The Man Behind the Microchip

by Leslie Berlin

  Hwoschinsky agreed to help Noyce determine his assets. “By far the bulk of it was Intel stock,” Hwoschinsky recalls. “But there were also the properties in California and Maine and the Noyce Air Force [Bob’s planes].” When the settlement agreement was signed, Bob got the house in California, Betty got the Maine estate, and the Intel stock was split down the middle, with Bob and Betty each receiving roughly a half-million shares worth nearly $25 million.42

  At some point while Hwoschinsky was doing his calculations, Noyce rather sheepishly handed him a shoe box. “What’s this?” Hwoschinsky asked.43

  “Just some little companies.”

  Hwoschinsky opened the box. It was filled nearly to the lid with IOUs and legal paperwork granting Noyce shares in various young companies in which he had invested. “Noyce could not just let his money sit,” Hwoschinsky explains. “He wanted to use it to stir up new adventures. He was in all sorts of little deals with people whom he had met socially. He just loved doing projects, loved people.”

  If an entrepreneur asked Noyce for money for a project that struck him as interesting, practical, and not technically impossible, he would usually agree to help. “He was a very generous person,” explains Intel’s Les Vadasz. “He did not have money worries, and … even if his heart wasn’t in [an investment], he probably felt, ‘Well, so what? There’s some upside.’ He didn’t take it so seriously.”44

  The type of private investing in which Noyce engaged is today called “angel investing” in high-technology circles. It took root in the United Sates in the 1930s, when wealthy benefactors such as Laurence Rockefeller helped their protégés to start new companies. Angel investing appeared in Silicon Valley around the time William Shockley started his company. By the early 1960s, for example, a dozen Hewlett-Packard executives set up a business-based “drinking club.” Each man promised to keep his door and ears open to electronics entrepreneurs and paid monthly “dues” of about $100. At the monthly “meeting,” held at one or another fellow’s home, the group would share a few drinks, and anyone who had heard of an exciting investment opportunity would talk about it. The group would then decide how much, if any, of their pooled money to invest in the company. After a half-dozen years, this little drinking group—which adopted the name “Page Mill Partners”—split roughly $30 million in profits.45

  Such returns were far beyond those of Noyce’s shoe box startups, Hwoschinsky soon learned. Paperwork from Advanced Micro Devices (AMD)—a successful company started by Fairchild salesman Jerry Sanders in which Noyce had invested a small amount—was stuffed in the box, but otherwise, Hwoschinsky had never heard of most of these operations, whose names are at this point lost to history. He had to determine the value of Noyce’s holdings and how much they had appreciated, based on some of the sloppiest record keeping he had ever seen.

  He grabbed a random sheet of paper from the box. “Noyce, how much did you pay for this stock?”

  “$16.95 a share.”

  “How do you remember? Do you have a canceled check?”

  “Paul, it was $16.95.”

  Hwoschinsky tracked down every entrepreneur whose name appeared in the box. When he met with the entrepreneur, Hwoschinsky also double-checked Noyce’s recollection of the price he paid for the stock against the founder’s paper work. Always, Noyce’s numbers were right. “He was a bloody elephant,” says Hwoschinsky. “I’ve never seen anything like it. He could remember anything and everything.”

  “Every now and again,” Hwoschinsky reminisces, “I’d get another call: ‘Uh, Paul? I found another shoe box.’”

