You are here

Dr. Rodgers' Speech: An Argument Against Antiturst Exemption for the U.S. Memories Cartel, November 14, 1989 | Cypress Semiconductor

Dr. Rodgers' Speech: An Argument Against Antiturst Exemption for the U.S. Memories Cartel, November 14, 1989

Last Updated: 
May 24, 2012

Dr. Rodgers' Speech: An Argument Against Antiturst Exemption for the U.S. Memories Cartel, November 14, 1989

T.J. Rodgers

Wescon Keynote Speech
San Francisco, California
November 14, 1989
STAN BAROUH / For Electronic Buyers' News
T.J. Rodgers and Congressman Jack Brooks:
"You're a good eggheaded boy."
T.J. Rodgers
President & CEO
Cypress Semiconductor Corporation
Address to the Armed Forces Communications & Electronics Association
Washington, D.C.
January 18, 1990

The press has labeled me as opinionated and cocky, but I'm not going to take them on here today since I see many familiar faces in the audience. Instead, I will talk about my other side, about my own version of "Mr. Smith Goes to Washington." Let me tell you what happened at the U.S. Memories hearings in Washington, what my concerns are about American competitiveness, what my fears are about the proposed consortium solution to the problem, and then offer a tried and true entrepreneurial alternative.

First of all I'm the first to admit that our country has competitiveness problems. One would have to have his head in the sand to not see the problems facing the semiconductor industry, in general, and the semiconductor equipment industry, in particular. The question before us is not whether we have problems, but how to solve them. The solution in vogue today is to treat competitiveness problems with consortia. Consortia are unwieldy groups, usually assembled by large companies (themselves with a good share of self-inflicted competitiveness problems), which are then supported by government money, granted tax advantages and allowed relaxations of antitrust laws-the Amtraks of our high technology industries.

The formation of 50 Amtraks certainly will not make America competitive. What will is the entrepreneurial solution. This approach capitalizes on American ingenuity, start-up companies, and fast-paced technology development-the basic driving forces of Silicon Valley. This is the right way to remain competitive, and later I will tell you what needs to be done to turn loose that entrepreneurial energy.


I did not want to go to Washington. I run a company in an era of "take-no-prisoners" investor mentality, but a staff member of a House Judiciary Subcommittee invited me to testify after he had read a one-sentence quote of mine in The Wall Street Journal which criticized Sematech as a poor and inefficient solution to the semiconductor competitiveness problem. He told me that the next consortium, U.S. Memories, was running through Congress unopposed. U.S. Memories had landed in force and rolled through the first of three Subcommittee hearings; now, I was being requested to present a balancing opinion in the second hearing. I figured that testifying was probably a misuse of my time, but the more I thought about our country's technology industries heading toward government control and the inherent threat posed to the entrepreneur, the more I became convinced that someone had to speak out against these potentially catastrophic governmental social engineering efforts.

The second hearing was held by the House Judiciary Subcommittee on Economic and Commercial Law-the Antitrust Committee. The issue before the Committee was: "Should we pass a law that exempts production consortia from triple-damage penalties, in the event that they get convicted of violating the antitrust laws?" The National Cooperative Research Act (NCRA) of 1984, the first law of that sort, gave immunity from the triple damages to research and development consortia such as the Microelectronics and Computer Technology Corporation (MCC) and Sematech, the semiconductor consortium. The new issue before Congress was whether we should extend the NCRA's antitrust protection to production consortia like U.S. Memories, by allowing them to produce and sell product in the open market, without liability for the potential triple-damage lawsuits to which all other companies are liable.

That was the technical issue, but it was really only the latest skirmish in the larger battle between big and small-between large, entrenched corporations and entrepreneurial companies.

I thought I would get clobbered in Washington. My opponent in the two-man debate, Gordon Moore, is not only a likeable person, but a semiconductor industry legend who has had a lot of practice testifying in Washington. Gordon's pitch is quite formidable. He makes it very clear, when you are before Congress, that you are on his turf. Gordon opens by saying, in effect: "I represent Intel (we invented the DRAM; we invented the SRAM; we invented the EPROM; we invented the microprocessor); I represent AMD, LSI Logic, National Semiconductor, Hewlett-Packard, DEC, IBM, the Semiconductor Industry Association (that's almost all of us) and the Electronics Industry Association (that's the rest)."

