RW: Wilf Corrigan is chairman and CEO of LSI Logic Corporation, a leading system on a chip supplier. Wilf has been in the semiconductor industry since 1960 where at age twenty-two, he made seminal advances in epitaxy at Motorola. He was chairman and CEO of Fairchild following the Les Hogan regime. In 1980, he and three others co-founded LSI Logic. I visited Wilf at Villa Isabel, his Los Altos Hills estate.
RW: Have you been active in, and here's it's 1998, have you been active in the semiconductor business now just about longer than anybody else? Gordon Moore's retired.
WC: Well I don't know about that, I mean I think there's plenty of other people still involved, maybe not as directly in the business as I am, but they're still active in the business. You know, a lot of the old semiconductor guys went into the venture capital business. I think you'll find many of the smaller companies, the specialty niche companies, people are involved either on the board or so on. And then of course there's still a lot of the old guys still running companies. You know, they, you know, Bob Swanson at Linear Technology was with me, at Transitron in 1960. Ray Stata at Analog Devices is still active and is now active in the semiconductor industry association as well. So I think there's still a few guys around. Of course there's Jerry Sanders with AMD, Andy Grove at Intel is still obviously very much involved. So, but I've been in the business since 1960, well the ranks are getting a little thinner.
RW: Yeah, well the integrated circuit was invented in '58, so you've been there since the beginning. So tell us about your family life in England and what sort of a family did you grow up in?
WC: Well, you know, my father worked on the docks in Liverpool England. And of course I was born in '38 so he was off at the war from '39 on, you know, he volunteered for the navy as soon as war was declared and I didn't see my father till 1946 since he was involved just about every theatre of the war. My early childhood was mainly in the bombing in 1940 and '41 and so on, so it was kind of a wartime experience. And then I went to grammar school in England and focused on science and then when I went to Imperial College of Science, I originally was going to take chemistry, but that quickly changed to chemical engineering. And when I graduated in 1960, the, at the time I had three job opportunities. I knew I was going to leave England, so I had one in a gunpowder factory in Australia, another job offer was in Germany and the third one was a company called Transitron in Boston. And the job in Boston paid twice as much as the other two, so I went to Boston. And, I was curious to find out what it was that this company did when I got there. So, that's how I got into the semiconductor business.
RW Because it paid twice as much.
WC: Only reason.
RW: And it was out of England.
WC: And it was in America.
RW: Well it was a pretty good choice looking around. So what did you do at Transitron?
WC: Well, you know, the on the job training in those days was pretty rapid. And I got off the plane in Boston on a Sunday night and a friend of mine that also worked for Transitron, had actually got me the job, he picked me up at the airport and as he pulled the car away from the curb, he said you know, I can't be taking you around all the place, you've got to get your own wheels. You know, I'm just one hour into the U.S. and so we lived in a boardinghouse just down the street from the factory, which was in Wakefield Mass. and there's a lot of young immigrant, mainly European immigrant engineers at that time and so we all kind of lived almost like, I wouldn't say it was like a fraternity house, but it was, each of us rented a room and so on. In fact George Wells, who you know, he was the only married person there. So he was married to Rona and she was kind of like the den mother. And then we'd walk to work in the morning. So I got in that Sunday night, first thing Monday morning I checked into their version of personnel office, which was very minimal, and by nine o'clock that Monday morning I was issued with a desk, and a seat, and so on, and I was told that I was now the production engineer for the 2N338 transistor line. And so being logical I said do we have a library here. So I went downstairs, found the library, checked out Transistors 1 and I got the book and I was on about page three of the book when a big guy came bursting into the room and he said,"Are you the goddamn new production engineer?" And I said, yeah, I was twenty-two years old at this time, and he said "Well I've got two hundred women outside here on the line sitting on their asses with their hands folded, this lines shut down and you're sitting here reading a book? Get your ass out on the line and go get that line started again." So that was my introduction to the semiconductor business.
RW: And did you?
WC: Oh yeah, sure.
RW: So how long did you stay at Transitron?
WC: Well actually it was not that long because the way in which we entertained ourselves at that time was, you know, we were all bachelors. And we quickly found out that there were lots of people, things were starting to get very hot, so there was lots of interest in hiring people for the semiconductor business on Route 128 in Boston, that was kind of like the center of the semiconductor business. I mean there was Dallas, of course, and things were starting to happen in Silicon Valley around Fairchild, but they, Route 128 was kind of hot as a semiconductor location. So you could always get a free dinner by being interviewed. So almost every night we'd go out and be interviewed and then, and of course the weekends, when we're all bachelors, nothing much to do, didn't have that much surplus money, I was still paying, you know, for buying my car and all the other stuff, and so an ad came in the newspaper from Motorola in Phoenix and they were going to be in town interviewing and they wanted people that knew something about epitaxy, which was growing crystals. In the case of epitax you can grow crystals from the gas phase. And I said well gee I know something about that because I had done a project in my last year at university. So I said yeah I know about that, so I'll go and interview with these guys. And again it was just another, well we'll go get a, I'll be wined and dined by these guys from out of town. And it turned out that the guy that was interviewing was another English guy who was a very charismatic individual called George Russell and he said hey we're ahead of everybody in this particular technology, you know, you're a perfect fit and so on, and plus we'll give you thirty percent more money. So, to me, you know, I just arrived, you know, a few months, now suddenly I'm going from five-fifty-five hundred dollars a year to eighty-five hundred dollars a year, which was a big jump. And so I said okay, I'll do it. Now meanwhile my wife to be was back in Norway waiting to get her visa. So I figured, and her parents that have never met me. So I figured if I were to tell her that I was about to move jobs, she was in Norway, her parents one way or another were going to talk her out of coming over. So I couldn't really tell her that I was going to move jobs because I had only been there six months. Obviously her parents would say, who the hell is this guy that you're thinking of marrying? So basically she, I said hey you have to be here by this date because Motorola, I didn't have the extra money, so Motorola would not pay in those days for girlfriends to travel. So I said, hey you've got to be there by this day and so one way or another I persuaded her to come by this state, which was right at the end of December. And so we got married one day and then we flew to Phoenix Arizona the next day. So that's where I finished up in Phoenix.
