely a thousand computers. Now there are 10 million. For each one, the fundamental building blocks are integrated circuits.
Industry pundits claim that these semiconductor industry has grown an average of 19 percent each year of the last decade. The US, longtime emperor of the world market (currently worth $11 billion), is facing a tough challenge from relative newcomer Japan. Great Britain, West Germany, and France are struggling to develop domestic production in a business expected to top $100 billion in 20 years.
And all the fuss and bluster is over a fingernail-size chip whose main ingredient, silicon, can be found in ordinary beach sand.
The IBM plant at Essex Junction is on the frontier of this highest of high technologies. It is not the only factory that makes chips, or even the largest. But one of its production lines makes the most powerful IC available in the world.
Like family dynasties, IC models come in generations, and Essex Junction is one of two plants in the US that produce 64K random access memories (RAM) -- a component containing a raisin-size chip that can store 64,000 pieces of information.
Shoehorning this much electronic power onto such a small space requires technology that seems beyond the ken of human knowledge. Designing the chips has been likened to playing chess with 40,000 pieces, and requires banks of computers to help scientists juggle all the possible combinations.
The industry's ability to pack more complex circuits on a given chip size in accelerating so fast that manufacturers have to run hard just to stay in place. The top 10 American companies (Advanced Micro Devices, American Microsystems, General Instruments, Intel Corporation, International Rectifier, Intersil, Motorola, National Semiconductor, and Texas Instruments) spend millions of dollars on new products, returning an average of 94.5 percent of their profits to research and development.
Currently, a normal-size chip contains about 200,000 transistors. One industry source predicts the norm of 1985 will be 1 million, and that by 1990 integrated circuits will contain more than 4 million transistors.
Today's memory chips could be supplanted by "bubble" memories, which store information in tiny magnetic domains or bubbles. Packaged with a permanent magnet, bubble memories won't lose their storage when power fails or if the device is removed from its computer.
"Mainframe" computers could be housed in a broom closet. A microprocessor might manage an entire factory. Consumer products from washing machines to sailboats could be controlled by one tiny piece of silicon.
The possibilities are endless. But while the nation's front pages focus on the prospect of computerized homes and robot secretaries, scientists and businessmen are scrambling in the shadows for what may turn out to be the most gripping story of all.
If the world will be run by computers, and computers are run by integrated circuits, will the country that makes the best chips run the world's economy?
It is 1947. At Bell Telephone's research laboratory in Murray Hill, N.J., William Shockley, Walter Brattain, and John Bardeen are not trying for a technological breakthrough that will found a billion-dollar industry.
They are just interested in why substances known as "semiconductors," such as silicon and germanium, don't conduct electricity very well. After much research they have found that injecting impurities ("doping") into a semiconductor increases its ability to conduct electric current. Building on this discovery, the team notices that when a piece of silicon is doped in specific ways it acts like a "gate" through which electricity can pass only one way.
By doping a third section, the men notice that the resulting device allows variations in a small amount of current in one circuit to superimpose identical variations on a much larger amount of current in another circuit, the principle behind amplification. Shockley, Brattain, and Bardeen have discovered the transistor -- an invention that will amplify signals and act as an electronic switch, will for the most part render vacuum tubes obsolete, and will win themselves a Nobel Prize.
now it is 1977. For the last 30 years US corporations have dominated production of semiconductors. National Semiconductor, Texas Instruments, Intel, Advanced Micro Devices, Motorola, and other companies have taken Shockley, Brattain, and Bardeen's discovery and used it to develop a wide range of solid-state electronic products.
National Aeronautics and Space Administration and US Defense Department contracts have pumped money into the industry at an average of $55 million a year. The immensity of the American computer market allows assembly-line production and drops costs as manufacturing techniques are perfected.
Most of the industry if huddled south of San Francisco, an one-time plum orchards now dubbed "Silicon Valley." The weather is good, the profits are high, and scientists change jobs often, diffusing technical advances.
But out of the West are coming the first sounds of competition. Japan, behind a maze of tariff barriers, has built a semiconductor industry now girding to go mano-a-manom with the US.
The Japanese government has poured more than $350 million into basic research. The major producers -- Nippon Electric, Hitachi, and Fujitsu -- are large corporations that can absorb losses in their semiconductor divisions and still survive.
Today the Japanese incursion into the semiconductor marketplace is an established fact. They have captured 23 percent of the world market, while the US holds 69 percent.
High-technology analyst Handel Jones of Gnostic Concepts predicts that by 1990 the Japanese semiconductor industry as a whole will reach parity with the US.
"I think this willm happen," he says. "The US semiconductor industry is too cocky."
Other industry watchers aren't quite so sure about the outcome of the US-Japanese battle.
"The Japanese are very aggressive competitors," says Dr. H. Gunther Rudenberg of the noted consulting firm Arthur D. Little. "They represent a real threat. But with hard work, the US can hold even."