Damn, Gordon, so I'm opinionated and cocky!

Despite my initial feeling of being overwhelmed, I testified on what needed to be said: that consortia damaged competitiveness, rather than helping it, and that you can be a patriotic individual and still vote "no" on stop-gap measures, such as the consortium "du jour." Since little was known publicly about U.S. Memories at that time, I used Sematech as an analogy. Sematech is, of course, the semiconductor consortium, founded with direct federal subsidies, and located in Austin, Texas. Its current charter is to advance semiconductor and semiconductor equipment technology and solve our industry's competitiveness problem. Sematech is a typical consortium story: an industry drifts into a non-competitive position, wraps its spokesmen tightly in the American flag, runs to Congress to claim the Japanese are taking the market away and states, in effect, if Congress does not do something immediately, the industry will fail. The spokesmen imply that if you (Congressmen) love America and you want our industry to prevail, you should give us money. And that argument works, especially when there is a panic. (Incidentally, both Sematech and U.S. Memories have red, white and blue logos. Sematech has also spent money to design a modern Sematech version of the U.S. colonial "Don't tread on me," coiled-snake flag.)

Sematech was formed with a $100-million-per-year federal subsidy. But, if you compare Sematech's Congressional hype to what really came about, you get two very different pictures. After the patriotic fervor subsided and the attention of Congress went away, Sematech got down to the day-in-and-day-out life of running a consortium. I pointed out that Sematech was funded to help all semiconductor companies develop technology for our industry. However, Sematech quickly became a very exclusive club. There are 923 semiconductor companies in the U.S. and somehow Sematech only managed to snare 14 of them. When the $100-million-per-year subsidy went into Sematech, its directors-who are all from large companies-simply created a dues structure set so high that small entrepreneurial companies, including Cypress and most other Silicon Valley companies, could not afford to join. The net result is that Sematech is exclusively a big company's subsidized club, used to fight both the Japanese and entrepreneurs, rather than the national institution it is supposed to be. (During its membership drive, Sematech even indicated that non-member companies were not guaranteed licenses to Sematech patents, and that Sematech members would be given preferential access to Sematech-developed manufacturing equipment sold in the open market.)

Sematech was chartered to develop next generation technology. Today, Sematech is working on 0.8-micron CMOS. We've been shipping 0.8-micron CMOS since 1986. All the competitors I worry about daily-mostly entrepreneurial companies-produce 0.8-micron CMOS, as well. So, Sematech is not working on the next generation of technology; it is working on today's technology, which is already being widely shipped, to help some of the big guys in the Sematech club catch up. The organization launched to save America from the Japanese ended up as government-subsidized research and development for 14 large companies, 12 of them with sales greater than $1.5 billion per year.

I concluded my congressional testimony by stating that although Sematech was a bad idea, U.S. Memories was a worse idea, since it would compete unfairly in the open market against real companies. I asked the Congressmen to look hard at the promises of U.S. Memories to make sure that they would not inevitably drift into an unfair attack on entrepreneurial companies-generally, the ones still competing successfully against the Japanese.


About the Committee: Jack Brooks runs it. He's a wiry Texan good-old-boy, a World War II veteran, and a very tough, competitive person who runs his committee tightly. He came down before my testimony, indicated that he had read the written version and said, "This is a real rip-snorter, a real rip-snorter, gonna enjoy hearing from you today."

Thank you very much Congressman."

And he reached up and he patted me on the top of my head and said, "You're a good egg-headed boy. We need more like you."

I later asked Jon Yarowsky, Brook's chief counsel, "Is that good or bad?"

Jon said, "That means he likes you."

Something unexpected happened that day, and very few have had a chance to hear about it. I'd like to share with you some transcribed notes of a recording I made. The press reported the flow of the hearing in the following way: Gordon Moore, spokesman for the semiconductor industry, stated what had to be done to make America competitive, while cocky, opinionated Cypress's Rodgers made some reasonable points.