RW: A little different in temperature and environment than Transitron.
WC: Well yeah, particularly for her because I mean she was coming from Trondheim, which is in the far north of Norway and suddenly she turned herself, just within days, she was sitting on the wrong side of town in Phoenix, Arizona. But fortunately it was at Christmas time of year, which was cooler in Arizona.
RW: So what did you do at Motorola?
WC: Well the, what he told me was actually correct, it was the very early days of epitaxy but at that time the... and it's always true I suppose, the one-eyed man is king. That most of the people in the semiconductor business tended to be more electronic engineers and are relatively few people with a chemical background, when actually the weakest link in the chain was the chemistry. So, I was actually able to very quickly get involved and then Gordon Morre has got the same sort of background as I have, chemist, and at that point in time, the people that were trying to do this technology really hadn't had the background in general so that, and it was at a point in time when everything was changing. So we were able to do a lot of the original work on epitaxy and actually get it into production both in germanium and in silicon in a very short period of time. And so, it gave us at Motorola a jump which was needed because at this point in time in transistors germanium, alloy transistors and germanium power transistors would have been the main product line for Motorola and they were trying to get into what we call small signal, which was germanium mesa and also get into mesa, what's called mesa silicon transistors. And we were starting to get these rumors that these people at Fairchild had got this planar process and of course nobody knew what the planar process meant and nobody could quite visualize what it was when all the rest of us were just working on doing the mesa. So we really had to do planar and so Motorola was looking for a difference and so we jumped on epitaxy as the difference. Fairchild was making triple diffused chips which where you're worried about the VCSAT, the saturation resistance or saturation voltage, and so you have to somehow reduce the series resistance and there's two ways you can do that, you can make the chip bigger or you could dope the back of the wafer with, you know, impurities or this new way of doing it which was you could start with a very low resistivity substrate and then deposit very high resistivity silicon on top in a very thin layer, five microns, maybe ten microns that for high voltage device maybe fifteen or twenty microns and nobody quite knew how to do this on a production basis, even though Bell Labs and IBM was demonstrating it. We figured out how to do it in high volume before anybody else did. And it was interesting to me that when I got to Fairchild in '68, which was probably seven years later, Fairchild had not fully addressed this epitaxy approach even then seven years later. So the, epitaxy was kind of an important thing to me, I mean I got a number of the basic patents on that. One of the people said now what was the high point of your existence in the semiconductor business and I'd say there was this kind of period of time in the early 60's when, you know, for a moment in time, you know, we were, there was a small group of us, maybe four of five of us, that we kind of knew that we knew more than anybody else about how to do this. Of course these advantages dissipate. And this was like twenty-four hours a day and a total involvement and when you're totally involved in this, you get so that you really believe that you can think like the silicon. We were able to be doing experiments, but they almost didn't feel like we needed to do the experiment because we knew what was going to happen. And you could make the adjustments to the temperature and the gas flows and just on your kind of intuition. And you felt like you knew exactly what the atoms were going to do as they came down on the surface and so on. So, it was a very exciting time and then through that, because we had the epitaxy process, we were then able to make transistors that were literally a quarter of the size of the competitor transistors, which, you know, as we've gone on in the industry, that's always been the most critical thing, how did you get that chip size down for the same functionality. So then in, I guess in '62, I was suddenly given the job of why don't you run the whole transistor line? And part of the reason for that was because Charlie Spork offered me a job to come work for him at Fairchild, this was in '62. And so I was going to go take this job and they said well, don't do that, why don't you come and run the transistor line because up till then I was the epitaxy guy...
RW: But you'd only been out of college two years.
WC: Oh yeah, I was like twenty-three, twenty-four at the time. So, and that was very typical and if you go back and look at the ages, you know, how old was Bob Noyce and how old was Gordon Moore and in fact how old was Les Hogan. I mean Les Hogan's first job he was a Harvard professor at thirty-eight was to go be General Manager of the Motorola semiconductor business. So that was very common at that time that most of the key players were very young, you know, in the early 60's. So then I got very much involved in silicon transistors and that was a great growth time for silicon transistors and then that's essentially what I did for that whole, for the next six years, was just growing the silicon transistor business at Motorola. And then in '68 Les Hogan was given the job as president and CEO of Fairchild Camera and so he invited seven of us to go with him to that, to California, so we did. And if you remember they called us Hogan's Heroes at the time, it was a very popular television series called Hogan's Heroes at that time. And so of course that - because his name was Hogan, this became Hogan's Heroes at Fairchild. And of course then we realized we were into a totally different environment, Phoenix in '68 and Silicon Valley in '68 were two totally and different environments. And it was quite an adjustment.