During the stormy 1973-75 recession, US integrated circuit makers quickly hauled down their mainsails -- slashing capital spending 52 percent and trimming payrolls and product lines. When the economy brightened in 1976, the Japanese quickly snatched 40 percent of the US market for the popular 16K RAM, while domestic makers struggled to hoist production.
"They started way behind the US in memory systems," says Handel Jones, "But now they've caught up." He believes the Japanese will eventually capture 60 to 70 percent of the still-infant market for 64K RAMs.
US semiconductor manufacturers traditionally have behaved somewhat like river-boat gamblers, betting large stakes on new products, hoping for a quick profit. Intel, for instance, is wagering much of its future on a new generation of microprocessors.
The Japanese take a less swaggering approach. "They operate on a much longer time frame," Jones says. "They realize chips are becoming a system instead of a part."
US semiconductor executives say the Japanese can afford to be patient because for the most part they are simply divisions of large; cash-flush corporations. The majority of American IC producers are independent entrepreneurs whose businesses succeed or fail with the success of their semiconductor circuits.
Quality is a more controversial issue, familiar in the annals of Japanese-American trade. A recent Hewlett-Packard "road test" compared three different Japanese 16K RAMs with three comparable American models. The most reliable American chip failed nine times more often than the best Japanese one.
"In the US, quality control means weeding out failures," says one industry executive.* "In Japan, it means building quality in."
It takes a dust particle only one micron wide to ruin a circuit under production. At all semiconductor manufacturers, a seemingly high percentage of ICs fail final inspection.At some plants, as many as four out of five ICs fail.
But with visions of Chrysler dancing in their heads, almost all American companies say they are reorienting their product lines to prevent defects.
The men at the top of the American semiconductor industry have varying responses to the Japanese challenge. Some view the Japanese as just another competitor.
"We see great opportunities in front us. And we know that strong competition is hard on our heels," says Honeywell's chief executive officer, Ed Spencer.
Others are abandoning this stiff-upper-lip approach and accusing the Japanese of erecting discriminatory trade barriers and using unfair business practices.
"There is strong evidence which suggests our Japanese competitors may have employed discriminatory pricing practices in their penetration of the crucial US market for 16K RAMs," testified W. J. Sanders III, president of Advanced Micro Devices, before the International Trade Commission last summer.
"Government needs to know the situation and be prepared to act appropriately as we move into the '80s," says Semi-Conductor Industry Association director Tom Hinkleman.
Fourteen US semiconductor firms have quietly been snapped up by foreign conglomerates in the last few years -- a move that makes industry executives even more edgy. Western European countries are scrambling to develop their own semiconductor industry before Japan and the US blot out all hope of competition. Britain will commit $430 million over the next decade; France will spend $300 million; Italy ans West Germany, more than $100 million each.
"If we are unable to maintain a technological superiority in the semiconductor field," Mr. Sanders says, "we could soon find ourselves dependent on foreign sources for the electronics necessary for defense and communications systems -- at risks and costs comparable to our present dependence on foreign sources of energy."
The political implications of the growth of the semiconductor industry go beyond mere flag waving. An oft-quoted study by Worldwatch Institute points out that the technological revolution spurred by integrated circuits is likely to increase the wealth gap between developed and underdeveloped countries.
If integrated circuits are the "oil" of the future, then the US and Japan may well become the OPEC of the 1990s. How will they handle that role? Should effort be made to disseminate such valuable technology? Will the automation brought about by semiconductors throw people out of work?
A study conducted by the West German semiconductor firm Siemans estiamtes that 40 percent of today's office work could be done by computer. The electronics industry itself, according to arthur D. Little, will open 1 million new jobs over the next seven years in the US and Western Europe.
Experts say it is impossible to predict whether the explosive growth of America's semiconductor industry will produce a net job loss or gain. But the Worldwatch report reaches this conclusion: Because microelectronics provides such a tremendous competitive advantage, "more jobs will be lost in those countries that do not pursue the technology vigorously than in those that do."
At Essex Junction, the second shift is just leaving work. Clambering into Jeeps and pickups, the workers melt back into the hills, going home to their children and wood stoves.
Perhaps a new industrial revolution is being launched in this aggressively rural area. The social changes wrought by semiconductors may literally match those of the steam engine. Already, without integrated chips, space travel would be impossible. Computers would be the size of stamping plants. And watches would all have hands.
The imminent arrival of such exotic products as videophones and home entertainment holograms has been trumpeted for years. But the real microelectronic advances will be less esoteric. Word processors are already as ubiquitous as office water coolers. This year, $1 billion worth of these systems will be sold in the US alone. Production lines for new Ford, Chrysler, and General Motors cars now are partly staffed by "smart" welding robots. The litany of "the computer revolution" is fast becoming a journalistic cliche.
And it all comes from a substance found in a few grains of beach sand, melted and molded into a silver of almost unfathomable complexity.
The remaining articles in this series will deal with agriculture and the textile industry.