In a narrow corridor here just south of San Francisco, a cluster of some 40 companies, small to medium-sized, performs one of the world's modern industrial miracles.
They produce -- with a labor force that is 98 percent unskilled -- a major share of the key component of the $100 billion electronic equipment industry. This is the awkwardly named "semiconductor,"" or tiny silicon chip, which, by holding many thousands of bits of information, forms the heart of computers and "smart" appliances.
Appropriately enough, this area, where it all got started, is called Silicon Valley. The low-lying, heavily guarded production facilities dot the landscape from here south to Sunnyvale, Mountain View, Cupertino, and Santa Clara.
Mythology has it that closeness to Stanford University and the West Coast defense industries resulted in the heavy concentration. But reliable sources say the bunching of plants in Silicon Valley was caused by nothing more complicated than the desire of the key original scientists to work in a sunny climate.
It has been a booming industry without parallel. "What we are talking about is a semiconductor industry providing the components for an $800 billion electronics industry by the late 1980s," said Jerry Sanders, founder and president of Advanced Micro Devices Inc.
Over the past decade, the compound annual rate of productivity growth for the U.S. semiconductor industry has been a spectacular 22.5 percent -- and prices have been reduced steadily. The first digital electronic computer, MIT's ENIAC, weighed 50 tons, cost $1 million, and occupied 3,000 cubic feet of space. A computer equivalent to ENIAC's relatively primitive functions could be held in the palm of a hand, and could be reproduced today for less than $50.
Sanders pointed out that just a few years back, a single-function (known as a discrete) integrated circuit sold for as much as $50. By 1978, a semiconductor with 4,000 circuits on one chip, having random-access memory, sold for only $5.40.
That chip, known as a 4K RAM, thus costs about one-tenth of one cent per function. But the hottest item in the business today, the 16K RAM, which stores 16,000 bits or functions, sells for about half the price per function of the 4K RAM. The cost reduction is thus on the order of 100,000-fold from that single-function $50 semiconductor.
This technical spurt has staggering implications for industry and individuals. The auto industry, a Johnny-come-lately to semiconductors, will be a major consumer. And individuals can look forward to a pocket-sized personal computer (not a mere calculator) more powerful than any business office machine now available. Probable cost: under $200.
But despite the golden technological era that seems to lie ahead, and their own super record of performance, the small companies here in Silicon Valley -- where it all began -- have banded together to ask for government help.
Over their shoulders, they see the Japanese coming.
"The U.S. color T.V. industry, didn't know what hit it until it was just about over," says one semiconductor analyst. "And we don't want that to happen to us."
They complain, moreover, that in Japan's protected home market, only Texas Instruments has won the right to invest in semiconductor manufacture. That successful penetration by TI came about when the big Texas company found Japan infringing its patents -- and threatened to close down Japanese production of calculators unless Japan allowed it in.
Thus, what one finds here is an international trade war story with a new twist: The high-technology Silicon Valley producers, through their Semiconductor Industry Association (SIA), are looking for protection of the kind normally sought by a declining industry.
Interesting enough, large semiconductor producers independent of the SIA, such as Texas Instruments, vieww the problem in a different light. They suggest that any headaches now being suffered by California producers were brought on by short-sighted management.
In the past four years, the Japanese have made vast inroads into the 16K RAM market here, going from a 2 percent share to 35 percent in 1979. According to the SIA, Japan's share of the world market was 42 percent last year, with the prospect of hitting 50 percent in 1980.
There is a dispute within the industry on how the Japanese managed to get such an acceleration in volume in such a brief period. The Silicon Valley producers make two claims: first, that the 1974 recession prevented them from boosting capacity adequately; and second, that the Japanese at one time undercut their domestic market prices for 16K RAM's in order to squeeze American companies' profit margins.
But at Texas Instruments, high officials say that the Silicon Valley producers misassessed the breadth of demand -- and when they failed to expand sufficiently, the Japanese sensibly took advantage of the opening.
Earl Rogers, president of Precision Monolithies of Santa Clara, Calif. -- one of the smaller companies -- agrees. "I think in many cases, our chief executive officers have overreacted," he said in an interview.
"What it really gets down to is that a tremendous shortage of 16K RAMs developed in the last few years, and the Japanese were there with a high-quality product, fairly priced, and were able to deliver parts on time that met the 'spec.'
"If you go back, you find that the fundamental problem is that the U.S. companies did not expand to fill the need -- not because they couldn't raise capital, but because they didn't have the guts to go out and build capacity during the (1974-75) slowdown."
But for whatever reasons, the competition is there. "The U.S. semiconductor industry now stands at a crossroad," according to an exhaustive Commerce Department study.
The fabulous growth of the semiconductor industry so far has been based on the theory that for any given semiconductor there is a determinable "learning" or "experience" curve in production. That means that for every doubling in production volume, the average selling cost can be reduced by one-third to one-half.
That's why market share is the all-important credo in semiconductors, accoring to Lionel Olmer of Motorola Inc.