RW: Well, I recall that the Motorola people were quite shocked with Jerry Sanders, his flamboyance.
WC: Well that was an interesting phenomenon. Well as you know, Jay Sanders is a very close friend of mine and has been for twenty-five years, but the Motorola management sales force was almost kind of very mid-west in its culture and its attitudes and the sales force at Motorola, if you, was very much like an IBM sales force at the time and people were very correct and wore a certain sort of suit and a certain sort of tie and so on. Whereas the Fairchild salesman under Sanders, who was the sales manager, were kind of notorious within Motorola, which we'd all come from because we kind of moved as kind of like a family. And so they all had to be subjected to the same inputs and so on. And they called the Fairchild sales force at that time the "Suede Shoe Boys" because they were very flamboyant and a few of them had gold chains and open necked shirts and very Hollywood. And of course Jerry, you know, certainly at that time was very Hollywood. And so I think the, I mean I came, I was thirty at the time and so there was almost half a generation with the other, ex-Motorola managers, George Scalise was about the same age as I was, most of the other guys were at least ten years, maybe fifteen years, older. And pretty conservative. But I remember George Scalise and I and the rationale was hey we could have really changed the management here because we need to get some real sale guys and, so on and so forth. And after we'd been at Fairchild for about a month, suddenly we realized that this sales force was younger, smarter, quicker, more aggressive, and better and was much more oriented towards selling leading edge technology than the old Motorola sales force. And, but I think at that time Les Hogan had decided to bring in Joe Van Poppelen for that job and then of course there ended up being a parting ways. But by the time that happened, we had decided that the - George Scalise and I in particular were very vehement and hey you really got to keep Jerry Sanders in this job, he's been the best sales guy in the industry and so on. But anyway, Jerry then left and that's how AMD started.
RW: Well I recall that it was, that the Motorola people didn't go down to the Wagon Wheel with the rest of the people.
WC: I did all the time, every Friday night.
RW: Not quite as much. Well so what went on? You got promoted more at Fairchild as well.
WC: Well originally, because I was the ,kind of the transistor guy, I took
over the transistor operations at Fairchild and also the offshore manufacturing
and, which was the offshore plants, which I was very interested in that offshore
strategy which really was at the time we called it jet age automation because
in Hong Kong the people were being paid ten cents an hour when the going rate
was, for direct labor, was maybe two-fifty an hour in the States. And that was
a strategy that was implemented by Fairchild. Fairchild was the real innovator
in that and they'd done that in the early '60's and there was a positive and
negative to that. It was the seeds of the success of the US industry over the
next ten years and it was also the seeds of the problems that the industry got
into later because the Japanese decided to automate. But things were changing
so fast that by the time they'd automated something, they'd automated for the
last generation. This happened to the Europeans a little earlier with the "push
out base transistor" that was highly automated in germanium, and then germanium
ceased to be the driver. And then there was the MADT which was an automated
alloy transistor and by the time it was automated and it was wonderful but it
was obsolete. Well the Japanese went through a whole series of things like that
from the mid '60's to the mid '70's and because the Americans had adopted this
Fairchild - copied Fairchild and gone with this offshore cheap labor, you could
move with this very quickly with very modest capital investments because you're
just using peoples fingers to do a lot of the work, and so your tooling was
very low cost and so as the technology changed, you could tool something else
and move on. And later the disc drive industry did exactly the same thing.
RW: But didn't that push back the automation that maybe should and could have occurred by using foreign labor?