Here is what was really said in that room that day. We had completed the formal statements, which I have already synopsized for you, and the Q & A session came up. Chairman Jack Brooks got to ask the first questions. And Gordon Moore, the senior person testifying, answered first.

Mr. Brooks: Dr. Moore, did you consider inviting Dr. Rodgers' firm, Cypress Semiconductor, into the new U.S. Memories consortium?

Dr. Moore: Well, I'm sure that given any indication that he might accept, he would be enthusiastically invited.

Mr. Brooks: You did?

Dr. Moore: It turns out that Cypress Semiconductor is one of the leading semiconductor companies that is not a member of the SIA. So he didn't participate in the original formation, where he certainly would have been invited.

Mr. Books: He was not invited.

Mr. Moore: He will be. All U.S. suppliers and users of DRAMs or semiconductor devices will be invited to participate in U.S. Memories.

Mr. Brooks: But Dr. Moore, I don't understand. I thought that they already had that meeting and set up the members of this U.S. Memories.

Dr. Moore: I ought to clarify that, then, Mr. Chairman.

And that warm-up exchange went on for about five minutes. In earlier testimony I pointed out that the government-subsidized research at Sematech was really unfair because one of their member companies, namely Gordon Moore's Intel, had made $600 million in profit in 1988, and that Intel hardly needed a helping hand from the taxpayers.

Mr. Brooks: Dr. Rodgers indicates that your company actually invented the DRAM and the static RAM only to lose or abandon the market position on these products. Why hasn't your company been able to capitalize on such innovations under its own control?

Dr. Moore: Well, first of all I'd like to say that of the semiconductor companies, I think Dr. Rodgers would admit that Intel was the most successful.

Dr. Rodgers: Intel is number one, I'll agree.

Dr. Moore: We have developed several opportunities along our way-

Mr. Brooks: $600 million in income last year?

Dr. Moore: Pardon?

Mr. Brooks: $600 million-profit-


A new questioner, Congressman Smith from Florida, cut in:

Mr. Smith: The sales are what?

Dr. Moore: About $3 billion.

Mr. Smith: So, you got a 20 percent net profit?

Dr. Moore: We had an exceptionally good year last year because there were other product areas where we have a strong leadership position. In addition to the two memory products mentioned, Intel also invented the microprocessor-

Mr. Smith: Would you think at all about subsidizing the U.S. government?

Dr. Moore: Well, Mr. Smith-

Mr. Smith: We need a subsidy more than you do, Dr. Moore.

Dr. Moore: I appreciate the problem. On the other hand, we lost $200 million in 1986. This is a high-risk business. Anyhow, we have a very sound position in the microprocessor/complex logic device area and are a major participant in one of the other memory product areas that we use as a technology driver as well as an important part of our business-

Mr. Brooks: How did you just blow it on the DRAM and the static RAM? You had it; you made money on it; you just up, quit and got out, and you all were the big animals in the park.

Dr. Moore: Well-

Mr. Brooks: How did that happen? I don't understand it. Why didn't you step on them?


When the questioning got around to Congressman Lawrence Smith of Florida, U.S. Memories was put under the microscope.

Mr. Smith: One of the things the cartel is trying to do is to avoid illegitimate, unfair competition right now from the Japanese and other consortia in Europe on the very same issues- not on the same product, but on other products. Isn't that true?

Dr. Moore: The cartel?

Mr. Smith: Yes.

Dr. Moore: Which cartel are you referring to?

Mr. Smith: U.S. Memories.

Dr. Moore: Well, that's not a characterization I'd use for U.S. Memories. U.S. Memories is a shot to try to re-enter the DRAM business which is very important to the entire electronics industry. A cartel certainly implies power far beyond that of a five percent market participant…

Mr. Smith: Well, let me ask you this. You're a public company, Intel, correct?

Dr. Moore: Yes.

Mr. Smith: In 1985 and 1986 you lost the market on DRAMs, Instead of chasing after the market, you made a decision corporately to move your money elsewhere and enter other product lines, right?

Dr. Moore: That's right.

