Era ends as Gordon Moore quits Intel's board

Era ends as Gordon Moore quits Intel's board
MANHASSET, N.Y. — Technically sound, no management potential. That assessment of a man who would go on to lead one of the most powerful corporations in the United States ranks right up there with a Paramount executive's first take on Fred Astaire ("Can't act. Can't sing. Balding. Can dance a little") as among the most spectacular misevaluations in the history of human resources.

The man in question is Gordon Moore, who at the end of May quietly closed the final chapter in an extraordinary career that is inextricably linked to the history of the semiconductor industry itself. Hewing to Intel Corp.'s mandatory retirement age of 72 — a policy he imposed himself about 20 years ago to assist in succession planning for the board of directors — Moore retired from the board of the company he co-founded in 1968. The inventor of Moore's Law and one of the last true giants of the electronics industry can still "dance a little." He will continue to work with the Santa Clara, Calif., chip maker as a consultant to the board, spending at least one day a week in the office. But he is no longer a voting member.

After earning a PhD in chemistry from the California Institute of Technology in 1954, Moore underwent a psychological evaluation while interviewing for a position with Dow Chemical. "The psychologist told me I was OK technically, but that I'd never manage anything," said Moore.

That prediction, of course, proved ironic in the extreme. Moore went on to co-found Fairchild Semiconductor Corp. just three years later, in 1957, and then Intel Corp. with Fairchild colleague Robert Noyce, 11 years after that. Moore has served as president, chief executive officer, chairman and chairman emeritus of the chip giant, which has just estimated second-quarter revenue of between $6.2 billion and $6.8 billion.

Often characterizing himself as an "accidental entrepreneur," the self-effacing and modest Moore first achieved fame in 1965 while working at Fairchild, with his prediction that ICs would double in complexity every year, pulling potential computing power along with them. When he became Intel's CEO in 1975 he revised this rule to every two years. The industry — and not Moore — chose to split the difference to every 18 months, and that is where Moore's Law has remained ever since.

But fame was the furthest thing from Moore's mind while growing up in the relaxed atmosphere of Northern California. Born in Pescadero, just south of San Francisco, on Jan. 3, 1929, Gordon E. Moore saw his future at the age of 11 when he caught sight of a neighbor's chemistry set. "When I saw all the neat things you could do with it, that's when I decided I wanted to be a chemist," said Moore. "Though I didn't quite know what that meant."

The "neat things" he envisioned were not for the faint of heart. Moore made explosives and rockets — fragments of which would pepper a neighbor's roof, "which wasn't much fun from their point of view." He tinkered with a variety of explosives, "but nitroglycerine was one of the easiest ones," he said. "You couldn't get away with that now."

The interest in chemistry would stay with him throughout high school, which was lucky, since athletics would not be his path to college. "In school I was always out for some kind of sport or other — but I was mediocre in everything. I think I got letters in four different sports in high school," he laughed. "I wasn't really very good at any."

Moore enrolled at San Jose State University, where, just before his sophomore year, chemistry took on a whole new meaning when he met his future wife, Betty. Moore went out of his way to get acquainted with her, and they started dating soon after. When Moore transferred to the University of California at Berkeley a year later, Betty moved too, living with her uncle during Moore's final year there.

Between completion of Moore's undergraduate degree (a BS in chemistry) and his first term at Cal Tech, they married. "We actually got married on a Saturday and I had to start taking exams in Cal Tech on Monday, so the honeymoon was driving from Northern California to Pasadena," said Moore.

After Cal Tech, Moore used his PhD to do basic research at Johns Hopkins University's Applied Physics Laboratory in Silver Spring, Md. But that didn't last long. "I found myself doing research and calculating the cost per word in the articles we were publishing, and wondering if the government was getting its money's worth. That's when I decided to get closer to something practical," he said.

As Moore tells it, he then "just sort of lucked out" as he began job hunting and bumped into William Shockley, the Bell Labs genius. Hot off having invented the transistor, Shockley wanted to set up his own company in Northern California. He needed a chemist to help him make a cheap silicon transistor and recruited Moore.

"The skills I had as an experimental physical chemist worked pretty well within the kinds of things that had to be done to set up the processing for silicon, so I just kind of fell into it, I guess," Moore said.

Moore also longed to get back to California, and it sounded like a perfect match. But as Moore would tell it, Shockley's erratic management style doomed the Shockley Labs group from the start. "I suppose maybe I should have been suspicious when none of the people who had worked with [Shockley] at Bell Labs joined his new venture," said Moore.

After realizing they couldn't get around Shockley, who by then had nabbed a Nobel Prize, Moore and seven co-workers — including Bob Noyce, with whom he would go on to found Intel — left the company. Moore pegs this event as the start of his "accidental" entrepreneurial career.

The eight wanted to continue to work together and to stay in Northern California, so they decided to set up their own company with the help of Arthur Rock, a young MBA out of Harvard. With an investment of $1.3 million from Fairchild Camera and Instrument and $500 from each of the eight, Fairchild Semiconductor Corp. was born in 1957.

Fairchild was where Moore and Noyce learned much of what would sustain them through their Intel years about how a large, successful manufacturing company should be run.

"We learned from the Shockley experience that none of us knew how to run a company, so the first thing we did [at Fairchild] was hire our own boss," he said. But when that boss — Ed Baldwin — ran off with the company secrets (which he later had to return), Noyce stepped into the general manager's position and Moore sidestepped into Noyce's slot as director of research and development.

Fairchild had one big trick up its sleeve: planar transistors. The concept was defined by Jean Hoerni, another Cal Tech post-doc who also had worked with Shockley. Hoerni came up with the idea of leaving the silicon oxide in place, after it formed, in the spots where the silicon junctions came to the surface. Normally this area would have been kept clean — at great expense.