WC: Well there's different conclusions on that, I mean the, if you automate too early, you'll lose your flexibility and I think that's happened to the Japanese certainly at that time from I would guess from '65 to '75 and then round about '75 their approach with a - certainly the assembly process they had completely automated so they could make it in Japan and meanwhile the labor rates had increased in Hong Kong and Hong Kong was evolving to the next level for them and so was Singapore and you know, the American response was to move to Thailand, to move to wherever the lower cost labor was, you know, so that. I know at Fairchild we had plants in the Philippines, we had plants in Indonesia and so on where the labor rates were what they had been in Hong Kong ten years prior. So it's, I think the Americans certainly won the round of linear circuits, T Squared L, bipolar. The Japanese were a non-factor in the main markets with America and Europe in bipolar products and it was only in the mid '70's that they really started to become a factor based on, as MOS became important and in two ways, one it became important in memories and it became important in consumer products. Now by this time, the consumer products business had moved to Japan because I think there was a conscious decision on the American governments part after the Korean War to just kind of feed certain industries into Japan and let Japan become economically viable because they felt strategically it was very important to have a strong ally someplace in the Far East and Japan was the logical thing and so I think they kind of turned a blind eye to, that the consumer electronics business had moved to Japan and the focus of the electronics industry in the US turned to military and to computers. So that, I mean in the early '60's at Motorola, you know, twenty-five, thirty, forty percent of our business was consumer related mainly supplying companies in Chicago. You know there was Zenith, there was Motorola itself, there was Warwick, a lot of major TV manufacturers centered around Chicago and then in the mid west and that was a big source of semiconductor revenue. And then in the late 60's that rapidly started to go away as the Japanese took over first the tuners and then the black and white TV and then the color TV and so on. So, the Japanese as a factor in the American and European market was not really important and that their market was very closed so nobody really knew what was happening within the Japanese market, but they actually were building a fairly large semiconductor business kind of behind the walls there in Japan, they might - like as Mao Tse-tung said in his Little Red Book, you have to have a secure gorilla base that when you're losing, you better have a place you can go back to lick your wounds and so on. And I think that's what the Japanese were doing in the 70's, I mean they were totally beaten in the World Market or were the Americans in the bipolar products. And the main strategy against them was continually changing technology and using the low cost labor of the Far East. You could completely change the packaging, you could change the cost structure at the back end of the process very quickly with the minimum amount of capital. And, but now when you got into the '70's that changed and they suddenly all of the automation they had been doing within Japan suddenly became very visible. And at that time, most of these companies, there wasn't a third party equipment industry until well into the '70's, so that all this automated equipment was a competitive weapon that was kept within a company. That was the important thing was you couldn't go out and buy a wire bonder but Hitachi, Toshiba, MEC, all built their own inside. Now, it's turned out since then that it's a third party equipment industry has emerged, which then became important in the '80's when suddenly a lot of these advantages went away as well. But then of course when they collaborated with the DRAMS, with the VLSI project in Japan, which was very much subsidized by the Japanese government. Now on the other hand, there was a similar VLSI project that was subsidized by the US government that had very much a military focus. But the Japanese companies very much collaborated with each other as opposed to competing with each other, pooled resources and made a great leap forward in DRAMS and that suddenly started to become apparent in the late '70's.
RW: Well back to Fairchild, you were making your way up the ranks, so tell us about that.
WC: Well, you know the semiconductor industry is cyclical, still is today and, but the, I was running the transistor operation and then Fairchild was eighty-percent semiconductors by this time. I mean even though it was, it had equipment divisions, eighty-percent of the revenue was coming from semiconductors all the growth was coming from semiconductors and then of course 1970 was a big down period and then at that time Les Hogan decided we had to make a change in the management structure and he had to have somebody run the semiconductor division. Right up until that time, it had been almost a little bit of a collegial sort of structure, so he decided to give me that job and so I ran the semiconductor division from 1970 'till 1974 and then in '74 the board decided to make me the CEO of Fairchild Camera, you know. I guess you and I were together in the early '70's at Fairchild and that was when I was running the semiconductor division.
RW: Well I interviewed Les Hogan and I asked him Well what went wrong. Because Fairchild started to decline in that period and he said that he'd had an operation in, before he left Motorola that hadn't worked out and that he in fact had been ill that whole period of time and that he just wasn't running at his usual speed so he couldn't stem the problems. So, what started the decline? What was the cause of this?
WC: Well, the, I mean if you go back in time, you know, I've studied the Fairchild history quite a bit, and the, it started I think when people started spinning out and then if you look at the archetype of the startup company and so on and I know you've talked to Bob, not Bob Noyce, I guess - Gordon Moore certainly, and to other people, you talked to Art?
RW: Not yet.
WC: Ah okay. I mean of course he would give you all of that. But, you know, the way it was started up, the, Art Rock talked to thirty people, thirty companies before he found somebody that was with Fairchild to put the money in. And then of course they started this on the basis that the company could buy out the entrepreneurs, which was Noyce and a bunch of people that came out of Moore and a bunch of people came out of Shockley. And then what happened was they almost instantly invented the planar process and the planar transistor which was very successful and then also very soon after that, Noyce invented the metal over oxide technology, which in my mind was the integrated circuit. We talk about co-sharing of the integrated circuit and so on and so forth, Noyce invented the integrated circuit.
RW: Kilby's thing was nonsense.
WC: You know, I mean from my standpoint anyway, and I'm sure there's, a school of thought from TI will tell you something different, but once you have the concept of the metal over oxide, because I remember both the planar process when I really understood what it was, you know, that sense of wonderment, wow, you know, that solves so many problems and then not long after that was this metal over oxide concept that made integrated circuits possible. It was the same thing, it was like, wow, you know, this was a superior intelligence that could figure this out. And it was just so fundamental and of course we applied it to transistors very quickly, but the real important thing was that it worked for integrated circuits. But, I lost track there a little bit.
RW: What went wrong with Fairchild, why did it start to decline?