Mr. Smith: And you agree that your moving to those product lines resulted in 1988 in a huge $600 million profit on $2 billion of sales. Now, that's to some degree promoted by shareholder loyalty, I assume. I mean, you've got shareholders who want a return on their investment, correct?

Dr. Moore: Yes.

Mr. Smith: All right. Now, why should the U.S. taxpayers subsidize to protect your shareholders?


By that time, it was nearly my turn. I looked over at Gordon. He had so many arrows in him that he looked like a porcupine. And, I was next.

And guess what? Not one hostile question was asked of me by anybody on the committee, even though two members of that committee have sponsored bills supporting U.S. Memories. And here's the final question put to me.

Mr. Smith: Mr. Chairman, I have just one other question. Dr. Rodgers, how old are you?

Dr. Rodgers: 41.

Mr. Smith: Well, first of all you're a breath of fresh air around here. Secondly, I commend you for what you've done with your company. How old is Cypress?

Dr. Rodgers: 6 1/2 years.

Mr. Smith: 6 1/2 years at a time when the Japanese are making their heaviest penetration against us, starting in about 1982, correct? I really do commend you. I don't know why we have lost the entrepreneurial spirit. There are people in the country who still have it.

So if any of you think that U.S. Memories will have a cakewalk through Congress, it's because you aren't aware of the strong sentiments in Washington against consortia.

The point is that Mr. Smith went to Washington expecting to get clobbered, and they actually listened to him. We entrepreneurs are dealing with a formidable opponent, with formidable lobbying capability. I wrote my own testimony. And we have zero lobbying effort. The jury is still out on who will win, but the point is, Congress is not reacting favorably to the creation of Amtrak-like consortia to solve our problems. No one gets a comfortable feeling when you say, "Let's have more Amtraks. That'll fix the problem. When we have a lot of Amtraks, we'll be competitive against the Japanese."

People respond favorably to entrepreneurs who take risks, start companies, and win--just like Intel did when they really did invent the static RAM, the dynamic RAM, the EPROM and the microprocessor. Back then, Intel did not need to travel to Washington to ask for anti-trust exemptions. They were in Silicon Valley, kicking butt.

I got another surprise. A lot of people heard and read about the hearing--that somebody actually had the nerve to say something against the conventional wisdom of the establishment. I received many letters. And all but one of them was favorable. They said, "We agree with what you say, and we're glad somebody finally said it." The one unfavorable letter came from the government relations director of a corporate member of the Sematech club.


The third hearing was held about a month ago. The Congressional protocol is that you can testify only once. As a result, whichever side can produce more good, fresh witnesses is the side that carries the day--with the other side going underrepresented. I went through the letters I received and shot back letters to their authors that said, "Look, if you feel so strongly about this, you'd better get on an airplane and fly to Washington and tell Congress why it's a bad idea, because I can't go back." I expected most of the people to be too busy or to feel that you can't change the system. Unexpectedly, I had six volunteers-and they were impressive people-Lou Tomasetta, the president of Vitesse Semiconductor Corporation, a gallium-arsenide company in Southern California; David Coelho, the chairman of Vantage Analysis Systems, a Silicon Valley software company; Tom Peters, the author of In Search of Excellence and Thriving on Chaos; George Gilder, the author of Microcosm; Ed Sacks, the president of Zilog; and Dick Shaffer, formerly a technology columnist for The Wall Street Journal, and now the president of Technologic Partners. It turned out the pro-consortia people did not produce as many witnesses, and we had to cut back from six to four. Those new witnesses brought new depth to the concern that America's successful entrepreneurial culture is being threatened by governmental intervention.


Let me tell you, in short form, what Lou Tomasetta said. His testimony was titled, "Maintaining America's Strength Through Competition, Not Consortia." Lou has spent most of his professional life working on gallium arsenide (GaAs), a semiconductor faster than silicon, and his company is one of the first to have it working for real. Vitesse Semiconductor Corporation is one of the two largest merchant gallium-arsenide companies in the world. Their products work; they're fast; and they're succeeding in the marketplace. Lou's argument is different from mine; he wants to prevent a consortium that has yet to be formed. Here's the gist of what he said:

