The Man Behind the Microchip

by Leslie Berlin

  Noyce spent a good part of the end of July working with Rock to identify potential investors. Intel was too risky for banks, pension funds, and insurance companies who had to abide by the “prudent man” rule of investing, and so Noyce and Rock targeted individuals. Noyce personally received several unsolicited offers of “financial assistance” equivalent to hundreds of thousands of dollars, but he, Rock, and Moore decided to limit the investor pool to people they knew well. They invited each of the other six Fairchild founders, as well as Dick Hodgson and Arthur Rock’s old investment firm Hayden, Stone (now chaired by Bud Coyle), to invest up to $100,000 each. Two associates of Rock’s—Max Palevsky (the founder of Scientific Data Systems) and Rock’s business school classmate Fayez Sarofim—were included on the investors’ list, as was Paul Cook, the CEO of Raychem who had recently offered Noyce and Moore the opportunity to start their company as a subsidiary of the materials sciences firm. Noyce suggested also including the Rockefeller investment group, whose manager he knew through Grinnell contacts, and Gerard Currie, who had co-founded a successful young firm called Data Tech that used Fairchild circuits in boards sold to the military. Noyce was a director of Data Tech and an early investor in the company. Noyce also considered inviting Sherman Fairchild to invest, but Rock or Moore must have objected, because Fairchild did not make the short list.38

  Noyce also wanted Grinnell College, whose board of trustees he had chaired for the past two years, to have a chance to invest. A few years earlier, Joseph Rosenfield, a wealthy Des Moines financier who served on the Grinnell board, had told him that the college would want a stake in any new venture that Noyce started. When Rock began lining up investors for Intel, Rosenfield and another trustee, neither of whom knew technology, each wrote $100,000 checks on Grinnell’s behalf. The college endowment invested another $100,000. The trustees, who hoped the investment in Noyce’s company would add $10 million to the college’s endowment, planned to donate the stock to the college if Intel were successful but to eat the losses if the company failed. Even Warren Buffett, who joined the Grinnell board shortly before the decision to invest in Intel was made, was willing to abandon one of his fundamental rules of investing—only put money into things you understand—in this particular instance. As Buffett put it, “We were betting on the jockey, not the horse.”39

  In case an investor had questions, Rock had Noyce draw up a business plan, which Noyce did by typing up three pages, the meat of which said that Intel was going to try to build semiconductors “not of the types of integrated circuits now on the market” and that Intel “will seek to extend the technology to higher levels of integration.” The word “memory” never appeared in the plan. “Frankly,” said Noyce, “we didn’t want people to know what we were going to be doing. We thought it would attract too many competitors too soon.”40

  What the plan did or did not say was irrelevant in any case. Not a single investor asked to see it. “People had known Bob and were kind of lined up to invest in the company,” explains Rock. Fewer than 48 hours elapsed between the moment Rock began formally calling investors and the moment every dollar needed had been committed. While Rock concedes that raising $2.5 million in two days was an “exceptional” feat in 1968, he makes it clear that had the search been launched in an age of email and cell phones, the response would have been even swifter. “Back then,” he explains, “people had to return calls.”41

  Many more people wanted to invest than were given a chance to do so. Thanks to a news story that printed Noyce’s and Moore’s home addresses, Betty Moore received several calls at the house from people who wanted to put money into her husband’s company, even though the callers had no idea what it would do. For years after Intel’s founding, Noyce was approached by “keenly disappointed” friends wanting to know why they hadn’t been invited to invest. He told them Rock had been in charge of financing the launch.42

  Rock, never an emotionally expressive man, demonstrated his own excitement about Intel by augmenting his original $10,000 purchase by another $300,000. This made him the firm’s single largest investor after the founders. Years later, after he had funded dozens of successful companies (including Apple Computer) that together were valued at billions of dollars, Rock averred, “Intel is probably the only company I ever invested in that I was absolutely, 100 percent sure would be a success—because of Moore and Noyce.”43

  THE EASE WITH WHICH INTEL WAS LAUNCHED offers quite a contrast with the frustrating attempt, just 11 years earlier, to fund Fairchild Semiconductor. The founders’ track records can be credited for some of the difference. Noyce and Moore had built and run a company that grew into a $150 million enterprise, third biggest in its industry. Arthur Rock, too, had developed a reputation for picking winners. His imprimatur meant a great deal.

