Peering into a microscope at a greatly magnified computer chip one day last August, Peter Stoll of Intel Corp. saw something startlingly familiar. In one of the tiny cells, two transistors were disconnected from the rest of the chip, and dangled uselessly in their bed of silicon.
Stoll, 33, a chip designer, recognized the defect as a small, last-minute repair job he had performed on Intel's 8086 microprocessor several years earlier. It had worked, correcting the minor flaw in the chip's logic, and the 8086 went on to become phenomenally successful as the "brain" in a wide range of business computers, robots and industrial machinery.
But what startled Stoll was that the chip under the microscope was not Intel's. It was a product of Nippon Electric Co. (NEC) of Tokyo. Stoll concluded that he was looking at a Japanese copy so perfect that it even repeated the small imperfection in the original chip.
Intrigue of that kind in the $13 billion-a-year global market for computer chips has led to U.S. accusations of unfair Japanese practices, ranging from copying to protectionism. Critics of Japan say that its efforts to gain supremacy in computer chips, perhaps the single most important technology of the Information Age, are typical of the methods employed by "Japan Inc."
"We're at war, no doubt about it," said a computer scientist from a large U.S. research laboratory. "If I had money in 'Silicon Valley,' I'd get it out . . . . It's just like any other war zone."
U.S. politicians are in a mood to strike back.
Democratic Reps. Don Edwards and Norman Y. Mineta, from California's so-called Silicon Valley area, have introduced a bill to give copyright protection to chip designs. They say the measure is needed to stop "pirate firms" from "flooding markets with copied designs that undersell the innovating firms."
But some trade specialists caution that there is a Japanese side to this story. For one thing, U.S. companies are holding their own in the competition.
Japan, whose share of the U.S. chip market is well under 10 percent, has made inroads in some kinds of chips, such as memories, that store information. But the United States is dominant in microprocessors, the "computers on a chip" that serve as brains for computers and controls in dishwashers, jet aircraft, missiles, industrial robots, telephone systems, traffic lights and hundreds of other products.
Many experts insist that Japan's progress is not attributable to copying.
"The basis for the Japanese taking an ever larger share of the chip market is not transfer of American technology," said a patent attorney for a large U.S. company. "It's Japanese management, equipment and a degree of cooperation between firms that's prohibited in this country."
Even the issues in the Intel-Nippon Electric dispute about alleged copying of the 8086 microprocessor become fuzzier on closer inspection. Intel contended that NEC wrongfully copied the chip's microcode, the set of internal instructions laid out as a pattern of transistors on the chip's memory. Intel counsel Roger Borovoy said the microcode was copyrighted and could not be used without Intel's permission.
Officials from NEC's U.S. sales company acknowledge that the microcode on their chip is identical to that on Intel's, including the flaw engraved onto the original.
"If you're not 100 percent identical, you're dead. If you take the fatal flaw out, it wouldn't be compatible. We have chosen to be as close to the original as possible," said NEC's David Millet, who is in charge of nationwide marketing of microprocessors.
But NEC officials in Japan and the United States deny that the company did anything wrong, contending that they had a right to produce their own version of the chip under a 1976 agreement allowing both companies to use the other's patents.
NEC officials in this country say the question of whether the microcode can be copyrighted has never been decided in court, and Intel agrees. And they say that NEC even sent Intel a 1979 announcement of NEC's version of the 8086.
The story of the NEC-Intel dispute is representative of the suspicion, tension and, often, grudging admiration that characterize the competition between the two countries. It begins with the markedly different cultures and societies from which the two have emerged. The Roots of Competition
Compared with the 84-year-old NEC, Intel is an upstart company, an example of American boldness and nerve that began with a few dozen employes in Santa Clara, Calif., in 1968 and grew into a business with 19,000 employes worldwide.
Intel's stock in trade has been innovation. Since it was founded, the company has spewed out firsts, including the first microprocessor in 1973. A founder, Robert Noyce, is one of the inventors of the integrated circuit, which became a basic component of modern electronics.
Intel is also a sort of corporate melting pot that, like the nation itself, has drawn its brain power from all over the world. Its current president came to America as a refugee from Hungary in 1957; a senior vice president was born in Hungary, and an Israeli, an Italian and a Japanese are credited with helping to develop several new Intel products.
NEC has succeeded in typical Japanese fashion: through dogged determination, aggressive marketing and initial reliance on U.S. technology, including that of Intel.
