You are here

Dr. Rodgers' Testimony: Some Concrete Proposals to Make the Semiconductor Industry More Competitive, July 23, 1991 | Cypress Semiconductor

Dr. Rodgers' Testimony: Some Concrete Proposals to Make the Semiconductor Industry More Competitive, July 23, 1991

Last Updated: 
May 24, 2012

Dr. Rodgers' Testimony: Some Concrete Proposals to Make the Semiconductor Industry More Competitive, July 23, 1991

T.J. Rodgers

"Some Concrete Proposals to Make the Semiconductor Industry More Competitive"
July 23, 1991
By: Dr. T.J. Rodgers
President and CEO
Cypress Semiconductor Corporation

THE DEBATE FROZEN IN TIME Two summers ago, I appeared before Congressman Brooks' House Judiciary subcommittee. Last summer I appeared before Senator Metzenbaum's Senate Judiciary subcommittee. Both of those hearings, like today's hearing, concerned the health of the U.S. semiconductor industry, the future of the antitrust laws, and the role of consortia in building American competitiveness. This is a debate that just will not go away. It is also a debate that seems frozen in time. In many respects, Washington has been conducting the same discussion about semiconductors for the last five years-even as some of the industry's most basic technologies and economics have experienced profound transformations. I would like to take a new approach this morning and begin with an overall perspective on the history and future of the U.S. semiconductor industry. After all, Congress cannot prescribe the right solutions if it is working on the wrong problems. I will then discuss how and why some of today's consortia help the semiconductor industry-and explain my continued opposition to the structure and policies of Sematech. Next, I will offer two new proposals to address the most serious competitive problem facing our industry-the damaging financial environment in which we operate. I will conclude with some general comments on the antitrust laws and the explosion of litigation that is damaging our industry.

THE SEMICONDUCTOR INDUSTRY IS HOLDING ITS GROUND Let me begin with a proposition seldom voiced on Capitol Hill by the CEO of a semiconductor company: The American chip industry is quite healthy right now and has been for the last five years. We have all heard the conventional view of the "crisis." According to industry-watcher Dataquest, the United States held 54% of the worldwide semiconductor market in 1982 while the Japanese held 34%. By 1989, the Japanese had increased their share to 52% while the U.S. share had slipped to 35%. Thus, in just seven years, there was a 37-point reversal in market share between Japan and the United States. Those figures are troubling, but they obscure as much as they illuminate. This dramatic market-share reversal is based on nominal revenues rather than volumes. In fact, the figures reflect currency exchange rates more than they do underlying competitive strength. If we use a constant yen-dollar exchange rate and restate the figures, the trend from 1982 through 1989 looks much less dire. Using the near-worst-case 1990 rate of 144 yen to the dollar, Japan held 47% of the world semiconductor market in 1982 while the United States held 43.5%. By 1989, the Japanese share was 51% while the U.S. share was 36%.

In other words, using an apples-to-apples comparison, it turns out that 25 points of the 37-point market-share reversal cited by the "doom-and-gloom" school reflect currency fluctuations rather than true competitiveness. These adjusted trends are nothing to celebrate, but they offer a less distorted picture than the nominal statistics.

I would argue that the United States and Japan have been in a state of dynamic equilibrium since 1987. During this period, Japan has gained no ground on the United States in semiconductors: Japan's worldwide market share has hovered within 2 points of 50% while the U.S. share has hovered within 2 points of 37%. The current equilibrium represents a balance between two ongoing battles. One battle is over manufacturing, which the Japanese have been winning but in which we are making a strong comeback. The other battle is over innovation, which the United States has been winning and continues to win.

The good news is that America's progress in manufacturing is moving faster than Japan's progress in innovation-it has proven easier for us to learn well-documented manufacturing techniques than for them to restructure their rigid economic system. Indeed, last year, the United States gained back semiconductor market share from the Japanese (based on either variable or constant yen-dollar accounting). I expect the United States to continue to gain ground over the next five years-optimistically, to win back as many as six market-share points; pessimistically, to hold what we have. America's slide began in 1979 and ended in 1985. Let's stop trying to rectify ancient history!