  As important, however, was timing. The American economy was booming—and the electronics industry was soaring. As the war in Vietnam escalated to once-unthinkable levels, the military spent record amounts on electronics. The industrial and consumer markets were also vibrant. Two months before Noyce and Moore launched Intel, failures of companies in the electronics industry dropped to record lows, with manufacturers failing on average only one-third as often as they had five years earlier. The industry was mature enough to have developed a network of suppliers, employees, lawyers, advertising firms, and other service providers familiar with semiconductors. At the same time, the microelectronics field was still new enough to support fresh entrants. It was, said one magazine, “the Age of Electro-Aquarius.”44

  This ideal moment would not last long, however. Business Week declared 1969 “the last year to start an integrated circuit operation … and make it big,” in light of increasing competition and rising capital costs. In that same year, Congress increased the maximum tax on long-term capital gains from 28 to 49 percent—a move aimed at getting a piece of the venture-capital pie for Uncle Sam and one that had the unintended effect of devastating the venture capital industry. By 1970, many people familiar with the industry believed the semiconductor market was fully saturated.45

  In the summer of 1968, however, life was so good for most electronics companies that firms up and down the Pacific Coast, flush with success, feeling generous, and hungry for employees, trained “young people from disadvantaged minorities” and “hard-core unemployed” as part of President Johnson’s War-on-Poverty programs. In Europe, student protests turned violent. In Chicago, police clubbed Vietnam War protesters at the Democratic National Convention, while in Washington, the United States Army, citing the “large number of civil disturbances” around the country, began construction on an “emergency action headquarters” anchored by a series of computers running on integrated circuit technology and designed to coordinate military action at as many as 25 domestic “hot spots” simultaneously. But in the electronics labs and fabs dotting the San Francisco Peninsula, it was a time of unbridled optimism.46

  Noyce and Moore were among the most optimistic. They knew that Intel would lose money until production got up to speed, but their business plan assumed only two years of losses. In effect, Noyce and Moore were saying that in 730 days their company would design a circuit, build a production line, produce in mass quantities a technology that had never before gone beyond lab prototyping, and then sell enough of those circuits to turn a profit. The plan has Noyce’s fingerprints all over it. Looking back 30 years later, Moore said the agenda was “more aggressive than I had ever planned on.”47

  IN EARLY AUGUST, Noyce left the exhilaration and furor of the Intel startup to spend two weeks with his family, who had passed almost the entire summer in Maine. Noyce joined them in time to close on a house they bought on the coast near Bremen, about an hour-and-a-half north of Portland. The house was modest, but the property was fantastic, nearly 30 pristine acres that ended in a promontory jutting into the Atlantic. From the porch of their new summer home, the Noyces could see only ocean, sky, and the small islands dotting Muscongus Bay.48

  The house w
as a concession to Betty, who had wanted Bob to start Intel in Boston. The region around Boston’s Route 128 had become a center of high-tech manufacturing when the Santa Clara Valley was still best known for plums. When Intel was started, more high-technology workers could be found in the minicomputer companies clustered along Route 128 than in the low-slung semiconductor buildings on the Peninsula. Noyce had briefly considered a move east at least seriously enough to ask Dick Hodgson, on whom he had come to rely as a mentor, what he thought of the idea. Hodgson immediately shot it down. “The whole concept of Intel is built around Fairchild, and your reputation and the people [on the Peninsula]. What’s Boston got to offer compared to this?”49

  And so, the Maine cottage. Noyce, of course, could not sit on the porch and watch the days meander by. He called back to California. He built a dock, bought several boats, made multiple trips into town for provisions and tools, met the neighbors, tried to swim in the ocean, and sailed with the family, the sun beating on his back and shoulders while Betty, who disliked getting wet, sat carefully in her life vest.