Gallium arsenide really works. I've been working on it for a long time. A few years ago, when we were still trying to get market credibility for gallium arsenide, I approached virtually every major American semiconductor company and made them a sweetheart offer-gallium-arsenide technology at favorable prices, just in return for second sourcing so we could gain initial credibility in the market. And I was turned down by every American company I approached. And so, Congressmen, a few quarters from now, when Vitesse and a few Japanese companies are the only people making profits in the now-critical gallium-arsenide industry, please do not "help" our industry with a U.S. Memories-type consortium, which provides a reentry vehicle for large U.S. companies who were too short-sighted to invest in gallium arsenide when they were given an excellent opportunity. Large U.S. semiconductor companies have chosen not to invest in gallium arsenide. So be it, but don't subsidize them back into the business after they realize their mistake.


Following Lou Tomasetta's testimony, we heard testimony from a different industry--software--and it dealt with a different consortium, not with Sematech or U.S. Memories, but with the Microelectronics and Computer Technology Corporation, or MCC. David Coelho, the chairman and one of the founders of Vantage Analysis Systems, a Silicon Valley software company, was the speaker.

David sent me a letter, which said, "I read your letter about Sematech and U.S. Memories, and I feel exactly the same way about MCC." Of course, he soon received my letter back saying, "Fine, now's your chance to tell Congress." And he took me up on it. Vantage makes a very advanced Computer Aided Engineering, or CAE, software program that is used by design engineers to enhance their productivity. There are only two companies in the world that have this unique software tool: Vantage and the MCC.

The MCC was formed when the Japanese announced that they were forming a consortium to produce a fifth-generation computer--one with artificial intelligence, a computer that could learn and program itself. Sensing that golden opportunity, the consortium boys went to Congress, decked out in the red, white and blue, and said, "The Japanese are going to have a fifth-generation computer. We need a consortium for fifth-generation computers. If you don't fund us, the U.S. computer industry will be in trouble." They got their subsidy, and it's called MCC. It's partly supported by government money, and it is housed in a building leased for $1 a year on land also leased for $1 a year. The MCC membership includes 19 huge companies, including IBM, DEC, General Electric and Honeywell. They have single-damage anti-trust immunity, provided to them by the National Cooperative Research Act of 1984. And they're supposed to work to make sure America has a fifth-generation computer before the Japanese get there.

You are probably asking yourself. "How could this national effort on fifth-generation computers possibly compete with a start-up software company in Milpitas, California?" The answer is very simple: the MCC is out selling software in the open market. Vantage has some of the hottest CAE software in the world; the only other company which has it is the MCC, and they're selling it in direct competition with Vantage.

But, you say to yourself, "How can they sell it? They're government-supported; that's unfair." Well, the answer is that the MCC doesn't sell software. It sells "subscriptions." For a nominal "catch up fee," you can get a subscription to the MCC. Once you are in the MCC, you are allowed to use their products. By analogy, the MCC doesn't sell books-they sell library cards-but once you've paid for the library card, you can use their books for free. It seems MCC is a little tight on its budgetary constraints, and a little extra cash never hurt a starving consortium.

Then, you might say, "Okay, government intervention in an open market is an undesirable by-product, but isn't the MCC an extremely important national effort?" But David Coelho methodically laid out the implications of this situation for Congress and demonstrated that the damage MCC could do to our software industry could easily be worse than my worst nightmares about the damage Sematech or U.S. Memories might do to the semiconductor industry.

Think about the potential. The MCC strikes, and hurts Vantage. The venture capitalists get the message to avoid funding any company that intends to produce any product that MCC is working on or might work on, because if they do, three years and $10 million later, they might get clobbered by a government-subsidized competitor that sells software below cost. In one of the worst trades I could comprehend, we would then replace our entrepreneurial software industry with the MCC.

Our software industry has been extremely successful-we have about an 80 percent world market share, export 55 percent of what we make-and we don't need any "help" from consortia. Why are our software companies so successful? George Gilder points it out in his writing. In every industry in which we have more competitors and more start-up companies, we beat the Japanese and vice-versa. The U.S. has 14,000 highly competitive software companies. The Japanese, on the other hand, have only 30 software  companies which are mired in large, vertically integrated companies. In the software industry, the Japanese are not competitive, and we are.