WC: Well what happened was that you had this enormous success of this company, but then it was bought back by a New York Stock Exchange company, so it was like this wonderful entrepreneurial effort and then suddenly it's a blob into a bigger company. And I think a lot of the people made some money in that. There was a fair number of people that, in the process of getting bought out, they suddenly had capital at a time when there was very little capital around and they'd seen the cycle happen and they said Hey, maybe we could do this again. And so you had a lot of spinout companies and interestingly enough Noyce and Moore who were the guys that you'd expect to be the most likely to do this didn't. So you had a lot of people, lesser people in my mind, that did this. And so you start to have a drain of people away and then in 1966 the, when Charlie Spork went off to form National Semiconductor and there's a whole story on how all that happened and so on. But the, when Charlie left to kind of revive National Semiconductor, which was a tiny failing company, he took with him, you know, one whole layer of operational management. Charlie was kind of the manufacturing guy at Fairchild and Noyce and Moore were much more the, R & D, the development. You know, even in '66, even though Bob was, you know, technically the person running it, I think Charlie Spork had become very important with this offshore strategy and the volume manufacturing and so on. So when Charlie left and he took a lot of the key people with him, including a lot of linear circuit people, it's not long after they had invented the linear circuit, and very key design engineers like Widlar who basically developed most of the original linear circuits himself just came out of one brain. And they, that had to weaken it. But I think there was a moment in time in '66 at Fairchild when virtually every area of integrated circuit technology was there at Fairchild, whether it was digital circuits faster than anybody else, whether it was linear circuits doing things that nobody else could do, PMOS circuits, it was kind of all there, but it was kind of too much in one confined place. I mean in retrospect Monday morning quarterbacking, if I could take, you know, my brain of today, I would have said look this is what we're going to do, we're going to set up five separate companies, spin them all out, and what we call Fairchild will be a holding company and we'll give pieces of equity to, you know, five different groups of people and it will be a holding company and the primary investor and let these companies go. And, but that's Monday morning quarterbacking.
RW: Well that happened but it wasn't planned.
WC: Without the benefit of Fairchild. And then you got into, now Fairchild was still basically run from New York and they had a series of managements and the guy from Corning, John Carter, probably did the most damage. And he had an idea of leveraging the success of the semiconductor business to go acquire other little companies and I remember when Hogan and the group got there and we talked to the Fairchild people and went into the finances, what was happening, just when the semiconductor industry was exploding, they were asking the semiconductor division to operate, to maximize cash flow when to support these other losing divisions on the East Coast making equipment and so that was why Fairchild was under invested with Hogan and I and the other people arrived we were puzzled that there was no capital investment, that all of the assembly and test equipment, all of the diffusion equipment, was ancient, it was very backward to haven't done any of the things that we'd been doing at Motorola and so our initial conclusion was we just have to put a massive infusion of capital into the company to revive it. And of course as Hogan arrived, Bob Noyce and Gordon Moore left and of course, and Andy Grove and of course they took the silicon gate process with them at the time. I mean...
WC: ...if somehow magically I could come back in time and say what I would have done, I would have sued them and taken all the strategies that Intel later employed to get their position in the market place, I would have sued them because, and prevented that company from taking off with the basic technology that was developed at Fairchild.
RW: But Roger Borovoy, who was the council at Fairchild and would have done the suing, was a good friend of Bob Noyce's and shortly there after joined Intel.
WC: Now Roger was the patent attorney at the time, he worked for Nelson Stone. And you know, there was a very open situation that time and the people really didn't feel that you could necessarily successfully sue people on trade secrets or on patents and I think this all changed some years later when Polaroid won the lawsuit on patents against Kodak on instant photography and after that lawsuit, suddenly the whole strength of patents litigation changed. But in retrospect, if I had taken Intel, if I had been in a position in '68 to take an Intel like approach to this technology extraction that went on when the group set up Intel, it would be a different story today I think. And then over time, at Fairchild, the, you know, we went through the '74 recession, we never lost money in the '74 recession. And through '79 we finally had to back up to doing seven hundred million or so a year and making, you know, fifteen percent pre-tax profits. And then there was a hostile takeover from Gould, which, you know, I think you detailed in your book, and the, and I think at the time that hostile takeover started, the price was sixteen dollars a share and by the time we sold the company to Slumberger, it was sixty-six dollars a share. So, certainly the shareholders made out okay and the, most of the managing people did okay on their stock options and so on. But then after that, the Fairchild story was more or less over. But it's interesting to see that the remnants of the Fairchild it was is recently now spun up as a new company and again still doing five hundred million dollars a year.
RW: It's amazing isn't it? So you had, when you left, you had a do not compete clause. In '79 or so?
WC: Yeah, right.
RW: And you were going to just go and enjoy life because you'd been poor, worked all your life, so this was going to be your chance to just relax, you were independent...
WC: Not really, I was going to go into the venture capital business actually at the time and so I spent a little time learning the venture capital business and the, and you know there's a lot of opportunities at that point in time in the venture capital business. And, but the more I got involved in it, I realized that, again, all these things are cyclical and whereas the venture capital business was looking very good, the more I got involved in it, the more I realized that their problem was they didn't have vehicles to invest in. And so it became clear to me, probably by the middle of 1980 that the, it was probably better to catch the money than to pitch and a lot of the venture guys that I talked to said, you know, why don't you start a company, don't try and get in competition with us, I mean you got to go learn a whole new skill set. If you want to start a company, we'll put the money in it. And I remember I came home and talked to my wife Sigrid and I said, you know, I think I'm going to go start a company. And she said, but you've been telling me for years that all of the semiconductors the world's ever going to need exist and there's probably going to be fewer of them, what's changed? And I said, a lot of things have changed. And so she thought I was crazy, but, and that's when I went through this whole series of events that finally ended up talking to you about a custom integrated circuit company.