That said, we should question even the value of aggregate statistics. Using a single market-share figure to measure the health of the entire "U.S. semiconductor industry" is an exercise in dangerous over-simplification. The DRAM memory chip market, the largest market segment and the one most thoroughly dominated by the Japanese, greatly distorts the figures. DRAMs are a brutal commodity business in which even the most efficient Japanese producers are suffering from internal competition and recent Korean entry. On the other hand, microprocessors, the central engine of the computer revolution, remain the most technologically exciting and consistently profitable segment of the chip business-and their design and production are thoroughly dominated by American companies. Finally, when you look at present and future generations of specialized logic chips-math processors, digital signal processors, video processors-innovative American companies are leading the way at every turn. Viewed as a country or as a group of companies, the United States is positioned precisely where it should be-dominating the high-value, high-margin, innovation-driven parts of the business.

I certainly do not deny that our country has competitiveness problems. But the most severe problems are concentrated in the large, established, high-profile companies associated with the industry's early days-giant companies that will have to change or become dinosaurs. Indeed, last year, despite gains in the worldwide market share by the United States, six of America's eight largest chip companies lost money. These results are vivid evidence that the old guard companies that have been identified as "the U.S. semiconductor industry" can no longer be used as a true barometer of our industry.

The balance of power in semiconductors-as in the whole of electronics-is shifting away from many of the big, established companies toward smaller, nimble companies focused relentlessly on innovation and quality. Corporate size is no longer an intrinsic asset in competing with Japanese conglomerates. In many cases, it is a liability.


Think about microprocessors. Five years ago, Intel (then a $1 billion company) and Motorola (then a $3 billion company) talked publicly about how the Japanese threatened their leadership position in microprocessors. What really happened? Small, aggressive U.S. companies-not the Japanese-took market share. The most exciting developments in microprocessors are the new RISC architectures from Sun Microsystems and MIPS Computer Systems. My company is shipping a powerful new microprocessor based on Sun's SPARC architecture. Intel and Motorola have been unsuccessful in responding with their own RISC processors. Meanwhile, the Japanese giants are still trying to figure out how to respond to this blizzard of innovation.

Five years ago, if someone had asked me whether Cypress Semiconductor could enter the microprocessor market, I would have said we could not afford the massive resources such a move would require. But the game has changed dramatically. Ever-increasing computer power (essential to designing new chips) means that ever-smaller companies can play and win against the giants. With 22 employees and a total investment of only $7 million, Ross Technology, a tiny subsidiary of Cypress, brought to market a chip set more powerful than the Intel 80486. More than ever, the future of U.S. technological competitiveness rests with this country's entrepreneurs.

There is a new strategic logic for success in the semiconductor business: think small, think flexible, think efficient. Unfortunately, this is not the logic accepted by the old guard. These companies continue to rely on the incorrect premise of economies of massive scale and political protection-in manufacturing, in trade laws, and in wasteful litigation over intellectual property. Even the Japanese are due for a come-uppance. Japanese factories do very well with high-volume, medium-performance products engineered for relatively easy manufacturing. But our business is getting more fractionated, more specialized, and more complicated all the time. Japan's fabulous achievements as disciplined mass manufacturing will provide its companies with less and less of an advantage in the future.

Competing head-on with the Japanese does not require a few more government-subsidized "battleship" fabs and a handful of giant companies. Rather we need hundreds of small, flexible, innovative companies to wage slashing raids at Japan's weakest points. The lead article in the July-August 1991 issue of the Harvard Business Review makes just this point. As the authors state: "Dramatically changing technology and economics have made obsolete the traditional model of semiconductor production-a model built around high-volume wafer fabrication and vertical integration. Most large, established U.S. manufacturers have chosen to ignore or resist this unsettling new reality. They have suffered dearly as a result. But a new breed of companies and a few members of the establishment have embraced the industry's transformation. These companies have succeeded spectacularly in revenue growth, profits, and technology leadership."