  Betty Noyce, at long last and at least for those two weeks on “her side” of the continent, was happy. She had wanted a home on the East Coast for so many years that even her mother-in-law, with whom Betty preferred to share nothing, knew of her dream. “So Betty’s home on the coast is a reality,” wrote Harriet, who thought such luxuries mildly sinful, in a letter to Bob. Harriet rather unfairly believed that Betty had made Bob overly interested in money, and she left unsaid what she clearly was thinking: now maybe she’ll give you some peace.50

  Shortly after Noyce returned from Maine, Intel began moving into the Union Carbide building, a 30,000-square-foot manufacturing facility at 365 Middlefield Road in Mountain View, not far from the original Fairchild Semiconductor headquarters. The building was a distinctly Spartan arrangement: even the offices of the “high muckety-mucks” (as Noyce laughingly referred to the management team) were rooms appointed with furniture cast off from the previous occupant. These offices, which lined the front of the building, were so small and crammed together that Noyce estimated that nobody was ever more than 20 steps away from another person—close enough to call meetings by hollering. It was quite a change from Noyce’s office at Fairchild, which Syosset management had insisted on having professionally decorated by an interior designer. The arrangement of the offices of Intel perfectly reflected the roles that the young company’s executives would play. Noyce sat nearest the front door. R&D director Grove sat closest to the middle of the building, near the research area. Moore’s office was between these two.51

  Intel’s first official communications center was a long table with three telephones (each with its own number) lined up across the top. The corporate “cafeteria” was a bare room with about six tables, a couple dozen chairs, a few vending machines, a bulletin board, and a sink with a rack behind it for coffee cups.52

  Union Carbide had not entirely moved out of the fab area in the back of the building that Intel intended to upgrade and use as its own manufacturing facility. Old Union Carbide equipment punctuated the cavernous space at odd angles and in random locations. Pipes poked from the ceiling. Wires drooped from the walls. The tile floor had more holes than tiles. Below the floor, things got even worse. The sewer pipe running out to the street had been completely eaten away by the acids Union Carbide had washed down the sinks.53

  It was either a disaster or an opportunity. Gene Flath decided the disarray gave him carte blanche to choose precisely where and how to build the fab that Intel needed. He began drawing up a plan. When Wescon, the annual trade show for which almost a decade earlier Noyce had asked Last to demonstrate a rudimentary integrated circuit, opened in Los Angeles, Flath attended, “with the equivalent of a checkbook” in hand. Flath went straight to the suppliers and equipment manufacturers’ booths at Wescon and proceeded to order equipment for the Intel fab right off the show floor: “I’ll have one of those, two of those, three of these.” A bank of furnaces, an evaporator, a lithography machine. The astonished salesmen, far more used to wooing and haggling that could stretch weeks, were writing makeshift purchase orders on the backs of random sheets of paper.54

  When it came time for the first Intel Christmas party, the company was still small enough that all 30 employees and their spouses could fit into Gordon Moore’s living room, where they drank themselves silly on a sweet, fruity Boston Fish House Punch brewed with a potency only a chemist like Moore could devise. Gordon and Betty Moore rarely hosted parties, but Betty Noyce had already announced at an earlier and smaller gathering that she was never again planning to engage in the sort of business socializing that had been expected of her at Fairchild. Her announcement had struck many as rude, but Betty had always found superficial conversation—particularly with those who knew her only as “Bob’s wife”—stressful. In her own mind, she was simply exercising her prerogative, as the wife of the man who made the rules, not to put herself into that situation again.55

