Bruised and bloodied by the Japanese, the American semiconductor industry is preparing to fight back with something many experts call one of its Pacific competitor's special weapons, joint research and development.
With government blessing, 13 major U.S. companies have founded a consortium to try to make discoveries in manufacturing technology -- and to reserve them for Americans only. Preliminary work at the organization, known as Sematech, has begun from second-floor offices in Santa Clara, Calif. Assuming that $100 million a year in funding is not stalled in Congress, it will get down to business in earnest late this year.
Sematech brings under one roof members of a diverse industry long known as a stronghold of go-it-alone entrepreneurs. But fear that U.S. makers could eventually be reduced to little more than bit players in the world market for semiconductors -- about $31 billion of them were sold last year -- is bringing about a change of heart.
"This industry pulled together," said Alan W. Wolff, an attorney who has represented many American semiconductor companies. Executives came to realize, he said, that "no single firm is going to make the difference."
Semiconductors are silicon-based devices about the size of a tie clasp. Crowded with microscopic circuitry, they are the basic building blocks of most of today's electronic products -- such as computers, videocassette recorders, weapons control systems and digital watches. They are also showing up increasingly in everyday articles like children's toys and washing machines.
U.S. companies pioneered the industry and for years dominated it around the world. But in the 1980s they have been thrown into severe recession in the face of foreign competition, most of it from Japan.
U.S. makers lost an estimated $2 billion in 1985 and 1986, as prices plunged and foreign chips poured into the United States. Tens of thousands of workers lost jobs. Things have eased somewhat since the trough of 1985, but the industry is nowhere near being out of the woods and is searching hard for long-term solutions.
Semiconductor executives argue that their industry is special, a key to health in the electronics sector and therefore to the economy in general. Lately, military concerns have entered the debate, too: With chips playing an increasingly important role in weapons systems, more and more people in Congress and defense circles are saying the United States cannot risk depending on foreign suppliers for them.
Japanese companies have done best in standardized, mass-produced chips used in computer memories. They now dominate the U.S. market for these chips, with only two U.S. companies surviving in the so-called "merchant" field, production for sale to other companies rather than internal consumption. South Korea has also entered in a small but fast-growing way. The United States is generally regarded as the world leader in lower-volume but higher-value "customized" chips and in general chip design.
U.S. producers allege that the Japanese captured much of their market share here by cheating. In recent years, the Commerce Department has repeatedly declared various Japanese makers guilty of "dumping," or selling at prices below their production cost. U.S. officials also have ruled that some foreign companies pirated U.S. designs.
Enforcement and a year-old agreement on overall semiconductor trade with Japan are helping reduce these practices. But many U.S. semiconductor experts agree that even without them, Japanese sales would have thrived. "The brutal truth is that they manufacture better and with more productivity than we do," said W.J. Sanders III, chairman of the major semiconductor producer Advanced Micro Devices Inc., and a member of the Sematech board.
The U.S. industry is characterized by large numbers of comparatively small and in many cases financially shaky companies that do nothing but make chips. In Japan, production and consumption take places within "vertically integrated" multibillion-dollar conglomerates like NEC Corp., Sony Corp. and Toshiba Corp. One division will develop new production equipment, trying out its new creations in the company's chip division. Then, many of the chips are used in finished products made in yet another division. Financial losses in chips and production equipment can be offset by profits elsewhere in the company, giving more flexibility.
The U.S. corporate culture stresses maximizing quarterly profits; Japan's is more concerned with stability, market share and the long view. The Japanese never stop modernizing their equipment, even if there is no immediate return in sight. One recent study found that about 75 percent of all U.S. semiconductor production lines were using equipment of a type two generations old. In Japan, the figure was 40 percent.
Japanese plants have lower turnover of employes -- unmarried ones often sleep in dormitories on the plant grounds -- than do American factories. By many accounts, too, the Japanese workers are more highly motivated and attentive to the painstaking details of chip production.
Sematech intends to focus on manufacturing techniques, not chip design, and close some of that gap. Its projected cost is about $1.5 billion over six years, split roughly equally between member companies and the federal government, with a few local governments also contributing.
Sematech, which will eventually involve about 800 people, is meant to pool resources and knowledge in hopes of devising important new production equipment and processes. The industry has earlier jointly funded university research, but the new approach would put people from the companies themselves shoulder-to-shoulder. A prototype assembly line is intended to produce no chips for sale, just processes and equipment designs. Member companies would have first rights to buy resulting discoveries for use in their factories, with nonmembers next in line to get them, under license.
But not if they are Japanese or South Korean or from any other foreign country -- for a period of time, at least. Sematech will be an Americans-only club, an oddity in a country that traditionally has allowed free flow of commercial technology to its allies. The idea is to keep its discoveries from leaking long enough to give Americans a leg up. Sanders said the industry evolves so quickly that Sematech will pay off for his company if things are kept back just one or two years.
Sematech in many ways mirrors a project the Japanese pursued from 1976 to 1980 to close the United States' sizable lead in a chip field known as very large scale integration. Japan's electronics giants worked together, spending about $300 million and sharing resulting patents. The effort is generally credited for greatly speeding Japan's gains in the world chip market in subsequent years.
The Sematech idea encountered skepticism in Congress and the Reagan administration when first unveiled. Wasn't this just another ailing industry seeking a federal handout? (The federal government is already set to spend an estimated $400 million to $500 million on chip research this year.) What if the findings leaked to Japan quickly, meaning U.S. taxpayers would be underwriting Japanese companies' new equipment? Or, in the other direction, what if the findings became captive to Sematech members and didn't provide the general social and economic benefit promised?
Moreover, U.S. industrial strength has always been said to lie in its competitive spirit. Sematech might result in narrowing the field of inquiry. Cooperative research has a mixed record in the United States, critics felt. Companies sometimes send mediocre people and save their best ideas for the home lab. "It's a trait of the American character to want to be independent," said Edward E. David Jr., an industry consultant who was involved in evaluation of the project
Despite reservations, David supports putting federal money into the consortium. "After all, Sematech is the only game in town," he said. "We haven't seen any other suggestions."
Sanders and others in the industry say that Sematech cannot solve all of the problems they face, nor is it intended to. But it will give breathing space, Sanders said. "Sematech can do what the little boy did when he put his finger in the dike."
Other industry shortcomings are being addressed elsewhere, he said. For example, centralization is being achieved through acquisitions, such as his own company's recent buyout of Monolithic Memories, he said..
Congress now has before it trade and defense bills containing money, and Sematech says it is optimistic. "We have very, very good support on the Hill," said Managing Director Larry W. Sumney, who is searching for a permanent site and a chief executive. But much of that backing seems to be the result of military rather than economic concerns.
Earlier this year, the Defense Department, citing national security, helped block a Japanese attempt to acquire Fairchild Semiconductor Corp. It also circulated a study calling semiconductors a strategic defense industry and recommending more aid. Sematech's federal funding, assuming it does not fall victim to objections to other provisions of the bills in which it is contained, will be handled directly by the Pentagon.