Think of the grave implications of allowing the MCC to hurt our software entrepreneurs: the top four software companies in the United States account for 60 percent of all software sales and every one of them is a start-up that commenced operations after 1980. It would take less than a decade for a government consortium, pushing venture start-ups out of the way, to create bureaucracies that could gravely injure our software industry, one of our greatest entrepreneurial successes.


Tom Peters' testimony was entitled, "Just Say No to U.S. Memories." Tom's testimony blasted the underpinning assumptions behind U.S. Memories. The basic fallacy of their argument is that the Japanese form massive production consortia which are willing to lose money until American competitors go out of business. U.S. Memories' supporters argue that unless we have organizations of equivalent mass ready to meet the Japanese head-to-head in their sumo ring, we haven't got a chance. Peters points out that, in fact, the Japanese don't have any production consortia. The basic premise of the production consortium argument-that we are competing against the Japanese manufacturing consortia-is absolutely wrong.

Take a look at the automobile industry-an industry where we're not too competitive. Unlike the software industry, the reason we're not competitive in automobiles is that the U.S. has three slow-moving car companies, while the Japanese have fourteen of them, each of which has been successful in attacking our industry. They've won because Nissan, Toyota and Honda have beat GM, Chrysler and Ford one-on-one, not because they have ganged up on us.

Tom's second point was that Japan is currently trying to become more entrepreneurial. I know that: they invited me to give a lecture series in Japan on the American entrepreneurial system in 1985. The Japanese want very much to become entrepreneurs, to become more westernized so that they can strip away our most important competitive advantage. You read all the time how Japanese companies are becoming more entrepreneurial. Ironically, just as they realize that their sumo cartels are obsolete, we're about to run off and copy them-and with government subsidies, no less. (And, by the way, that Japanese fifth-generation computer-the one we were so worried about that we had to form the MCC-was just introduced, and it was an embarrassment. The thing is the size of two refrigerators; it is less powerful than many commercially available supercomputers; it doesn't talk; and it doesn't learn. Japan and America both wasted money on fifth-generation computer consortia.)


The next witness, George Gilder, has the intellectual, rumpled look and disheveled hair of a Harvard professor. George, the author of Microcosm and a guru of the microelectronics industry, got up and said, "First of all, Tom Peters and I disagree on about everything. I'm a conservative, supply-side Reaganomics Republican, and he's a lifelong Democrat." And then he said, in just about these words, "The one thing we do agree on is that U.S. Memories is a turkey-looks like a turkey, walks like a turkey, smells like a turkey." After that, he went into his discussion of the real problems confronting the American semiconductor industry:

Capital gains tax:  It is too high; therefore, people avoid capital-gains-oriented investments. As a result, they put their money in banks and government loans instead of buying stock. Thus, stock prices are low and technology companies cannot easily raise the money they need to grow.

Education:  We do a poor job managing it and we don't spend enough money on it.

Non-productive Consortia:  He listed the ill-conceived consortia I've been discussing-Sematech, MCC, and U.S. Memories.

Regulations:  He showed how government regulations are killing the telecom industry.


After the July and September rounds of congressional testimony, I think the consortium fad is starting to fade. You have read recently that the Bush administration is considering reducing funding for Sematech. Such ventures are increasingly viewed as expensive and under-productive.

Time is running out for U.S. Memories. It started up months ago with seven member companies, and it's been stalled there ever since. Although U.S. Memories keeps promising to announce new members, such companies as Apple, Compaq and Tandy, big users of dynamic memories, have announced that they're not going to join the consortium.

IBM, which still strongly supports U.S. Memories, has hedged their bet. They just announced a deal with Micron Technology, an entrepreneurial company in Boise, Idaho. IBM agreed to transfer their 4-megabit dynamic RAM to Micron. Now, IBM is going to win either way in its bid to help the American semiconductor industry.


My question is: Wouldn't we be more competitive, wouldn't we be better off, if we had a thousand companies competing in the semiconductor industry, each a tough, fast moving, entrepreneurial competitor? Wouldn't that make us more competitive than having two or three "Amtrak-Semi's" vying for international competitiveness?