  THE NOYCES OFFICIALLY BEGAN LIVING ON OPPOSITE COASTS in September. “Da is a little sad, but he seems resolved,” Penny wrote to her mother, an apt description of Noyce’s emotional state. At a dinner with his parents, brothers, and a few other relatives shortly after the decision to split from Betty, he started talking in a quiet, vulnerable tone no one in the family had heard before. His own parents and grandparents would never have gotten themselves into the situation in which he found himself, Noyce said. He was concerned about the effects of this marriage and its dissolution on his children. Already his youngest daughter was forced to choose which parent would attend parents’ weekend at her school. “Nothing else I’ve done matters,” he said, “because I’ve failed as a parent.” He had spent his life finding problems he could solve in innovative ways. This one he could not solve—he had helped to create it, in fact—and he worried that his children would pay the price for his failure.46

  Adding to his woes was the situation at Intel, which had changed dramatically. Over the course of three months, the semiconductor industry had fallen apart. Some portion of the excellent results Intel and its competitors had posted for the first half of the year could be traced to customers’ stockpiling semiconductor memories in the wake of concerns about the energy crisis. But by the middle of the third quarter, with the world economy beginning to spiral into recession and demand for electronics dropping rapidly, these customers had no need to buy more semiconductors.

  IN JULY 1974, the share price of Intel stock fell 30 percent—from 63½ to 44½—overnight. When Noyce arrived at work on the day after the precipitous drop, he learned that several Intel employees who had exercised their options when the price was in the sixties had not been able to sell their stock before the price drop due to inefficiencies in the option-exercising process. This struck Noyce as profoundly unfair. The employees would be short by one-third not because they timed the market wrong but because the elaborate stock-option protocol required by the SEC had delayed their sale. He wondered if it was possible to change the system.47

  Converting stock options to cash in the mid-1970s was a drawn-out, tedious process. At Intel, for example, employees first went to the on-site options desk to submit a form and a check for an amount equivalent to their strike price times the number of shares they were selling. An employee exercising an option to sell 1,000 shares with a $5 strike price, for example, would hand in a $5,000 check. Intel employees who did this in early July 1974, when the stock was at 63½, expected to receive $63,500 from the sale (1,000 shares at $63.50), thereby clearing a profit of $58,500 ($63,500, less the $5,000 paid to exercise the options).

  But submitting the form and check did not immediately sell the stock. The sale would be completed only when the employee signed the back of a paper stock certificate. These certificates were held by Intel’s transfer agent, Wells Fargo Bank in San Francisco, which did not release them until it had cashed the check. A week could easily pass between the time the employee exercised the options at Intel and the time he or she received the stock certificate and, by signing it, sold the stock. The sale was processed at the price on the day the certificate was signed, not at the price on the day the option was exercised.

  In general, the weeklong delay while paperwork traveled among Intel, Wells Fargo, and the employee was not problematic. But when the stock dropped 30 percent overnight, that delay would have cost the employee in the example above nearly $20,000. Of course, an employee could choose to exercise and hold the stock until the price recovered, but anyone who needed the money immediately was out of luck.

  Noyce soon was on the phone with his broker, Bob Harrington of Dean Witter. “We had five execs who exercised options in the last week but haven’t been able to sell the stock yet [because they had not received the certificates],” Noyce explained. “Two of these guys were planning on using the money for down payments on houses. Now the stock is down nearly one-third, and escrow is closing. Do you have any ideas about how we might help these guys?”

  Harrington could sympathize. To drum up business for his brokerage services, he had begun broadcasting stock reports every morning on a local Silicon Valley radio station. Many people had told him that they never missed his 7:57 “semiconductor report.” They sat in their cars, listening attentively and mentally calculating how much their stock was worth that day.

  Harrington was always surprised at how real the stock
seemed to his listeners with stock options. If a man held options on 50,000 shares in a company and Harrington reported that company trading at $20 per share, then by golly, the person considered himself a millionaire! Never mind that this fellow’s options might not be fully vested, or that he might not have the cash to pay the upfront costs of exercising shares, or that prices might drop and leave him with nothing. People thought of themselves as already possessing the money that was theirs only if everything went according to the most optimistic scenario. Although the Intel executives Noyce mentioned to Harrington were certainly more sophisticated investors than many of his listeners, Harrington knew that for them, too, it must have been a blow to find themselves 30 percent short. Harrington told Noyce that he would check with the legal department and get back to him.