RW: That's true, up to that point, no one had done it right, and then with the big guys that didn't want to do it and in fact don't want to do it today because it turns out you can't really make any money at it as we found out and then there's the little guys that didn't have the capability of doing it right. And so it was open, the time was open for us to go and do it right and we did, and blew away everybody.
WC: No, but I don't think, you know, you should say people don't make any money at it. I mean this is a cyclical thing and I think in the, in the last up-cycle, I mean, we were getting up towards the, you know, the high twenties in operating margins and so on, and I think generally one can run this sort of business at, you know, a twenty percent operating margin. But of course the business changes, I mean, when we started, the environment was ripe for custom integrated circuits. I mean, I know at Fairchild, when major customers would come to us and ask us to do custom integrated circuits, generally we would decide not to do that. I mean, we did do custom circuits for the mainframe computer companies, ECL products which were basically ECL gate arrays, that went into the early Cray computers, Unisys, Burrows, I guess I want to say Unisys and it used to be Univac and Burrows. But Fairchild did the Cray, the Univac, the Burrows logic and, but that was not really going at the custom integrated circuit, it was doing a few projects for high volume things. And I think the big difference, what was happening was just as the Japanese had their VLSI program, the US some years before, the military, had financed what was called a VISIC program and that was to somehow accelerate the technology to what was then considered to be an unconscionable goal, which was sub-micron. And it was by fiat, it was determined that any of these major programs had to be designed into sub-micron technology. And it was kind of a very brute force approach to force the technology to progress. And we certainly hadn't planned on this, but it became apparent that by 1980 many of these major, major programs were in dire trouble, that even though either they couldn't make the sub-micron technology or everybody had forgotten about the design problem, they didn't quite know how to design the stuff. Each circuit was designed differently and so on and would have subtle problems with it and so on. And I remember when one of the first things you said to me was when we decided we're going to go down the road, and one of the most seminal things was that you said we're going to let the customer write the test programs. And if you remember at the time, I thought you were crazy. Because in the business, the idea of letting the customer define the specs on his own product would mean your yield would go to zero in my mind, that, you know, you have to have control of the specifications. But the way you thought that thing through I think was seminal to the early success of the company. And when we found out in, by like '82 that our two micron CMOS product would actually meet most of the requirements of all of these one micron vISIC programs, it was magic how so many military customers just said, well we'll get a waiver from the defense department because this solves the problem.
RW: Well, you know, in, you brought up government support, both in Japan and in the US but in Europe there's been government support as well.
WC: Oh yes, sure.
RW: That has largely failed.
WC: Well I don't think you can necessarily say that. Because the question is, what is the long range outcome. I mean if you look at the European government support of the Airbus, which has been ridiculed in the past, but now suddenly Airbus is a very viable competitor to Boeing. They kept alive a strategic industry. If you look at all the government, the millions and billions and billions that they fed into the semiconductor industry in the '50's to '60's to '70's and '80's nevertheless, you've now got three, it's kind of merged and merged and merged. So you got SGS Thompson, which is certainly still a major contender in the worldwide semiconductor business. You know, it's a five billion dollar a year semiconductor entity. You've got Phillips, which really does not get as much visibility as it should but yet one forgets that Philips owns thirty percent of TSMC, the Taiwanese foundry. So Phillips is very viable as a semiconductor company and Siemens, there's almost a resurgence right now in the last few years of Siemens Semiconductor out of Germany and even though they shut some factories with the DRAM debacle that's going on right now, they are a force to be reckoned with. So, you've got three powerful semiconductor companies in Europe, you don't have a lot of little ones, but you do have three big powerful companies. And I'm sure one way or another there's still a certain amount of government subsidy going into it. But certainly look at SGS Thompson, that's now, you know, listed on the New York Stock Exchange as very big viable company. I mean they just, six months ago they raised close to two billion dollars on the New York Stock Exchange.
RW: Well another sort of industrial policy that we don't' really talk about is Taiwan, Singapore, Korea; there's been government support in those areas as well.
WC: Oh yeah, very much so. I mean if you, it's interesting, there's some interesting contrast there. I mean they, the Korean model, not just in semiconductors but generally, has been the Japanese model. They copied the Japanese model. Taiwan essentially copied the Silicon Valley model. There's some pretty good arguments that say a lot of Silicon Valley happened because of the environment of the money that was being fed into the space program and the military programs and so on. I mean the military business was a great source of sustenance for the semiconductor industry in the 50's and the 60's and the 70's and even well into the 80's.
RW: And even at LSI Logic at one point in time it was almost half our revenue.
