The Man Behind the Microchip Read online

  With his young family, rising income, high education, and new home, Noyce was an avatar of the changes sweeping the San Francisco Bay Area at the end of the 1950s. More than 3,000 people moved into Santa Clara County every month. Some came explicitly for opportunities in the electronics industry. Others were drawn to the Bay Area by Stanford and Berkeley or by jobs at Lockheed, Admiral, or Kaiser Industries, among the area’s largest employers. Almost everyone made the trek for the same reasons that had attracted Noyce five years before: the weather, the proximity to the mountains and the oceans, the easy access to San Francisco, and the near-mythic allure of the Golden State. Because the immigrants to the areas closest to Noyce’s home tended to be young and nearly as well schooled as he was, Noyce could watch his neighborhood and the towns around him grow younger, wealthier, and better educated with each passing month.60

  Even the buildings were edgy, new, modern, and built on idealism. In towns up and down the Peninsula, Joseph Eichler, a liberal developer who lobbied against discriminatory housing practices, was building airy community centers surrounded by affordable, high-concept homes with open floor plans, flat roofs, glass atriums, and radiant heat. IBM’s new facility in San Jose, designed by modernist architect John S. Bolles to encourage workers’ contemplative activity, was 63,000 square feet of steel and floor-to-ceiling windows with views of the lushly landscaped grounds, banded with a frieze of rectangular yellow, brown, and gray ceramic tiles.61

  The year 1960 marked the first time that electronics sales on the Peninsula surpassed $500 million. Of the firms engaged in this now half-billion-dollar business, nearly two-thirds were less than a decade old. A test equipment company was born in Palo Alto, a printed circuit firm in Menlo Park. A new Palo Alto-based technical services operation designed and fabbed prototype components, while a crystal-growing facility in Mountain View (founded by another refugee from Shockley) specialized in the manufacture of pure silicon ingots. The number of tenants in the Stanford Industrial Park increased sixfold in five years.62

  The concentration of firms benefited Fairchild Semiconductor, which could use the mass spectrometer at Lockheed and ask the Bay Area Pollution Control Lab to perform a series of important experiments on silicon oxide. Fairchild could have a Menlo Park firm deliver de-mineralized water, purified to the precise standards the lab required for washing components and mixing chemicals. They could have lenses ground at a company a few miles down the road. It was hard to believe that only two years before, Moore needed to build his own furnaces and Noyce had to scrounge for photolithography lenses at a camera shop.63

  Sweeping developments unrelated to electronics also benefited Fairchild Semiconductor and Bob Noyce. The increasing mechanization of agriculture in California freed up thousands of low-skilled workers for work in electronics assembly plants. An aggressive state-sponsored infrastructure-building spree changed zoning regulations and installed a network of roads and sewer pipes to attract people and industry to California. In the two decades after the end of the Second World War, the state of California also established its consolidated system of 9 universities, 19 colleges, and 106 community colleges, which could provide an educated workforce for high-tech industry.64

  Noyce had never before felt as at home in a place as he did in this patch of California, where so much was new, and life changed so quickly. With the dashing John F. Kennedy just elected in Washington, not just this state but the entire country, Noyce thought, seemed full of possibility and promise. He felt, he said, as if the “world were [his] oyster.” It was a time when he truly believed “you could do anything you wanted.”65

  NOT EVERYONE AT FAIRCHILD was as happy as Noyce. His move to general manager and Moore’s accompanying promotion to lead R&D elevated these two founders above the other six in the corporate hierarchy. Only Vic Grinich, who was named associate director of R&D, held a management position even approaching the level of responsibility accorded Noyce and Moore. Marketing, another key part of the company, was under the control of Tom Bay, who was not one of the founders. Charlie Sporck, a blunt-talking pillar of a man who joined Fairchild from General Electric in 1959, had leapfrogged both Eugene Kleiner and Julius Blank for the top manufacturing position. Several of the founders chafed under the new stratification and resented feeling, as Jay Last put it, like “just another employee working in a research lab for somebody else.”66