Noyce added to the planar transistor idea with two key inventions that would eventually lead to the first practical integrated circuit. The first involved running interconnections as metal films over the top of the planar devices; the other put structures inside the silicon to isolate one transistor from the other. The IC industry was born.

Working in parallel over at Texas Instruments Inc. (TI), Jack Kilby was busy with his version of what an IC should look like. "TI was the established silicon transistor company at the time, insofar as there was one" in the early 1960s, said Moore. "We were taking a different technological approach that we were absolutely convinced would be far superior."

Fairchild was making, in effect, grown-junction transistors, whereas "Kilby essentially built a laboratory model showing you could make all the circuit elements out of germanium — but it was a long way from a practical device. TI just didn't have the right technology. What they had couldn't do the batch processing the way planar technology did."

Moore said he was never clear whether Noyce was aware of Kilby's work or not. Regardless, "Both their contributions were important — but significantly different," said Moore. Kilby agrees about the importance of Noyce's role, saying that "Noyce made a significant contribution. As a result of the planar process, ICs became a great deal more manufacturable and practical."

Moore soon got the job of director of the laboratory. "I guess I had to oversee the efforts to actually make the first practical IC," he said modestly. "Well, it didn't seem that different. It was just more of the kind of chemistry and stuff that I had been doing. We wanted to make a transistor and we had some naive idea that if we made a good transistor people would buy it."

Theirs was a strongly technology-driven approach to the business. "Nothing really appeared as an especially major challenge at that time," Moore recalled. "Run-of-the-mill engineering while continuing to advance things. A lot of fun."

Passed over

However, it wasn't long before management problems ended the fun. After going through two CEOs in six months, and then trying to run the company with a three-man committee of board members "who didn't know anything about the semiconductor business," the final straw came when Noyce was obviously going to be bypassed as CEO — despite being the obvious choice.

"He [Noyce] was a little unhappy about that. I could see that the company was going to change dramatically if they brought someone in from outside, and I thought I'd rather leave before it changed than after," said Moore.

So, together, Moore and Noyce decided that the only logical thing to do was to resign and set up another company. But to do what? Whatever the business plan was, it had to minimize the advantage of established companies such as TI and Fairchild.

"We thought we saw an opportunity in making much more complex ICs than anyone was making commercially," said Moore. "Semiconductor memories were the first bunch of circuits at that time." They were attractive because the startup could make a product of almost arbitrary complexity for use in almost all digital systems, riding on the clever processing of silicon, not low-cost assembly.

This was the founding premise of the company soon to be known as Intel Corp., set up on July 18, 1968. Chasing bipolar technology and multichip packaging options as well, Intel eventually hit its stride with silicon-gate MOS technology.

"At that time silicon-gate MOS had only been demonstrated on individual transistors and it certainly wasn't clear that that was going to be the way the industry went," Moore said. "But it had some attractive features, so we went after that."

The decision proved fortuitous: Not only did the technology work, it was also just hard enough to engineer to give Intel the head start a young startup needs in the face of established competition. "They [the industry] didn't pay much attention for a while and took a long time to turn around and have a competitive technology," said Moore. As a result, Intel had six or so years to itself. "That really let us get established," he said.

An early foray into electronic calculators led to the development of a central processing unit built on a single MOS chip. The result was the first microprocessor — the 4004 — which emerged from the labs in 1971. Though the calculator market never took off for Intel, for various reasons, the company's forward path had just been defined. And it never looked back.

During those years, Andrew Grove, whom Moore had hired straight out of grad school, was Moore's right-hand man. The first on the payroll, Grove also has the distinction of being the first to get an Intel paycheck, even before Moore and Noyce drew a salary. "Noyce and I weren't going to take any salary until the first of August, but Andy arrived before that and he wasn't in a position to live without salary at that time."

The two hit it off from the start. "He's very very bright, very direct, results-oriented. A terrific guy to get things done," Moore said of Grove, who is now Intel's chairman. "We quickly got to the stage where we could communicate amazingly efficiently — to the point where we were completing each other's sentences. Although he probably completed more of mine than I completed of his."

'Shades of gray'

In a lot of respects the two were complementary. "He sees things in black and white, I see them in delicate shades of gray," said Moore. "He was a good amplifier for me. I'm not a dynamic manager in the same sense Andy is. But I think we worked well together."

Moore's relationship with Noyce was complementary in another direction. "He was gregarious, the type of guy everyone liked the first time they met," Moore said. Noyce was also an idea man par excellence. "In fact," said Moore, "he had a frustrating number of good ideas. You'd be doing something and he'd come and suggest things and you really couldn't not consider his idea. He never pushed it, but he just threw out these pearls and figured people would just pick up enough of them and run with them — and that's generally what happened."

As a team, the trio of Moore, Noyce and Grove formed the core ideology that would help propel Intel into the technological juggernaut it has become. Though Noyce died in 1990, Moore and Grove continue the legacy. From the first bipolar memory product in August of 1969 — the first MOS product the following month, the first microprocessor in 1971 and right through to the latest, Itanium processor with 240 million transistors — Intel has never let up.

At the recent Intel analysts' meeting in New York, Grove reflected on the current semiconductor slump and quoted two lesser-known "laws" Moore articulated during the 1981 downturn (occasioned by the minicomputer market crash and ended by the arrival of the PC). The first, a quote from As You Like It, is "Sweet are the uses of adversity." The second: "It won't remain like this forever — optimize with R&D investment and in aggressive new products."

Those are Moore's — and Intel's — mottoes even now. The company has invested $4.2 billion in R&D this year alone. Moore may have retired, but his legacy at Intel just won't quit.

Gordon Moore sat down with EE Times to chat about his past, his present and what he thinks the future will bring. Click here to read the Q&A.