WC: Absolutely. So, I mean, that's downgraded at times I think for a lot of people, but the reality is if there hadn't been a big, viable military business in the 60's and 70's a lot of these companies would have been out of business at every downturn because when the economy went down that fed through directly into the semiconductor industry, but the military business, at that time, was not independent of the economy. I mean it just kept on going like a, you know, the basic carrier wave and so when the commercial, the consumer, the computer business slowed down, the military business was still there like a rock that was producing maybe thirty percent of the revenues and maybe much more of that of profit. I mean the way people do accounting, it's now always obvious that, but I think the, one way or another there was a fair amount of American government money fed in, that's not true today. And certainly in Japan, if their government had not had the closed market strategies in the 60's and 70's, and probably they would never have created a viable technology industry. So I think governments, you know, have to do that even though I spent a lot of my time arguing against it and so on, the reality is that there's times in with fledgling industries where the government has to do things, either it's got to help the environment or direct subsides or whatever.
RW: Well you mentioned Airbus, which is a huge success heading toward fifty percent share of world market and Japan and Asia have almost nothing in the commercial aircraft market. There's an example where it worked. Another aspect of the semiconductor industry that I'd noticed being a design guy and not a semiconductor guy, is the increasing importance of the design aspect to the business and intellectual property. Can you comment on that?
WC: Well, I mean there's several threads in the semiconductor industry that flow through. And one of it is raw technology, that you've got to be able to meet criteria, whether it's a speed performance criteria, or it's a density criteria. I mean clearly you can't play in a million gate plus chip market if you can't make million plus gate chips. But the increasing importance of intellectual property and design tools certainly LSI Logic design tools were seminal to our ability to create a business. And even today, design tools are very important and it's the way in which you arrange those design tools, which are very important. The availablity of equipment of the right kind is very important. The availability of money, when you could have built billion dollar fabs today, money is very important. So there's a lot of different things which are important but they all have to play together. Now intellectual property is one very important element and you know, we're talking earlier about why has Intel been able to dominate and be in this position and I think the - essentially you almost have to look back to IBM. And there was a point in time in the computer business where IBM achieved this position where they controlled the computer architecture, they controlled the computer software, which were bundled and they got to such an incredibly powerful position, such an incredible percent of the business, that finally the government had to take action and break it down. Otherwise if the government hadn't done that, I'd say there's a, you'd probably be totally IBM dominated today. Now you take that as the archetype and that took government action to make that happen, just like that happened in the telephone business that the government finally had to take action. You can go back even further with Rockefeller, that the oil trust had to be broken by the government, otherwise it's checkmate. One company controls the whole industry like DeBeers controls the worldwide diamond business. Now you've seen that happen with Intel today, who's essentially recreated the IBM position prior to the government antitrust suit in that they control the architecture and with alliance with Microsoft, the software is controlled and that walks in the IP and that IP becomes the controlling element in the PC, the way it's going it probably will be a controlling ailment in most computing - these architectures. And as you go forward, you say well, this is an IP control, it's not because necessarily Intel can manufacture better than anybody else though they're a very good manufacturing company today. It's really the control of the IP, the intellectual property, the patents, the legal position, the copyright control and so on, which nobody's been able to break. So, that is a radically different position than almost anybody else has in the industry today. It will be interesting to look back in twenty years to see what happens
RW: Yes it is, I think what people don't appreciate about Intel is, I think people understand that they dominate the desktop, but I don't think they understand that they now dominate, and as we got to 64 bits, completely control servers, workstations, super computers, everything that we call a computer; general purpose - not embedded. And the Japanese, we're going to take over, they're out, Europeans are out, and what we call computer manufactures take Intel architecture and put them in boxes today with a few minor exceptions and declining exceptions. Well, you brought up the cost of FABS, at a billion dollars a FAB, how can a billion dollar a year company afford that?
WC: Well you just make sure that you get to be a two billion dollar a year company in time to do that. It turns out in the semiconductor business, you know these numbers are big, a reasonable model is your twenty percent of sales a year invested in capital is a sustainable business model. The issue is really in the smaller stages, an established semiconductor company, that's been around for awhile, you know, as long as it's reasonably managed, will generate enough cash flow that it can keep investing at this sort of rate, twenty percent per year...now, it might be thirty percent some years and it might be ten percent other years, because the lumps of capacity - capacity is very lumpy in the semiconductor business, you know, it's like, you know, women having babies, you know, if you're interview them just at the moment when they're actually having the baby, they're saying, I'm never going to go through this again and then a few years later they're having another baby because they've probably forgotten that. And a FAB is very much like that, because you take this billion dollar entity and obviously the first wafer that comes out theoretically is taking on the cost of the whole thing and takes you some long period of time either to develop the business to fill it or the time just to ramp it up to get it to a reasonable percentage of capacity. I mean most of us today know how to get it to yield pretty quickly so that you can probably yield at very competitive yields almost right away when you start the wafer FAB, which wasn't the case. I mean at Motorola in the '60's and I'm sure it was true at the other companies, there would be times when the yield on a line would be one percent and you try and compare that with an automobile factory where, you know, you say Oh well, ninety-nine percent of the cars we try to make don't work. But nowadays, processes and technology paths are well enough understood that you can run