Entrepreneurial companies challenge the wisdom of established companies. Five years ago, if a semiconductor company executive had stated, "We will manufacture completely in the United States," the establishment would have said, "That's a going-out-of-business plan. You have to go to Malaysia, where labor costs fifty cents an hour."

The fact is that Cypress doesn't just make wafers in the United States; we do assembly and test here too. We do over 95% of all our manufacturing in the United States. And, guess what? We have better profitability than the big guys-the people who hire as many Malaysians as Americans. The entrepreneurial environment constantly challenges the conventional wisdom of established companies, and, in many ways, has changed the way our industry does business.

Let's not copy obsolete Japanese consortia. Let's get our government to start working on the things that the government is supposed to be working on, such as improving education and reducing capital gains taxes. I'll tell you a quick story about capital gains. I recently wrote to Congressman Leon Panetta to urge him to vote yes on a capital gains tax reduction. His reply was that he couldn't support a capital gains tax reduction because he could not vote for "a tax break for the rich." How long are we going to allow our representatives to hide in that Trojan horse?

Take a look at NMB-a third-tier Japanese semiconductor company, a good high-quality manufacturer-but a third-generation company. They just went public and raised $472 million. That's three times more money than all of Cypress's six public and private offerings combined. They did it in a Japanese market in which the maximum capital gains tax is one percent of the asset sold. Everybody there buys stocks with capital gains potential because their profits are almost untaxed. Perhaps the Japanese do give tax breaks to the rich. Perhaps unemployed semiconductor workers in the United States will take solace over the fact that no "fat cats" were given tax breaks for investing in their semiconductor industry. And if you look at the NMB offering, I doubt that the $18,600 price per share was intended for the small investor.

We need our interest rates lowered, so we can borrow money at competitive rates. And education: we need to spend less on missiles and airplanes, and more on our colleges to turn out more engineers to compete with the Japanese. The anti-trust laws: we should not relax our anti-trust laws. Just as the fair trade laws insure fair competition between countries, the anti-trust laws insure fair competition inside the United States. We must also enforce our fair trade laws vigorously to ensure a level international playing field between countries.

Let's compete using what we know works-our entrepreneurial system-strengthened by the American traits of risk taking, creativity, and diversity. In short, let's make America more competitive by enhancing the best tool we have-American ingenuity.


First of all, I would like to thank the Air Force Colonel Will Stackhouse for getting me interested in this conference and Rear Admiral Paul Tobin for inviting me. I think the issue of American semiconductor competitiveness, or as your conference is called, "Chip Wars," is a timely one.

Last weekend, I was standing in line in a supermarket in front of the news tabloids. One of them had a feature article with an interesting title, "The 24-Hour Quick Fix Diet." Nobody believes in 24-hour quick fixes; that's why the diet article is in a supermarket tabloid and why the U.S. Memories consortium, a quick fix diet for the semiconductor industry, is dead. There is no more a quick fix in the semiconductor competitiveness area than there is a quick fix for our education problem. Over the period of a decade and a half, through under-funding, parental neglect and management incompetence, we allowed our educational system to decay. We are now looking toward the next decade in which multiple efforts applied from kindergarten through graduate school will be required to restore the system.

Sandy Kane, the president of U.S. Memories, and I agree on one issue: there is a true competitiveness problem in our semiconductor industry. Certainly the problem is routinely overstated to achieve the sense of despair that drives people-and Congress-to action. If I asked you who is more productive, the average American or the average Japanese, how would you respond? If I asked from which country the United States imported the most, which would you name? And if I asked you what country owns the most assets in the United States, which would you name? If you answered Japan three times, you were wrong each time. The average American is actually twice as productive as the average Japanese. What is happening is that the average Japanese is getting more productive faster than the average American, closing the gap between us. That growth-rate difference is indeed a bona-fide problem, but we should not exaggerate it into a second-place claim for America. We import more from Canada than we do from Japan. If you were going to be hit by our largest import, you would be hit by a falling log, not a Toyota. Finally, the Japanese have only recently passed the Dutch in ownership of American assets to assume a relatively distant second place behind the British. Why have we not been screaming for the last decade about British ownership of our assets, as we have about the recent Japanese acquisition of the Rockefeller Center by Mitsubishi?