  Every founder, regardless of his position in the new management order, noticed that Fairchild Semiconductor was slipping from his grasp as it contributed an ever-increasing portion of the profits of Fairchild Camera and Instrument. Within the boundaries of Fairchild Semiconductor, Noyce and the men who worked for him determined how to allocate resources, whom to hire, which customers to pursue, which products to build, and when to introduce them. But the Fairchild founders had no influence over some of the biggest decisions affecting their company because Fairchild Semiconductor now belonged to Camera and Instrument and had no representative on the parent company’s board. John Carter controlled the size of the Semiconductor budget. He determined how its profits were used and whether or not its employees should be granted options on Fairchild stock. Stock options were a particularly sensitive area. Noyce wanted them for senior scientists, engineers, and managers, at the very least; Sherman Fairchild, schooled in the East Coast, big company, gold-watch-on-your-thirtieth-anniversary-with-the-firm approach to benefits, considered stock options “creeping socialism.”67

  The Fairchild founders could only watch with dismay as John Carter, misinterpreting the success of the Semiconductor acquisition as evidence of his own ability to rescue struggling technical businesses, expanded Camera and Instrument’s operations almost willy-nilly. Profits from Semiconductor were not reinvested in the division or shared with its employees; instead, they went to buy a cathode tube company and a firm that manufactured offset printing presses and other printing supplies. In short order, Camera and Instrument expanded into space research, oscilloscopes, office equipment, home movie cameras, and stamp machines.68

  What was hardest for the men who started Fairchild to swallow was that Sherman Fairchild owned roughly 100 times more stock than any founder. Camera and Instrument CEO John Carter owned nearly twice as much stock as did the group of eight combined.69

  BY THIS POINT, Noyce’s attention was focused almost entirely outside the lab. The new diode plant was slated to open in March 1961. Noyce gave a great deal of thought to the job responsibilities of plant manager (“o[ver-head] control, run prod[uction] show, marketing liaison, [good] relations [with headquarters]”) and personally interviewed every potential candidate for the job. At the same time, Noyce was trying to convince IBM to buy more Fairchild products. He orchestrated an elaborate meeting with IBM to feature Fairchild in the best possible light. There would be a presentation on the integrated circuit, another on the lab, and a third, by Charlie Sporck, to emphasize that Fairchild was an up-and-coming volume manufacturer (“running 3 M[illion devices]/yr [with] capacity of 25 M[illion]/ yr”). Noyce would close with a discussion of Fairchild as a trustworthy supplier: we are “not guilty of talking about non-existent products,” he planned to say. “Others will copy [us]—always have.” All this planning came to naught; aside from its purchase of Fairchild’s first 100 transistors, IBM was never an exceptionally important customer for Fairchild.70

  Meanwhile, a senior Camera and Instrument executive had heard that Noyce and Hodgson wanted Semiconductor someday to gain a foothold in the European market. The Camera and Instrument man happened to know the uncle of an executive at Italian business machine giant Olivetti. Olivetti had just started a joint venture called Societa Generale Semiconduttori (SGS) with Telettra, a young microwave company. SGS built germanium transistors and silicon diodes and rectifiers in a modern glass-fronted factory located in a northeastern suburb of Milan. If Fairchild took a stake in SGS, the joint venture would build and sell Fairchild silicon planar transistors, as well. Fairchild Semiconductor would not need to put up any
capital or divert significant resources away from United States markets, which were far larger than the fledgling European demand. In fact, all the American firm needed to do was teach the Europeans how to build silicon planar transistors. In exchange, the Americans would have exclusive rights to sell SGS germanium transistors in the United States.71

  Noyce and Bay liked the idea of penetrating the European market with no capital outlay, but they were unsure it merited trading their proprietary state-of-the-art knowledge about the most exciting development to emerge from Fairchild labs. They did not officially object to the plan, however, and the next thing they knew, Fairchild Camera and Instrument had signed the papers to establish the joint venture.