To document their argument, the HBR authors combine the 1990 financial results of five innovative, young semiconductor companies: Altera, Chips & Technologies, Cirrus Logic, Xilinx, and Weitek. Together, these companies had 1990 revenues of $620 million, net income of $65 million, a market value of nearly $1.4 billion and a pretax return on equity of more than 26%. These results hardly portray an industry under siege.

Cypress's history reinforces the argument. We founded the company in 1983 and went public in 1986. We have never had a losing quarter since we first turned a profit in 1985. We now generate revenue at an annual rate of $300 million and consistently register pre-tax profit margins of more than 20% of revenues. We have succeeded precisely because we have not needed half-a-billion-dollar fabs or $100-million microprocessor development budgets. Yet even the toughest of our American competitors would give credit to our technology and microprocessor capabilities.

Last year Cypress ran its entire business with just two factories that together cost less than $80 million. Our wafer fab in San Jose, California, opened in 1984. It was completed in seven months from start of construction to the time we shipped our first revenue wafer. On a total investment of $40 million, this fab generates $90 million a year in revenue, runs almost 60 different products manufactured in 26 different processes, and does all of Cypress's R&D.

Our fab in Round Rock, Texas, opened in a used building in 1986. Eight months transpired from the time the wrecking ball wiped out the middle of the building (we left the walls standing) until the time Cypress Texas shipped its first revenue wafer. Our investment is $40 million. The factory is now generating revenue at a rate of $210 million per year and still runs at only 50% capacity. Both fabs manufacture products in the latest state-of-the-art, sub-micron technology.

My point is not to boast about Cypress's performance. Many of our entrepreneurial rivals-companies such as Integrated Device Technology and Micron Technology-are tough competitors. My point is to underscore the new realities of competition in the semiconductor industry. You cannot win the competitive battles of the 1990s with a strategic playbook written in the 1970s. The game has changed. Unfortunately, the giants that dominate the debate in Washington are still calling plays from yesterday's playbook. I cannot emphasize the point strongly enough: If we continue to shape government policy based on what is required to prop up sagging companies, we will make the wrong policy decisions.

Consider an analogy with the computer industry. Ten years ago, the leaders of the U.S. computer industry were IBM and the five "BUNCH" companies: Burroughs, Univac, NCR, Control Data and Honeywell. Today the BUNCH has been virtually eliminated from the competitive scene-but the U.S. computer industry, driven by the rise of new companies such as Apple, Compaq, Dell, and Sun Microsystems, has increased its global leadership position. A similar process of "creative destruction" is under way in the semiconductor industry. I do not believe the giant companies in our business will disappear, but the main engines of innovation, growth, and profitability will be the new, hungry, entrepreneurial players.


This entrepreneurial perspective on the health of the U.S. semiconductor industry relates directly to the issue of consortia-an issue with an enthusiastic following in Washington these days. Although I do not oppose the basic idea of consortia as a tool to improve U.S. competitiveness, I do not consider them a powerful tool to improve our competitiveness position. But structured correctly, with the right members and the right agenda, they can make a contribution.

Here is a quick review of how some consortia help our industry:

• The Semiconductor Research Corporation (SRC) makes some important contributions to education. It helps bright students work on relevant projects and thereby increases the supply of graduates with advanced degrees in fields relevant to our industry. The SRC operates without taxpayer subsidies.

• The Semiconductor Equipment and Manufacturing Institute (SEMI) has made some real contributions in areas such as technical standards. For example, its Standard Machine Interface (SMIF) allows manufacturers to use robots to move wafers from one machine to another in an environmentally clean production environment. SEMI, like the SRC, operates without taxpayer subsidies.

• The Microelectronics & Computer Technology Corporation (MCC) began eight years ago with delusions of grandeur-massive budgets and annual membership fees of $1.5 million. Over the last few years its structure and research agenda have evolved to serve its market well in several areas of computer research, including the critical field of interconnection. Two weeks ago Cypress became an associate member of MCC-a clear sign of our support. Like SRC and SEMI, MCC operates without taxpayer subsidies.