There is another very important factor that the gloom and doom critics also routinely ignore-the yen-dollar exchange rate. Remember ten years ago when we used to watch the contest for the largest bank between the Bank of America and Citicorp? Now the stories are all about how the top ten banks are all Japanese. Of course, because banks have to be compared to each other in common currency, the exchange rate determines which bank is biggest. If you give me control of the exchange rate, I could make the Bank of Botswana the biggest bank in the world. Now look at semiconductors. If semiconductor market share were modeled hypothetically by a billion-dollar Intel representing the United States share and a billion-dollar Nippon Electronic Corporation representing the Japanese share, each country would have 50% of the semiconductor market. If we then devalued the dollar relative to the yen by a factor of two-just what has happened over the last few years-without any real changes at all, Intel would remain a billion-dollar company and Nippon Electronic Corporation would become a two billion-dollar company in devalued U.S. dollars. The predictors of doom would make headlines with that statistic: "U.S. semiconductor industry market share drops from 50% to 33%!"

Even with the mitigating factors I have discussed above, I am convinced, along with other U.S. semiconductor executives, there is a substantive productivity problem in the U.S. semiconductor industry. The supporters of U.S. Memories and I simply disagree on the solutions. I do not feel that government intervention-the propping up of companies which have become uncompetitive-will make America strong again. I really have no arguments against the formation of either research and development or production consortia per se. The unacceptable aspect of both Sematech and U.S. Memories is that they are supported partly with government money to the advantage of a few companies and to the disadvantage of successful entrepreneurial companies in the same market. The demise of U.S. Memories has highlighted the fact that there is no quick fix for our competitiveness problem. We must all look in the mirror and resolve to go on a diet-which may take a decade-to make each and every one of our companies lean and competitive. There is no law that can be signed, no subsidy that can be given, no further national declaration that can be made that will make our industry more competitive. The CEO of each semiconductor company must look in the mirror every morning at the only person who can bring our industry back.

I had little to do with the demise of U.S. Memories. My only contribution was to open the debate on the issue. I debunked U.S. Memories' unfair argument; namely, that if you do not support consortia like Sematech and U.S. Memories, you are somehow unpatriotic. The U.S. Memories logo, like that of Sematech, is red, white and blue-just to remind you not to attack the flag. I went to Congress and established that I am as patriotic as anyone within U.S. Memories, that the company I help to found, Cypress Semiconductor, succeeds in head-on competition with the Japanese, and that you can be a loyal American and a tough competitor and still think that U.S. Memories is a lousy deal. Once the American flag gets stripped off of the venture, it is easier to debate the merit of the U.S. Memories proposal objectively.

The demise of U.S. Memories was the result of an open debate and the objective evaluation which is required by investors when billion-dollar sums are involved. I have done a few back-of-the-envelope calculations to show that U.S. Memories missed the mark as a reasonable venture investment by factors, not percent. The most striking of these calculations shows that if U.S. Memories were required to achieve the nominal factor of 10 to 20 times stock price appreciation in the three- to four-year time frame required by most venture capitalists to trigger investment, that venture would have been required to grow from zero to five billion dollars in annual sales in a period of three to four years, more growth in four years than DRAM-inventor Intel achieved in 20 years! I think that even the people who supported U.S. Memories strongly would have to take a second look if their personal retirement funds were to be invested in U.S. Memories stock. Simply, U.S. Memories died because it was a risky investment which demanded a huge billion-dollar investment, but which promised a substandard rate of return.

Since I am in Washington, I will use the insightful words of a U.S. Congressman whose name I have forgotten to give advice to our government the next time they consider quick fixes like U.S. Memories, "Bravely, do nothing." The government should help to create a favorable environment for business, but it should not interfere with business, In the words of Clyde Prestowitz, "If government is given the business of picking winners or losers, we all know it will pick only losers."

Let us all take the failure of U.S. Memories as a positive sign that we admit there are no quick fixes to our problems and that we will go forward with resolve and hard work to improve our semiconductor industry.