  In very short order, Noyce found SGS a big headache for which even the trips to Italy could not compensate. Fairchild was obligated to send several top production people across the Atlantic to teach SGS engineers how to build silicon planar transistors—a loss Fairchild felt keenly. The SGS venture probably gave Fairchild a two-year lead in the European market relative to what the company could have achieved on its own, but Noyce and Hodgson, who served as directors of the joint-venture, had difficulty influencing decision making at SGS because representatives from the two Italian companies tended to vote together—and against Fairchild—on many issues before the board.

  Also preoccupying Noyce were nearly $8 million in military subcontracts with Autonetics, a division of North American Aviation, which was an associate prime contractor to the air force on the Minuteman intercontinental ballistic missile. Autonetics expected to use roughly 1,000 Fairchild mesa transistors on the missile contract and required Fairchild to establish a Reliability Evaluation Division (at a cost of nearly $1 million to Fairchild) that would demonstrate the reliability of the NPN transistors. This special reliability division of the company, which would be administered entirely separately from the rest of Fairchild, needed to accumulate 150 million hours of data on how the company’s transistors performed under stresses similar to those the devices would face in a missile shot. Noyce was trying to convince Autonetics to accept planar transistors instead of the mesa transistors originally specified, since he expected planar yields would soon match mesa yields and thus saw no reason to continue making mesa transistors given the greater reliability of the planar devices.72

  OF ALL THE DISCONTENTED FOUNDERS at Fairchild Semiconductor, none was more disillusioned than Jay Last. He thought Fairchild Semiconductor was giving the product on which he worked—the integrated circuit—short shrift. At the end of 1960, Gordon Moore had told him that he planned sizable cuts in the integrated circuits group. Indeed, the only significant personnel reductions in the R&D budget issued in mid-January, 1961, come from the “microcircuitry” group. It was at about this same time that Moore announced at an R&D staff meeting: “OK, we’ve done integrated circuits. What’s next?” Last, on the other hand, felt their work had barely started.73

  Jay Last also recalls a general staff meeting in November 1960, at which Tom Bay suggested that the integrated circuit project be scaled back or shut down “since Last has already pissed away a million dollars on it.” That estimate was probably exaggerated by a factor of two, and divisions other than Last’s were spending money on integrated circuits—Bay himself had approved advertising, sample giveaways, and trade-show marketing for the integrated circuit—but Last says that no one came to his defense. Noyce, who was running the meeting, remained silent. His unwritten policy was to allow people to thrash out their disagreements without his intervening. An employee once broke his wrist banging on a table to make a point during one of Noyce’s staff meetings.74

  Last stood up, announced he was taking a leave of absence—effective immediately—and walked out. He got as far from Fairchild as he easily could, traveling across the country to deliver a series of talks at his alma mater, MIT.

  At this point, Arthur Rock, the young banker who had helped with the initial Fairchild funding, re-entered the life of the company. Rock had recently left New York and Hayden, Stone, drawn to California by the “energetic scientists forming around Stanford.” Rock and a partner, Tom Davis, a former vice president of the Kern County Land Company who oversaw Kern’s investments, had just started a private investment company and raised a $5 million fund with plans to back new electronics companies. Almost all of this money came from Rock’s friends in the East Coast financial community, but he also invited the eight Fairchild founders to invest. Six agreed to do so immediately, but Noyce and Gordon Moore felt their positions required them to ask Camera and Instrument’s permission before investing. Camera and Instrument immediately forbade them from participating, citing potential conflict of interest. No one noted it at the time, but Davis and Rock were launching the first venture capital fund on the West Coast.75

  Rock had become a close friend to Jay Last and so knew of Last’s longstanding dissatisfactions at Fairchild. Rock had also made the acquaintance of Henry Singleton, a PhD engineer who had left his research job at Litton Industries two years before to start a high-tech conglomerate he called Teledyne. Singleton wanted to start a division of Teledyne to develop advanced semiconductor devices for military applications. In other words, he wanted to start an integrated circuits company.