From the outset, NEC had financial and structural advantages over Intel. While Intel makes more than 80 percent of its income from the sale of chips, NEC is a conglomerate that produces computers, electrical equipment and other products. Chips account for less than 20 percent of its revenue, so a temporary decline in that business can be offset by gains in other products.
As a member of the influential Sumitomo industrial group, NEC could draw on the financial resources of the Sumitomo Bank and on the marketing connections of the Sumitomo trading company. But Intel has depended for its financing on the vagaries of the U.S. stock market and bank loans. For most of the last 10 years, Intel has had to borrow money at much higher interest rates than NEC.
Until the early 1970s, NEC was no match for American chip makers. The U.S. computer chip industry was expanding rapidly, thanks in part to heavy government spending on chips for the Apollo man-on-the-moon space program and the Minuteman intercontinental ballistic missile.
In 1973, computer scientists in Intel's laboratory scored a major breakthrough with invention of the first microprocessor. This was a watershed not only for Intel, but also in the history of the information industry.
Until then, chips generally had performed only a single task, such as adding, subtracting, multiplying or dividing. Combining those tasks required wiring together several chips on a bulky board. But a single microprocessor chip could perform all those functions. This meant, for example, that one computing chip could run a pocket calculator, shut off a microwave oven, analyze blood or control traffic signals.
It was possible for general-purpose microprocessing chips to replace more expensive, customized ones previously needed by industry. As microprocessors became more sophisticated, they increasingly began to do jobs that previously had required large, cumbersome computers.
NEC claims to have developed an early microprocessor on its own at about the same time as Intel. This chip, the uCom 4, could handle simple tasks such as operating a pocket calculator. But Japanese officials acknowledge that they have had trouble keeping up with U.S. advances in microprocessors. To do so, Japanese companies have repeatedly relied on U.S. patents and "reverse engineering."
Industry representatives make a distinction between reverse engineering, a generally legitimate practice in which one company's designs are used as a model by another company's engineers, and copying, in which imprints of circuitry are taken by using photographic and lithographic techniques.
In the late 1970s, for example, NEC produced a version of Intel's 8080 microprocessor, the first chip complex enough to handle word-processing programs. A new generation of microprocessors was making possible the era of small, compact personal computers, and Intel was again in the lead.
Tomihiro Matsumara, NEC's senior vice president for research, acknowledged in an interview that NEC attempted to make and sell its own comparable chip, "but we did not succeed." So, he said, NEC engineers analyzed the 8080, then laid out their own "completely different" version, using NEC manufacturing techniques.
Roger Borovoy, Intel's general counsel until he left the company last month, said Intel had no objection because NEC had used the 8080 only as a model and not "copied" it.
Japan, he acknowledged, was becoming an innovator in chips in its own right. Between 1974 and 1977, the government had poured at least $300 million into a research consortium that included NEC and five other companies. "They had come a long way with their own development. They'd attained a status of their own," Borovoy recalled.
Evidence of NEC's progress came in April, 1976, when Intel and NEC signed an agreement that enabled each company to use the other's patents. In the next several years, Intel was to utilize several NEC patents for specialized types of chips.
By the late 1970s, NEC, Hitachi, Fujitsu and Toshiba were grabbing significant shares of the world market in memory chips, devices that store information but do not perform the complex tasks of microprocessors. But these companies still had problems with the far more complex microprocessors.
In 1978, a year before NEC completed its version of the 8080, Intel introduced a much more advanced microprocessor, the 8086. It crammed 30,000 transistors onto a quarter-inch-square piece of silicon, producing as much computing power as some 1960s' computers that filled rooms. The 8086 could handle not only word processing but also complex mathematics, and it and comparable microprocessors are being used in most sophisticated personal and business computers, such as IBM's popular personal model.
NEC's representatives recognized that the 8086 gave the United States a decisive edge in silicon brain power. In 1978 they approached Intel about supplying technical aid to produce the 8086 in return for a percentage of the money NEC would get from selling the 8086 in Japan.
But this time, Intel turned NEC down. NEC, in the midst of a U.S. expansion program, was preparing to enter the international chip market in a big way. It had just purchased a California computer memory company called Electronic Arrays and was planning a second California facility for making memories and logic circuits.
"We weren't anxious to help our competitor," an Intel official said.
Instead, Intel made a deal with NEC's Japanese rival, Fujitsu. Thwarted, NEC decided to go ahead with a version of the 8086 without special help from Intel.