Granting the contributions of these three consortia, Sematech remains the wrong answer to the wrong problem. It equates the health of our industry with the fortunes of a handful of giant companies, some of which are models we should avoid following-the companies that lost market share to the Japanese to begin with.

In the interest of brevity, let me offer just three specific criticisms of Sematech's structure and policies.

1. Sematech's membership policies discriminate against many of the companies that are leading America's comeback. Sematech's advertised dues of one percent of sales actually have a $1 million per year minimum. That means a $20 million company pays dues equal to 5% of its sales-perhaps all of its profits and five times the dues for a company with revenues of more than $100 million. What's more, Sematech's dues top out at $15 million per year, which means that companies with sales of more than $1.5 billion receive a dues discount. Thus, a Sematech member with $3 billion in sales pays dues equal to one-half of one percent of its sales-or one-tenth the rate of a $20 million company.

Sematech's dues structure deliberately discriminates against America's small semiconductor entrepreneurs. Is it any wonder that only 14 of America's hundreds of chip companies have chosen to avail themselves of the federal government's generosity? Or that 12 of Sematech's members are giants with annual sales of more than $1 billion? Ultimately, Sematech looks less like an "industry consortium" than an exclusive Corporate Country Club.
2. Sematech policies harm non-members-even American non-members. Sematech uses taxpayer money to aid its old-guard members at the expense of innovative new companies. There is no more troubling example of this phenomenon than the equipment holdback contracts that Sematech has imposed on some of the semiconductor-equipment producers it supports.

Let me share a brief story. A year ago Cypress attempted to buy a wafer polishing machine from a company called Westech in Phoenix, Arizona. We sent a group of engineers to meet with the company. They reported back to me that something seemed strange. They felt they were not getting access to Westech's most advanced equipment. Ordinarily, when a small equipment company like Westech gets a chance to do business with a new customer, it leaps at the chance to show its newest and best equipment. I called the president, but the issue was never resolved.

Recently, I got my answer. Sematech's development contracts with Westech on the new polishing machine require contractually that Westech withhold this equipment "for a period of one year from the time of normal introduction" from all but Sematech's members. This holdback restriction on taxpayer-funded equipment is outrageous-especially in light of Sematech's recent public-relations maneuvers. A month ago the Semiconductor Industry Association denounced a group of Japanese semiconductor equipment companies for withholding their most advanced production equipment from American companies-charges that have withered under scrutiny. Now we discover that Sematech itself is withholding from American companies equipment financed by American taxpayers!

3. Sematech uses kickback agreements to reduce the financial contributions of its members-and again tilts the playing field against successful U.S. rivals. Recently, Sematech contracted with one of its own member companies, Intel, to purchase and test equipment that Intel probably would have acquired itself. Under the agreement, Intel received-for free-a $1.5 million AMP-5000 wafer-etching machine. It also received a grant of $700,00 to install the etcher and an additional $1.2 million to evaluate it. In effect, this kickback agreement amounts to a 23% reduction in Intel's annual Sematech dues-a government subsidy hardly needed by a company that earned $650 million last year.

My company also evaluated the AMP-5000 etcher. Well before Intel acquired its etcher, we bought the second AMP-5000 sold in the United States and only the nineteenth sold in the world. We paid the full $1.5 million for the equipment. We spent $500,000 to install it. It took us two years-and about $5 million-to work out the bugs so the etcher would operate at the quality level demanded. Cypress paid dearly to make it work. Intel received the same equipment for free.

Unless Sematech's policies and practices receive careful scrutiny, they will inspire similar programs-like U.S. Memories, the ill-advised next step. Fortunately, U.S. Memories never took off because the industry saw it as a poor investment. Today Congress is being asked to evaluate another proposal, Microtech 2000, in which the U.S. government would put up $1 billion to create yet another Amtrak of the semiconductor industry.


Assuming the dynamics of our industry have changed and that high-cost government programs are ineffective, how can the U.S. government support not just the giants but also the nimble, entrepreneurial companies that are making gains in world markets?