  Rock, who would become one of the nation’s top corporate matchmakers, had spent the past few months urging Last, who wanted to oversee an operation devoted entirely to integrated circuits, to call Singleton. Rock had also asked Singleton to contact Last, but neither man wanted to make the first call. Finally, on the very day of the Fairchild Christmas gift exchange—somehow Last, who hated this sort of thing, had been roped into playing Santa—Rock called to say that he had Singleton waiting by the phone, and Last had better call him right now.

  After a short conversation, Last and Singleton agreed to meet in person. Jay Last said he would bring Jean Hoerni to the meeting. Hoerni, who saw no hope for advancement at Fairchild with the R&D director’s slot now filled by Moore, was also eager to try something new. He and Last drove to southern California—Hoerni hated to fly—and on the last day of 1960, they donned their best suits (“negotiating suits,” they called them) and met with Singleton and his co-founder George Kozmetsky at the Teledyne office in West Los Angeles. The meeting lasted several hours and ended well enough that Last and Hoerni left feeling fairly certain they would go to work for Singleton. Restless with adrenalin, they drove far out into the East Mojave, where they had planned a New Year’s Day climb in the Old Woman Mountains. They arrived at their campsite just before the calendar flipped to 1961 and stood together under the star-pricked sky, shouting and honking the car horn to welcome the new year and its new opportunities.76

  Back in Mountain View a few days later, Last had second thoughts. He felt a deeper connection to Fairchild than he had known. He called Noyce to ask him about the future of integrated circuits at the company. Noyce put him off. He still had not found a manager for the new diode plant opening in less than a month. Moreover, the general manager of SGS was in town for the next few days. Noyce hoped a relationship with the Italian executive would give him some leverage at the joint venture. Could Last wait until later in the week to talk?

  The question gave Last the answer he needed. In February 1961, he, Hoerni, and Sheldon Roberts (who long ago had told Last and Hoerni that he would leave Fairchild if the right opportunity arose) resigned from Fairchild to start the Amelco division of Teledyne.

  The departure of three founders opened the floodgates at Fairchild. In the next few months, a key researcher left to join Amelco. The integrated circuits development team departed en masse to Sunnyvale, where they, along with personnel expert Jack Yelverton, started an all-integrated-circuits company called Signetics with $1 million in start-up funds from three investment banks (White Weld, Lehman Brothers, and Goldman Sachs). Another integrated circuits company, Molectro, spun out, as did an operation, started by Gordon Moore’s assistant, to build furnaces for semiconductor firms. Moore, with the dry wit he sh
owed only to his closest associates, told Noyce, “If the personnel expansion rate of the last few weeks is maintained, instead of achieving a 60 percent expansion, we are likely to end up with about 60 percent of the personnel.”77

  Eugene Kleiner’s departure, in January 1962, left only half the founding team at Fairchild Semiconductor. The group of eight worked together as a group only one more time, and then not by choice. On Christmas Eve 1963, every founder received a registered letter at his home. The IRS had billed each of them nearly $250,000 in back taxes and penalties related to the acquisition of Fairchild Semiconductor by Camera and Instrument. Together the group of eight hired a top-notch legal team to contest the charges. (“I’m glad these fees are deductible,” said Last. “I’m glad they are divisible,” Moore shot back.) The case was settled out of court to the founders’ satisfaction, but not before giving them all a good scare.78

  The acquisition had caught the attention of the IRS because it seemed somehow wrong in 1959 for a young scientist to make more than a quarter-million dollars in a year and a half. This prospect, however, struck the men who left Fairchild Semiconductor in 1961 as not only fair but also potentially reproducible. If they started their own companies, they could run them and hold the majority of stock. Contrast this with their situation at Fairchild, where they had no managerial control, no prospect of moving into senior management, and no significant stake in the company, even though some of these men had been instrumental in its success. To stay on as an employee or to start something new as a founder was a choice between a twice-monthly paycheck and a potential fortune, between what one person who left called “just a job” and “a taste of blood.”79