Noyce and his fellow semiconductor executives worried that ripple effects from losses in the DRAM market could threaten the entire American industry. The DRAM was what the industry called a “technology driver”—a relatively simple chip produced in such large quantities that it enabled the manufacturers to hone skills that they could then apply to the production of more complex devices. In other words, fewer 16K DRAM chips produced by Americans in 1979 would mean less efficient production of other, more profitable chips in 1981. “Capturing the 16K market is like taking the top of a hill,” explained AMD founder Jerry Sanders. “Once you have it, all you have to do is shoot down.”7
Americans in other industries had learned in the most painful way possible not to underestimate the Japanese. Although “made in Japan” was once synonymous with shoddy workmanship, after World War II, the Japanese had developed into world-class manufacturers and business leaders. By 1985, Nippon Steel outproduced U.S. Steel, and a Japanese bank (Dai-Ichi Kangyo) was the world’s largest. Japanese televisions had driven American products out of the market. America’s trade deficit with the island nation ballooned to $40.7 billion—almost 40 times the deficit just one decade earlier. As the 1980s drew to a close, 68 percent of Americans named Japan as the number one threat to the nation’s future.8
As early as 1978, Noyce was warning that historians might someday note that “in addition to originating and nurturing a vibrant semiconductor industry, the United States also lost it—the same way we have lost the steel industry and the TV market—to foreign competition.” In just the previous eight years, 19 United States semiconductor firms had been partially or entirely purchased by foreign investors.9
The prospect that the industry he had helped to launch might soon die on American shores absolutely infuriated Noyce. He certainly was not a racist—he pushed for eased immigration laws, praised the “brilliant minds from places like India and the Orient which have made such major contributions to high technology,” and had been as harsh in his criticism of European semiconductor firms when Europe had a protected market as he was of the Japanese—but his comments about the Japanese could be spiked with uncharacteristic racial nastiness and hyperbole. He told Fortune magazine that the Japanese were “out to slit our throats,” a comment that he later admitted got him “almost thrown out of Japan,” and he warned that “the time to throw up a block to a karate chop is before it lands.” When a Japanese semiconductor executive referred to United States firms as “boutique semiconductor companies” (thus implying that Japanese firms were more efficient manufacturers), Noyce started yelling at him—highly unusual behavior for a man so shy of confrontation that his nickname was “Dr. Nice.” When as a joke, one of Noyce’s friends made him a cake covered with paper Japanese flags, Noyce pulled out his cigarette lighter and burned every one of them.10
He was not alone in his feelings. In 1977, Noyce, Fairchild president Wilfred Corrigan, AMD president Jerry Sanders, National Semiconductor president Charlie Sporck, and Motorola vice president John Welty decided to join forces to counter the threat to their industry. Three of these men had worked together at Fairchild, and all five had known each other for nearly a decade. This group founded the Semiconductor Industry Association (SIA) with plans to harness the industry’s cooperative spirit and innovative thinking in the service of what one founder described as the SIA’s “essential task”: “slow down what the Japanese government is doing in support of its industry, and speed up what our government [is doing].”11
The two governments’ approaches were indeed quite different. Where the Americans had taken a fairly laissez-faire approach to fostering business development, Japan’s Liberal Democratic Party, which had governed the island nation since the mid-1950s, actively fostered the country’s industrial sector. This meant importing raw materials—Japan had almost none—and exporting high-value-added finished products manufactured from those materials.
Until 1980, Japan’s Ministry of International Trade and Industry (MITI) held exclusive power to grant import and export licenses. In the 1950s and 1960s, MITI focused on two export industries: automobiles and steel. The government offered companies in these industries low-interest loans and developed national policies to shut established foreign (usually European and American) competitors out of the domestic Japanese auto and steel markets. This government-sponsored protection and support enabled the target industries to prosper, and eventually, to challenge the very European and American competitors once excluded from the Japanese market. As the steel and automobile industries strengthened in the 1970s, the Japanese government used many of the same policies to support consumer electronics and the semiconductors that made them work.
Partially as a result of their governments’ approaches, the Japanese and American semiconductor industries were structurally quite different from each other. The prototypical Silicon Valley semiconductor firm in the 1980s was the independent, “merchant” producer, which manufactured chips for end users other than itself (although Texas Instruments, IBM, and Motorola were vertically integrated companies). Merchant companies often got their start as entrepreneurial ventures; depended on continuous technological innovation and mass production; and before the 1980s, had traditionally regarded the government’s proper role to be that of an eager customer with very deep pockets.