Most importantly, government should not involve itself directly in the details of our business. The semiconductor industry is far too complicated for any one agency (or even a group of companies like Sematech) to make the right decisions consistently. Government support must be institutional-designed to create a rising tide that lifts all boats. The proposals that follow reflect this "generic" approach.

One critical factor to the success of the semiconductor industry has inspired much rhetoric, but generated little or no real action in Washington. The members of Sematech often warn (with justification) of the erosion of our "manufacturing infrastructure"-the chain of production from raw silicon through capital equipment, wafer fabrication, assembly and test. One major cause of the erosion of our industry's manufacturing infrastructure is the damaging state of America's financial infrastructure-the scarce supply of reasonably priced capital that successful companies need to build their manufacturing muscle. The federal government can help us achieve two critical objectives: to lower the cost of capital and to extend the time horizons of our investors.

The ambiguous and over-used term "cost of capital" means different things to different people. Troubled companies often use it to explain away bad management or poor investments. In our industry, however, the capital problem is real and severe. The Council on Competitiveness reports that Japan and Germany invest on the order of 3% of their Gross Domestic Product (GDP) in civilian R&D. The United States invests less than 2% of its GDP in civilian R&D. Thus, our two leading rivals out-invest us by more than 50%-a huge handicap in high-technology competition.

The single most damaging "tax" the federal government imposes on our industry is deficit spending. We all understand the vicious cycle in which massive government borrowing drives up the cost of corporate borrowing-with a major impact on a capital-intensive business like semiconductors. During the early 1980s, a period of steep market-share erosion for the U.S. chip industry, American companies were paying 12% interest rates, while the Japanese were paying 4%. The average cost of the silicon, plastic, copper, gold, and all other raw materials in a Cypress chip is only about 28 cents. Our capital costs-the money we borrow to build our factories-average about 56 cents per chip. In other words, our manufacturing costs are much more heavily influenced by the cost of money than by the cost of silicon. Doubling the cost of the raw silicon wafers we buy would have less of an impact on our bottom line than raising the interest rate we pay by only two percentage points.

Government borrowing also hurts our access to the equity markets. The average price-earnings ratio for a high-tech issue on the New York Stock Exchange is about 15. In Tokyo it is about 60. Therefore, U.S. companies must sell four times as much stock as comparable Japanese companies to raise the same amount of money. Consider the example of the NMB, a third-tier Japanese semiconductor company that is a subsidiary of a Japanese ball-bearing company. NMB manufactures well, but compared to Cypress it has a narrow range of products, no marketing, little process development, no product development, and almost no sales capability. Yet NMB raised $472 million in a single public offering two years ago. Cypress-a well-funded company by American standards-needed six stock offerings to raise just $150 million. How can our entrepreneurs compete over the long term when highly regarded U.S. companies struggle for capital, while third-tier Japanese companies effortlessly raise huge sums?

The short-term mentality of our financial markets is real, hurts American manufacturing, and is encouraged by our current tax system. A week ago we released financial results for the second quarter over the business wire and through a direct teleconference presentation. The results were very encouraging: record revenues, record earnings and good news on orders. I wanted Cypress employees to see first-hand how Wall Street thinks-in particular, how uninterested the markets would be in our strategy and achievements as opposed to our short-term outlook. One thousand Cypress employees gathered in an auditorium to listen to our teleconference report. Participating in the call were the representatives of about 65 of our major investors, accounting for more than half of our outstanding shares.

These investors spent literally 30 seconds talking about what we had achieved in the quarter. They spent only a little more time talking about our long-term strategy. Nearly all the questions concerned what would happen over the next 89 1/2 days-until our next quarterly results would appear: Did we have any yield problems? Would our RISC microprocessor continue to sell at current rates? Would we feel the usual summer lull in semiconductor orders? Our employees were taken aback by the short-term view of our investors, but this is the financial environment in which we must operate every day.