NEC's Matsumara acknowledged that the resulting chip is "interchangeable" with the Intel version, but he strongly denies that it was "copied." Similarly, Robert Hinckley, an attorney for NEC in San Francisco, contends that NEC had a right to reverse-engineer the chip because of the patent cross-licensing agreement of April, 1976.
NEC officials said it was no secret that they would produce the 8086. Electronic News reported it and, NEC officials said, they sent a copy of their announcement to Intel and received no protests.
NEC, however, had several problems.
For one thing, the Japanese company apparently had difficulties reproducing a version of the Intel device without American help. It was not until 1980, two years after Intel's 8086 appeared, that NEC's comparable chip was sold in the United States.
There was also the problem of Intel's copyright on the chip's microcode, a sort of brain within a brain. It is the part of the microprocessor that takes electronic commands from a keyboard and tells the rest of the chip's parts what to do with the commands and in what sequence.
Like a video-game cartridge, the microcode is a computer program that has been written by a programmer and then is built into the chip. In a Pac Man videogame, the microcode tells the Pac Man what to do. In a microprocessor, the microcode tells a computer what to do. Although the microcode appears in the 8086 as hardware--a pattern of 10,752 tiny transistors--Intel maintains that it is not a mere piece of electrical circuitry but is "intellectual property" covered by copyright law.
Copyrighting the microcode had seemed to Borovoy a way to protect the company's intellectual effort from infringement. Borovoy said his "knees wouldn't shake" at bringing a lawsuit against a company that copied Intel's microcode.
But Hinckley, NEC's San Francisco attorney, said no cases have been adjudicated establishing any company's copyright claim on such material.
"Copyright is designed to protect works of authorship--artistic works--and we don't think microcode qualifies," he maintained.
Whatever the merits of their respective cases, NEC and Intel reached a settlement on the 8086 in March after several months of negotiations and without litigation. Borovoy, who said he could not discuss details of the settlement, said the agreement would save hundreds of thousands of dollars in court costs. The Battle for Market Share
But the dispute over the 8086 is seen at Intel as only one chapter in what will undoubtedly be a continuing battle.
"The Japanese see themselves locked in a warlike struggle, determined singlemindedly to reach their objectives by any means, regardless of the impact on the U.S. . . . It's going to be a very, very bloody battle out there," Intel's Noyce said.
He argued that Japanese tactics have denied American companies the fruits of their innovation, profits that enable them to pour money into creating new technical breakthroughs needed to maintain the U.S. lead.
U.S. studies have accumulated a mass of evidence buttressing Noyce's contention that the Japanese government has shielded local chip companies from U.S. competition while they prepared for an onslaught on traditional U.S. markets. U.S. companies have never been able to capture more than 20 percent of the Japanese chip market even when their technological lead was overwhelming.
Before 1978, only Texas Instruments was permitted to establish a wholly owned manufacturing subsidiary in Japan, and even TI had to share some of its patents with Japanese companies to secure that concession.
Few deny that the Japanese challenge is serious. Japan is running a $250 million trade surplus with the United States in chips. And NEC and Hitachi ranked just behind Motorola and Texas Instruments as world leaders in sales last year.
A detailed study issued in February, 1982, by the congressional Joint Economic Committee warned that the main casualties of the relentless Japanese export drive could be small, innovative Silicon Valley companies. With them out of the running, it warned, Japan would be in a position to beat the United States at innovation.
Some industrial experts say the United States should keep its sense of perspective as it responds to Japan's challenge.
Robert B. Reich of the Kennedy School of Government at Harvard University said Japanese chip companies made headway after 1975 primarily because they plunged ahead while U.S. companies, hard hit by the recession, "stood still."
U.S. companies have recently regained some of their lost share of the world market in memory chips and still have an impressive lead in microprocessors. In typical U.S. fashion, Intel is on the verge of marketing an even more advanced microprocessor, the 80386, which the company claims will be far ahead of anything produced in Japan.
Intel has also announced that it will soon sell the first magnetic, bubble-type memory capable of storing 4 million bits of information, the equivalent of 240 typewritten pages.
"Despite trade barriers and protection and copying, we're still winning, although that's no guarantee for the future," said Bob Derby, who ran Intel's marketing operations in Japan.
That, free traders say, should be a warning to those in Congress who want to wield the big stick of government retaliation in the computer chip battles with Japan. NEXT: Hitachi's coup in lasers