In sharp contrast to this model, the Japanese semiconductor industry consisted of six huge, vertically integrated electronics firms (Nippon Electric, Fujitsu, Hitachi, Toshiba, Mitsubishi Electric, and Oki Electric Industry), which had been developed with direct assistance from the government for the express purpose of growing Japan’s high-technology sector. These firms manufactured not only chips, but also the electronic equipment—such as computers or VCRs—that used the chips. In fact, in 1979, only 7 percent of the companies’ combined revenues were derived from semiconductor sales.12
The Japanese government conferred substantial benefits on these semiconductor firms, imposing tight restrictions on foreign imports from competing manufacturers, subsidizing and organizing national research projects, developing programs designed to funnel talented students into fields such as engineering, and permitting a flexible application of antitrust laws. Equally important was the signal that such government actions sent to the Japanese banking community. Japanese banks, many of whose directors also served on the boards of the electronics companies, were willing to make loans to companies in such clearly “favored industries,” despite the firms’ debt ratios, which would have been considered astronomical by American standards. As a result, Japanese semiconductor firms found it much easier to obtain capital than did American companies.13
The difference in access to capital launched the Semiconductor Industry Association’s first serious lobbying push. In February of 1978, Noyce, who headed the SIA’s Trade Policy committee and would serve as chair of the organization’s board, traveled to Washington, D.C. on behalf of the SIA to testify before a United States Senate Committee on Small Business about the need to lower the capital gains tax, which stood at 49 percent in 1978. The fact that nearly half of any capital gains would go to the federal government had contributed to a general drying up of venture capital, and less available capital put American high-tech businesses at a competitive disadvantage relative to the Japanese. Many other business groups—including the American Electronics Association (formerly WEMA), which had invited Noyce to testify—had their own reasons for wanting the capital gains tax lowered and vigorously lobbied for a reduction.
Their efforts paid off. At the end of 1978, the Carter administration reduced the tax to 28 percent. The lower capital gains tax—coupled with nearly contemporaneous changes easing the “prudent man” rules that had restricted pension funds’ abilities to invest with venture capitalists—had dramatic effects. By one estimate, within 18 months of the changes, the amount of money flowing into professionally managed venture capital companies each year shot from $50 million to nearly $1 billion.14
But the decline of
the American semiconductor industry relative to the Japanese did not slow. Indeed, just as the SIA had feared, the American share of the worldwide semiconductor market began to slip in areas other than DRAMs. The Japanese began cutting prices so dramatically that Intel, accustomed to dropping prices approximately 30 percent every year, found itself forced to cut the price of its 16K EPROM, once the company’s cash cow, by 90 percent in just 18 months. The American industry was convinced that the Japanese were “dumping” their chips on the American market—selling them well below the cost of production—and then planning to raise prices once their foothold in the American market was established. The situation was desperate enough that Intel furloughed 2,000 employees and in 1982 allowed IBM to acquire a 12 percent interest in the company in exchange for $250 million. Other companies suffered as well. From 1981 to 1982, AMD’s net income fell by two-thirds, and National Semiconductor went from a $52 million annual profit to losses of $11 million. Around this time, Noyce started an SIA meeting with an offer to “lead the group in prayer.”15
EXCEPT FOR A BRIEF UPSWING IN PROFITS in 1983 and 1984, things just got worse. The Japanese market continued to be closed to foreign sellers: United States firms manufactured less than 10 percent of the chips sold in Japan, whereas in other export markets, American-made devices accounted for about one-third of the chips sold. Meanwhile the share of the total market for semiconductor devices supplied by Japanese manufacturers continued to rise until, in 1985, the once-unthinkable happened: Japan’s share of the total world market for semiconductor devices surpassed that of the United States.16
Noyce estimated that between 1984 and 1986, the American semiconductor industry lost $2 billion in earnings and 27,000 jobs. In that same period, 13 percent of the electronics jobs in Silicon Valley disappeared. Adding insult to injury, the Japanese electronics giant Fujitsu began maneuvering to take over 80 percent of Fairchild, the granddaddy of Silicon Valley semiconductor firms. Fairchild was already owned by a foreign company—French conglomerate Schlumberger had bought it in 1979—but the prospect of Japanese ownership led the San Jose Mercury News to lament, “The transaction seems to tell us, in one quick message, how far we’ve fallen and what we’re up against.” Industry leaders, including Noyce, briefly discussed some sort of united opposition to the deal, but abandoned the idea in light of antitrust concerns. One well-known semiconductor analyst, when asked about the potential sale, could do little more than shake his head and mutter disgustedly, “Talk about the world being turned upside down.”17
The phrase captured the overriding sense of the times. America was accustomed to dominating high-tech industry, but in many markets the country now lagged in second place. The semiconductor industry had seemed impervious to the recessionary forces of the 1970s, but now it was sucked into the vortex.