"As an industry, we are just about where radio was in the 1920s," said John L. Nesheim, treasure of National Semiconductor Corp. "All of a sudden, the world has learned how to use this thing. We have a reliability that is superior to electromechanical devices, which wear out. Plus, lower costs and reduced energy use."
Worldwide sales of semiconductors for last year are estimated by the SIA at $9.8 billion, of which the U.S. share was $6.1 billion, or 62.5 percent. Japan had $2.6 billion, or 26.5 percent, with the rest in Europe (mostly Dutch and German).
By any objective standard, Europe is so far behind the United States and Japan in current technology that most Common Market governments are in an open race to preserve their own markets by tax and other subsidized help to their national companies.
The growth of the U.S. companies has been spectacular. It all began in 1947, when three scientists at American Telephone & Telegraph Co.'s Bell Laboratories invented the transistor to replace the vacuum tube. All of the tube's capabilities could be put on a tiny silicon chip. In 1961, the scientific geniuses spawned by the transistor revolution -- some of whom are now the chief executives officers of companies here -- figured out how to put more than one transistor on the same chip.
By 1977, an inventor at Intel put an entire computer on a chip the size of a freckle, and the industry had its first microprocessor. Since then, there has been increasing sophistication, and there appears to be no upper limit to the number of functions or bits of information that can be stored on a single chip. The design of the complicated circuits is computerized, and the final drawing is created by a computer-controlled photocomposition process.
At the heart of the miniaturization process that allows hundreds of thousands of electronic circuits to be placed on a tiny silicon ship is a complex photo-lithography technique. Ironically, the $20,000 to $30,000 cameras used are -- you guessed it -- Japanese. "We need the very best," said Gordon Moore, president of Intel.
The chief concern at a top level in the industry is that all of this extraordinary technology will be overtaken as the Japances leap-frog their way forward. Currently, the government in Tokyo and the major Japanese companies are working together on a project known as "VLSI," for Very Large Scale Integrated Circuits.
The U.S. industry's nightmare is that VSLI could give the key Japanese companies -- Nippon Electric Co., Hitachi Ltd. and Fujitsu -- a lead not only in volume, but in quality. An associated concern is that the Europeans, now behind, will catch up through acquisitions. Of 28 American companies sold or consolidated in the past decade, 18 have been picked up by Japanese, British, German or other foreign investors.
Ropers of Precision Monolithics -- while he thinks that the SIA has exaggerated the overall competitive problem -- asserted the Japanese are already ahead in quality. He noted that Nippon Electric has a rejection rate of a mere 0.2 percent, or one-twentieth of the 4 percent rejection rate of the U.S. industry.
The ultimate target of the Japanese VSLI project is not only a big technological leap into 64K RAMs and 256K RAMs, but the U.S. computer industry itself, according to Texas Instruments President Fred Bucy. The general assumption is that the Japanese intend not merely to catch up with, but surpass, IBM in worldwide sales of the next generation of computers.
But for the moment, what worries U.S. semiconductor makers most is Japanese competition in the most advanced commercial state-of-the-art technoloagy, the 16K RAM. The essence of the complaint is that the Japanese have "targeted" the 16K RAM, now their most profitable item.
All agree that the 16K RAM is the take-off point for the next generation of the technological revolution, which will be based on 64K and 256K RAMs (which two U.S. and three Japanese producers already have off the drawing boards and into sample production).
The demand for 16K RAMs is so heavy for use in large computers, telephone switching systems, industrial controls, automotive electronic ignitions and a variety of consumer products running from watches to electronic toys that only the manufacturers' best customers get prompt delivery. Both Japanese and American producers keep some customers waiting six to eight months for their 16K RAMs.
In an interview at Intel headquarters in Santa Clara, Vice President Robert N. Noyce said that the trade battle between the United States and Japan could come down to "a battle of financing." If the U.S. industry doesn't get tax help its market share (already shrunken from a commanding 88 percent in 1963) will be reduced further, he warns.
A "national industrial policy" is needed if the semiconductor industry is to be preserved "as a dynamic force in our economy," he warns.
Noyce, one of the pioneers in the industry -- in January, President Carter awarded him the National Medal of Science in recognition of his work -- notes that "we protect the shoe industry when it's dying or the textile industry, when clearly it would be advantageous to move it to another part of the world." Helping the semiconductor industry "would cost a small amount, compared to bailing out the geriatrics," he urges.
What will happen if the industry doesn't get the help it feels it needs? Some Silicon Valley companies would jprotect themselves by concentrating on a "specialty niche" within the market.
"But you can certainly envision horror stories," Noyce said. "We had a very strong TV industry in the U.S. in 1965. It was hard to envision RCA going out of the TV business. But if I'm not mistaken, the RCA video camera that I bought the other day was built in Japan, and that would have been inconceivable 15 years ago. So maybe IBM's computers will be built in Japan, too. It's almost as inconceivable that that would happen as that I'd buy an RCA video camera made in Japan."