  Not more than an hour later, Dean Witter’s securities attorney told Harrington that the brokerage house could do nothing about this situation, although Intel could offer loans to its employees to cover the shortfall.

  But Harrington had an idea: if there were some way to sell stock immediately upon exercising, no one would ever get caught in this bind again. If an employee exercising options could sign a note promising to turn over the paper stock certificate just as soon as it was in his or her physical possession, could that promissory note itself not serve as the temporary equivalent of the signed stock certificate? And why stop there? If a promissory note could stand in for a stock certificate, why couldn’t it stand in for a check? Why not simply add a line to the promissory note directing that the cost of exercising the options be deducted from the proceeds of the sale? This way an employee would not need to have large amounts of cash on hand in order to exercise options.

  In the scheme Harrington imagined, the employee in the above example would not submit the $5,000 check to Intel. He would not need to wait to receive the stock certificate before selling his stock. Instead, as soon as he signed a promissory note, his options would be exercised, his stock would be sold, and the $5,000 needed to exercise the options would be automatically deducted from his $63,500 payout. In other words, he would sign the promissory note and almost immediately receive the $58,500 coming to him.

  Harrington contacted a law firm, and together they secured an “assurance of no action” from the Securities and Exchange Commission. In effect, the SEC said that if a company wanted to assume the liability of accepting promissory notes in the place of checks or stock certificates, a firm could do so, as long as it provided notice of these actions in its annual report.

  Harrington told Noyce that he planned to start an “Option Exercise—Immediate Sale” program for corporate clients. Intel signed up for the service, and within a few months Intel’s finance vice president was reporting that “virtually all” of the Intel employees exercising options used the “1-day stock buy/sell turnaround.”48

  Harrington had hoped to keep the program proprietary, but the idea was too good. It spread rapidly through Dean Witter and then around the country. “That program made life easier for tens of thousands of people,” Harrington explains, “and the seed for it came from Bob Noyce asking a simple question: ‘Isn’t there a better way?’”49

  The same-day sales innovation was one of very few bright spots in Noyce’s life in the second half of 1974. While his marriage was suffering its death throes in Maine, semiconductor companies in California began laying off workers. In July alone, 50 people lost their jobs at Intersil, Signetics laid off 100 workers, and AMI cut 230 jobs. In an attempt to avoid layoffs at Intel, Moore and Grove—with Noyce’s assent—decided to close two production facilities for a week before both July 4th and Labor Day, temporarily furloughing 600 employees, who were granted only half their normal pay.50

  Such efforts could not save Intel’s bottom line. The memory market had essentially disappeared in the recession. The Microma watch business was well on its way to disaster and would post losses of more than 1.5 million pretax dollars in 1974—nearly double the hit projected by the executive team at the beginning of the year. Intel’s 4K RAM was so late that the company began second-sourcing a competitor’s product. The only bright spots were the memory systems and microprocessor businesses, whose sales had increased both quarters, but these were too small to have a significant effect on overall results. Intel’s total sales dropped 5 percent in the third quarter—earnings slipped 48 percent—and would fall another 9 percent in the fourth. Although the company did better than many of its competitors—AMD’s $2.4 million net income in 1974 evaporated into a $2.5 million loss in 1975; Signetics went from $10 million profits to $4 million losses; and Mostek fell from $4 million profits to $1.2 million losses—a 48 percent fall in profits was a devastating performance for a company accustomed to tripling its revenues annually.51

  In October, Noyce signed his name to one of the most difficult letters of his career. “To all Intel Employees,” he wrote. “Continued reduced business conditions have made it necessary for us to have a reduction in our work force today. … We have tried to avoid such reductions in force by instituting transfers of people within the Company into areas which have not been seriously affected by the economy. We hoped that our overhead reduction in August and periodic reduced work weeks would be sufficient to cope with the slackening product demand until the economy recovered. This was based upon the hope that business would be improving by now. Unfortunately, it is not.”52