In one of the more shocking reversals, Gordon Moore and Andy Grove recommended to the Intel board of directors that the company leave the DRAM memory business altogether in 1985. A DRAM memory—the 1103—had been Intel’s first best-selling product. DRAMs had brought the firm from a two-man startup to the Fortune 500. But now the product line was acting as a net drain on profits from other areas of Intel’s business, particularly microprocessors, because Intel had to price the memories so cheaply. Andy Grove recalls “going to see Gordon [Moore] and asking him what a new management would do if we were replaced. The answer was clear: get out of DRAMs. So, I suggested to Gordon that we go through the revolving door, come back in, and just do it ourselves.”18
Arthur Rock calls the vote to abandon memories in order to focus Intel’s attention on the microprocessors once considered only a sideline business “the most gut-wrenching decision I’ve ever made as a board member.” But Noyce had no second thoughts about approving Moore and Grove’s decision. “He thought the Japanese were already beating the heck out of memories,” recalls Ann Bowers. “He hoped microprocessors would offer a way out.” By the end of the decade, seven of the nine American DRAM manufacturers had left the business.19
And still the problems persisted. In 1986, for the first time since it went public in 1971, Intel lost money—$173 million. AMD, declaring the year the worst in the company’s history, lost $37 million. National lost $143 million. Even after cutting 7,200 jobs (28 percent of the employee base), Intel was still in such trouble that several times its senior managers met in Noyce’s living room to discuss a topic no one wanted to contemplate: “how to shut down Intel, if it comes to that.” An April Fool’s edition of the company newsletter offered its own suggestion in a cover story headlined “Japanese Buy Intel: Grove Named Shogun.” Explained one SIA executive, “It’s hard for someone who did not live through it to imagine how bad things were. We had really begun to think that the American semiconductor industry might cease to exist, [having been] erased by Japanese competition.”20
NOR WAS IT ONLY THE AMERICAN SEMICONDUCTOR INDUSTRY that was in trouble. Noyce was convinced that everywhere he looked in the United States he saw signs of “the decline of the empire.” The federal government was running record deficits, Americans were saving less and spending more than at any time in history, the country was importing more than it exported, and math and science literacy among American students was declining with every passing year. “Can you name a field in which the U.S. is not falling behind now, one in which the U.S. is increasing its market share? We’re in a death spiral.” Noyce told a reporter. He predicted that Silicon Valley might one day become “a wasteland,” one of many around the country. “What would you call Detroit?” he asked. “We could easily become that.” To make matters worse, Noyce believed that the gloom overhanging the industry was dangerous in and of itself. He was convinced, he said, that “optimism is an essential ingredient for innovation. How else can the individual welcome change over security, adventure over staying in a safe place?”21
IF FREEING UP THE FLOW OF CAPITAL had not helped, and dropping unprofitable products and cutting jobs had not worked either, perhaps the answer to America’s economic woes lay in adopting a homegrown version of the Japanese government’s industrial policy targeting a few select industries. Noyce did not think so. He said that the government should “target entrepreneurs” in every field and not single out any particular industry for support. His voice was one of many engaged in the American debate over industrial policy. In one camp were those—generally supporters of President Ronald Reagan—who argued that the correct response to the Japanese threat was a redoubling of current supply-side efforts to revitalize the American economy: reduce regulatory burdens and capital gains taxes; create incentives for long-term investment; and increase government expenditures, generally through military spending. According to this argument, the most useful step that the American government could take was a flying leap out of industry’s way. Government should give more by taking away—via taxes and regulation—less.22
Leading the opposing camp were Democrats who argued that the Reaganites’ traditional means of stimulating the economy had already proven futile against the Japanese. The Republicans’ supply-side approach had led America into this problem in the first place, according to Reaganomics’ opponents: previously unimagined deficits had pushed up interest rates, thereby increasing the value of the dollar to the point that foreign manufacturers could easily undersell their American competitors anywhere in the world, including in the United States.23
Hewing to the philosophy that if you can’t beat an enemy, you should join him, many Democrats suggested that the United States government follow Japan’s lead and play a more active role in fostering its industrial sector. “It is time to moderate the national habit of blanket legislation and broad policy strokes with more attention to the ‘micro’ requirements of individual sectors, industries and firms,” argued Senator Adlai Stevenson III. “The United States is the only industrial country which does not attempt to do this and rejects ‘industrial policy’ in any systematic sense.” In 1983 as part of the LaFalce Plan, House Democrats called for a n
ew Bank for Industrial Competitiveness, capitalized with $8.5 billion in federal funds, which would “make and guarantee loans to older industries in need of modernization and to innovative businesses having trouble getting started.”24
Much more than money was at stake in this debate over industrial policy. At issue was the country’s understanding of itself as a stronghold of free-market, laissez-faire capitalism. Republicans couched their arguments in language calculated to strike fear into a nation historically suspicious of federal power: did Americans really want more government meddling in industry affairs? Did voters really believe that Washington bureaucrats should have the power to determine the “proper” focus for the American economy? (In a strangely self-defeating slap at industrial policy advocates, Republican congressman Dan Lundgren of California alleged that “Supporters of industrial policy have never been able to demonstrate that the ‘best and the brightest’ are in Washington and … can do a better job [than is currently the case] of making the economic decisions affecting our lives.”)25
Industrial policy advocates, on the other hand, claimed that laissez-faire ideals were hollow and that the United States had a de facto industrial policy, administered largely by the Department of Defense. For decades, the federal government, in the guise of defense contracts, had supported research and development in specific target industries such as electronics and aerospace. Moreover, as recently as 1983, the federal government had purchased, largely for military use, more than half of all aircraft, radio, and TV communications equipment; a quarter of all engineering and scientific instruments; and a third of all electron tubes manufactured in the United States. In the words of one industrial policy advocate: “If Japan’s industrial policy has been implemented by … MITI and for purposes of economic defense, then America’s industrial policy (to the extent that it already has one) has been implemented by the Department of Defense and for purposes of its own military defense and that of the free world.”26
The Man Behind the Microchip Page 40