pretty good yields from day one, the question is how do you fill this FAB. You take a wafer fabrication plant, you can't just kind of just turn it on and click, it's running at full capacity. Let's say you've got a five thousand-wafer plant, five thousand wafer a week plant, and today you know, we're talking eight-inch wafers. I started on four tenths of an inch wafer. You know, the wafer was like, the crystal was like a pencil and you slice the crystal and the final diameter would be four tenths of an inch and I remember when we moved from four tenths of an inch to seven tenths, seven inch, we thought, My god how huge these wafers are getting today. And now you look at an eight inch wafer it looks like a dinner plate and so now we're starting to worry about twelve inch wafers. But when you start cranking this up, you might have eventually at full capacity eight hundred people perhaps in that wafer fab so you don't just suddenly walk eight hundred people in on day one and boom two weeks later you're running at full capacity. So there's the cost of getting up that ramp and then of course you have to be creating business fast enough to support the output of this thing when it's fully running. You know, it's the same thing like filling seats on an airplane, you know, the airplane's got to run whether there's one passenger in it or there's five hundred passengers in it, it has to go. And the same thing is true with the wafer fab. But once they're up and running generally the margins are pretty good. And so the trick is really to figure out how to do this in a way that you can get continuous profitability and to some extent I think the foundry business that's emerged is a way of doing that. That you can either have a fabless company and in some ways at LSI Logic we were the first fabless company, we almost invented that strategy. I don't think we give ourselves a lot of credit for it at the time because it was kind of like knee-jerk reaction, we have to, because starting on six million dollars, we clearly at those, in those days you needed to spend a hundred million to build a fab and I remember the ridicule that people said, How on earth can you start a semiconductor company on six million dollars? And the, so we had to invent the fabless idea and then of course lots of other companies have copied that since then. But it's clear as you move forward the answer to the Fab and nonFAB company is if you want to be leading edge technology you've probably got to use, make your own wafers. So you need to meld a mix strategy that some use of foundries and some use of your own mainstream factories. So you buy wafers from a foundry at higher variable cost, but much less fixed cost. And the management trick as we move forward is how did you melt these two things together because quite often you design products in and a technology will last for much longer then you want to run those wafer fabs. So you can almost have, as you were talking earlier, a manufacturing strategy is to be the last guy with a fab rather than having, so that another five people can shut down their fabs and come to you so the industry has one fab of a given technology that runs economically rather than five of you or ten of you having to run an old factory at a fraction of capacity, which is very inefficient. And I think this has happened in the light bulb industry and lots of other industries over the generations.
RW: I think now that if you're in business, you don't have to make everything yourself no matter if you're in the automobile business or what.
WC: And you know that's been a big difference with the Japanese. I mean the Japanese have stayed with the vertical, they used to be the Zaibatsu and then it moved to being the Koetsu where you have a large, and it's like Henry Ford with the River Rouge Plant where he stared with his own coalmines and his own steel mines and his end product was an automobile. I mean the Japanese bought into that idea and stayed with it for a long, long, long time and that loses a lot of flexibility. You know, most of the electronic companies today don't stuff their own circuit boards, most of them don't make their own chips. Those that do try and make their chip business into a viable stand alone business if they can. So I think you know, outsourcing becomes much more important and not just in our industry but, I mean in terms of equipment, semiconductor equipment, when I started in the '60's we used to build all of our own equipment. I mean the first alignment tools that we used was a ninety-degree straightedge and we'd put two flats on the side of the wafer and you'd align it by just shoving it into the corner. It cost two hundred dollars. Now it costs me five or six million dollars for a alignment tool.
RW: Well, finally...since we got out of school, the status of the engineer, the chemist, the physicist, has changed a lot. When we got out of school, we expected to go to work for some company and, as an employee, and there would be management, people in suits of a high status that would tell us what to do. And that's largely changed and now the status of high-tech people is perceived and in terms of money, in terms of lovely homes, is changed. Have you thought about that?
WC: That was, that was very true in England and that's why I left England. I mean I always wanted to start my own company so when I was about fifteen, that's why I went into chemistry, being in those days I wanted to, I was going to start a plastics company. And remember the Dustin Hoffman movie where, The Graduate, where somebody says, Get into plastics, young man. I guess I was part of that generation because I was a chemist, or chemical engineer. I always wanted to go start a company, it just never occurred to me that it would be an electronics company or a semiconductor company. I was thinking plastics and so on, but the. That was one of the reasons I left England, maybe the main reason, was because there was no way in England that a technical person, a technical person, by being technical you were roughly in the same category as an electrician or a plumber or a mechanic. And very few people with perceivable scientific background would end up in a management sort of position. And that was why I left. And so maybe I saw this differently from you because I hadn't been brought up in this country, so I kind of arrived in this country in 1960 so I saw only what I saw in 1960. And certainly the industry that I happened to end up in, the key management people were all engineers. So I mean, and that was even on the east coast. So I mean David Bacalar who ran Transitron was a PhD who had been at Bell Labs roughly at the same time as Shockley and Bardayne and so on, Bardine. And so my experience of America, with my point of entry, was Wow, this is great, everybody here has got a technical degree, so I've got the right qualifications to succeed in this environment and certainly by the time I got to Motorola, kind of my model there was C. Lester Hogan and he was a Harvard Professor of Physics, he wasn't a Harvard professor of English Literature and everybody in that chain was technical. So you know, my perspective was a little different than your's because you were kind of brought up in this environment in the previous ten or fifteen years.
RW: Well, thanks Wilf for all your comments.