  With this letter, signed two days before his legal separation from Betty, Noyce terminated roughly 30 percent of Intel’s 2,500 employees, most of them in production. He found the move profoundly upsetting and miserably confided in a friend, “For a few goddamned points on Wall Street, we have to ruin peoples’ lives.”53

  His letter to employees promised, “When business does improve, our first action will be to recall employees affected by this cut,” but he did not anticipate any changes for at least six months. He told a meeting of the New York Society of Security Analysts that Intel was bracing for more layoffs as well as “tremendous price attrition.” When Thanksgiving came, Intel again closed part of its operations—and did not pay workers—during the long weekend. Hewlett-Packard, Fairchild, National, Signetics, and AMI did the same, some of them also furloughing employees for the week between Christmas and New Year’s.54

  By the end of 1974, nearly 20 percent of the semiconductor industry’s blue-collar workers in Silicon Valley had been laid off. The industry’s job losses and involuntary and unpaid sabbaticals sparked a protest by about 125 production workers and supportive college students outside the annual Western Electronics Manufacturing Association (WEMA) meeting at which Noyce received the 1974 WEMA medal of achievement—granted “to those who have made most significant contributions to the advancement of electronics.” The citation called Noyce “a dominant force in the development of the semiconductor industry on the San Francisco Peninsula,” but the protesters outside the meeting had other thoughts about him. Among the signs reading “No Short Work Week!” “Fairchild Workers Unite!” and “70,000 Electronics Workers Say No Vacations Without Pay!” was one wishing “Indigestion to Noyce from Intel Workers.” This was the last thing he needed at the end of the horrible fourth quarter of 1974.55

  SHORTLY AFTER DECIDING THE LAYOFFS WERE NECESSARY, Noyce told Moore that he wanted to leave his position as Intel’s president. Arthur Rock conjectures that the layoffs were the final stress that led Noyce to leave daily management at Intel, but Moore had seen signs of Noyce’s desire for a change even during the glorious first and second quarters of the year. Early in 1974, Noyce told Moore that he had suggested to Charlie Sporck that they merge National and Intel. “It was not unusual for him to talk to people about things without consulting me first,” Moore later explained, without apparent rancor. Moore agreed to meet with Sporck but ultimately decided that he would “like to try running Intel for a while” rather than merging forces with National.56

  Moore and Noyce agreed that if Noyce left, Andy Grove should be promoted so that he and Moor
e could run Intel as a team, continuing the tradition set by the founders. Grove had been pushing for more responsibility for several years. Explained Intel director Richard Hodgson, “Andy Grove wanted more and more and more—in a good way. You just couldn’t contain him.” Gordon Moore, who liked to say that “Andy had gotten over his PhD,” was certain that if Noyce moved to board chair and Moore stepped in as president and CEO, Grove would make an excellent executive vice president. Already he had moved beyond simply implementing the founders’ plans to refining, supplementing, and even convincing Moore and Noyce to change them. Grove had arrived at Intel a scientist, rapidly transformed himself into an operations manager, and would no doubt be capable in a general management role.57

  In fact, the attention to hard data and formal processes that one associates with Grove had been ascendant in the company for at least a year, ever since the success of the 1103 had made Intel’s success dependent upon the ability to manufacture this device in quantity. Beginning in 1972, agendas for the weekly staff meetings, which were once open-ended, ran on a five-week cycle that addressed a different set of topics each week—“1. Bookings, shipments, organization, personnel; 2. Prior months’ performance (financial, unit sales and revenue, production to finished goods, business indicators); 3. Key customers [the only subject Noyce underlined] and new activities; 4. Pricing review/Cost reports, new and old product scheduling and pricing; 5. General.”58

  As early as 1972, says Gelbach, “Bob gave advice but Andy and Gordon ran it.” Another employee explains that although from the outside it appeared “Bob was running the company,” it only took a few weeks at Intel to “discover that Andy was running the company.”59