Most of our investors are smart and well meaning. They are forced into the quarterly mentality so as not to suffer a disadvantage relative to other investors. Indeed, when we finished the conference call, our share price shot up almost 10%. We checked with our specialist on the New York Stock Exchange. He explained that the stock rose because of "short covering" by two groups of speculators who had sold our shares in advance of the teleconference, in essence betting that we would report bad news.


The three proposals that follow are designed to increase the availability of capital and to reward long-term investors.

1. Restructure the capital gains tax to punish short-term speculation and to encourage long-term investment. (Please note: I did not say lower the capital gains tax.) I have always been a public advocate of reducing taxes on capital gains, but I have always had certain economic and political reservations about that move. Roughly two-thirds of our investors are institutions. Since most of them are tax-exempt, reducing the tax on capital gains will create few incentives for them to invest more in our industry. As for politics, there remains the almost-insurmountable belief in Washington that any reduction in capital gains is a "tax break for the rich." In reality, a tax break that steers the money of the rich into America's manufacturing infrastructure is an excellent policy. Unfortunately, the political reality is that "tax breaks for the rich" will not pass.

Many people oppose lowering the tax on capital gains, but everyone should be interested in changing the behavior of investors and extending their time horizons. We should reform our financial infrastructure to encourage people to invest responsibly in America's manufacturing infrastructure. We need an environment in which investors have strong incentives to invest for the long-term-and in which American companies finally have the freedom to operate with the long view in mind.

This restructuring proposal is hardly the first to look at capital gains from a time-horizons perspective. For example, Representative Tom Campbell has introduced HR 2523. Under his plan, capital gains realized within two years would be taxed at current rates. Any gains realized between two and five years would be taxed a one half the current rates. Gains on assets held for more than five years would be completely exempt from taxation.

The Campbell proposal is clearly an improvement over the status quo. However, I would modify it to overcome the "tax break for the rich" argument. First we should increase the capital gains rate on assets held for less than six months by imposing a 50% surcharge over the nominal rate. This surcharge would be a stiff disincentive to wasteful speculation. It would allow us to "tax the rich"-at least the rich who play hunches, trade on rumors, and churn investments to the detriment of the American economy. Assets held between six months and three years would be taxed at the current rate. Capital-gains taxes on assets held for more than three years would be totally eliminated.

Think of how radically these reforms would change investor perspectives. A short-term speculator would pay a huge tax on a capital gain relative to someone willing to invest over a three-year horizon. Somehow I think our quarterly conference calls would take on a different character if these provisions were in place. The questions would gravitate toward determining whether our company was positioned, managed, and funded to be a leader of the next three years, rather than the next quarter. This reform would also lure the "little guy" back into the stock market. Small investors willing to make patient equity investments with their savings or retirement funds would enjoy returns free of taxes.

2. Abolish restrictions in the Glass-Steagall Act of 1933 and the Bank Holding Company Act of 1956 that forbid banks to own equity in industrial companies and limit bank holding companies to no more than 5% of an industrial company's total shares. The regulations governing our banking system have created a "missing link" in the financing sources available to fast-growing technology companies, especially manufacturing companies. These regulations may have been appropriate for the realities of the U.S. economy after the Depression, but today they do grave harm.

Venture capital is America's high-technology saving grace. There is no shortage of venture capital to get smart ideas off the ground. No good engineer in Silicon Valley will be refused his or her first $10 million to prove the feasibility of a new technology or product. The problem comes at the next stage, when the company needs its second or third $10 million to build a plant and acquire real manufacturing muscle. As a venture capitalist myself and as an outside director with other companies, I have seen this phenomenon first-hand. This "mezzanine" financing stage is where so many good technology companies run out of capital-or run into the arms of cash-rich Japanese corporations that acquire valuable American technology for no other cost than investing. My own company had to raise money in Europe, and I serve on the board of a company that sought Japanese funds after several futile efforts in the United States.