  Every one of the 30 employees who celebrated that December night was technically oriented with the exception of a Controller, a receptionist, and a senior secretary from Fairchild, Jean Jones, who agreed to help out after being promised that she would only need to work part-time. Several of the newest hires—Hoff from Stanford, Dick Bohn from Sylvania, Skip Fehr from Texas Instruments, and Bob O’Hare from ITT—had not personally known the Intel founders, but had jumped at the opportunity to work with them for the same reason Noyce and Moore had wanted to work for Shockley. “They were absolutely the gods of the industry,” explained one early employee, using language almost identical to Noyce’s claim that the first call from Shockley was like hearing from God. Minor deities though Noyce and Moore might be, they did not have the final word on hiring. Andy Grove had the power to veto any technical hire.56

  A move to Intel was not an immediately lucrative proposition. Intel engineers were brought over at an average salary of $1,000 per month, plus options on 1,000 shares. This package was competitive, but not spectacular—Noyce and Moore knew, in the words of one early employee, that Intel “really had, frankly, a lot of sex appeal.” This employee recalls being impressed with the team and the “technical challenge” they were pursuing: “These guys seemed to be holding the challenge in their hands and steering the future.”57

  IF NOYCE HAD WANTED Fairchild to be the anti-Shockley, then Intel was the anti-Fairchild. The decision to structure Intel as what Noyce called “a two-headed monster,” with power split evenly between him and Gordon Moore, was a direct result of Noyce’s having confronted his own managerial limitations at Fairchild. Intel’s board of directors, which included Rock, Noyce, Moore, Richard Hodgson, and senior executives from several computer companies who could bring a customer’s perspective to board deliberations, was an intentional response to Noyce’s experiences with the Fairchild trustees who had been unburdened with any knowledge of semiconductors.58

  The ghost of Fairchild cast long shadows over technical day-to-day operations at Intel, as well. Because the difficulties in transferring devices from development to manufacturing at Fairchild were legion, the Intel founders completely erased the distinctions between the two divisions. Where other companies had one line for developing prototypes and another for mass-production runs, Intel had a single line. Processes developed in R&D were tested and fine-tuned with the same equipment used for commercial manufacturing. By “co-locating” development and manufacturing in this way, Moore and Noyce hoped to ensure that Intel did not waste its time developing high-concept devices that could not be built. Noyce and Moore further ensured the integration of R&D and manufacturing when they gave overall responsibility for production to Andy Grove, who had a theoretical and research background.

  Noyce and Moore decided to orient Intel 180-degrees away from Fairchild in product mix. Where Fairchild built huge volumes of relatively easy-to-manufacture, low-margin chips and engaged in the price cutting and trench warfare that this approach necessitated, Intel
would be tightly focused on being first to market with state-of-the-art devices that could command high prices and generate high profits that could be plowed back into R&D. By the time other firms entered the market and drove prices down, Intel would pull from its quiver a next-generation device that would once again command high prices until other companies could join the fray and the cycle would begin again.59

  Secrecy was essential for Noyce and Moore’s plan to work. Intel scientists did not give talks that could benefit competitors. They did not publish technical papers. This, too, was in direct opposition to the culture at Fairchild, and at most semiconductor companies and research labs, where PhDs openly discussed their findings at conferences or in print. The only currency in the Fairchild lab had been scientific prestige, but at Intel, the researchers had stock, which they knew would be worthless if they spread secrets around town.

  On a more philosophical level, too, Noyce wanted Intel to be the anti-Fairchild. Fairchild, he said, had come to be managed by “group think,” which killed innovation. He admitted that his personal drive for consensus was partially responsible for “group think,” which he defined this way: “any decision to go along with a new product innovation had to pass through a narrow gate. A single negative vote could kill a project, and one positive vote was worth approximately zero.” He wondered aloud if he would do better to champion a policy in which “a single ‘yes’ could initiate action.”60