Venture capitalists cannot fund the large mezzanine investments that high-technology manufacturers demand. The two-to-one return associated with these risky investments is far below the return venture capitalists require to stay in business. Mezzanine funding is the "missing link" I discussed earlier. In fact, government subsidies for programs such as Sematech or the ill-fated Consumer Electronics Capital Corporation are essentially designed to bridge that financing gap. But why turn to taxpayer subsidies when government could help create a mezzanine funding industry every bit as successful as the venture capital industry is in seed funding?

Commercial banks could play a much larger role in funding new semiconductor companies. Today banks cannot lend to start-up semiconductor companies because they cannot feasibly charge the 20%-50% interest rates necessary to compensate for the risks of bankruptcy. Ordinarily, high-risk loans are accompanied by warrants for the company's stock to increase the rate of return on the loan while keeping the interest rate reasonable.

For the last decade such mezzanine financing has been an important source of funds for the leveraged buyout movement-insurance companies and other institutional investors have provided billions of dollars of convertible debt, bonds with warrants, and other debt-equity hybrids. Yet we do not allow our banks to offer similar kinds of financing to America's high-technology community.

3. Congress should make permanent the R&D tax credit. The "temporary" R&D tax credit has been in place since 1982. The semiconductor industry now takes it for granted and makes its investment plans accordingly. Making the tax credit permanent would not actually change the status quo. Reducing or eliminating the credit would be devastating and would exacerbate the industry's financial problems.


We continue to complain about "lawyers" and fail to realize that they only do what we tell them to do in the environment we have created for them. And while the most frivolous issues are litigated at great expense, our basic business fairness laws-the Clayton and Sherman Antitrust Acts-are being systematically undermined.
1. Congress should not gut the antitrust laws as they apply to production consortia. In 1984, the National Cooperative Research Act reduced antitrust penalties on research consortia from treble damages to single damages. Now Congress is being asked to consider the National Cooperative Production Act, which would reduce antitrust penalties for manufacturing consortia. The logic of this legislation is elusive: the penalty reductions are relevant only if consortia are convicted of violating antitrust laws. Do any of us really believe that America's competitive future rests with reducing fines on corporate lawbreakers? Moreover, what is the logic of demanding "fair play" and "fair trade" with our foreign competitors, while making it less costly for the giants to engage in unfair practices within the United States?

2. Congress should create disincentives for frivolous and harassing lawsuits filed under the banner of "protecting intellectual property." Many large companies in our industry are in the midst of a litigation frenzy. For example, AMD and Intel have probably spent at least $1 million a month during their three-year legal war over rights to the 80386 microprocessor. With that same amount of money, AMD and Intel could have funded Cypress and each owned half of what is now an $800 million asset! Texas Instruments has a stated policy to sue dozens of companies to generate revenues. It actually assigns particular patents to individual lawyers who run profit-and-loss centers designed to maximize earnings off those patents with massive legal activity. Meanwhile, Texas Instruments' semiconductor operations continue to run with heavy losses.

Cypress's own legal history illustrates how absurd these skirmishes have become. When we entered the RISC microprocessor market we needed a technology manager. After an introduction from a venture capitalist, we hired a well-known computer expert and played venture capitalist for his RISC company. When he and the four other founders of Ross Technology left Motorola they were immediately sued-both as a corporation and individually. All of Ross's employees were directly threatened with not being able to work for another company on microprocessors-the foundation of their careers.

The Motorola lawsuit was a typical warning shot fired across our bow. We settled the suit in the usual way, with no money changing hands and an agreement not to disclose the terms of the settlement. Then came a truly absurd development. Ross technology hired one engineer from Advanced Micro Devices. AMD sued us and him personally. Fortunately, it turned out that AMD's lawyers made an embarrassing blunder that I call "Xerox litigation." The lawyers had simply taken the Motorola lawsuit, which named five defendants, removed their names, inserted the name of the engineer hired from AMD, and then re-filed the lawsuit with minimal cost. AMD literally copied the Motorola lawsuit-and did not even have the plagiaristic precision to change the plural to the singular! Naturally, they did not prevail and we kept our engineer. Unfortunately, these stories are typical of the legal realities of life in our business. In more than 25 actual or threatened lawsuits, we have never lost a case or paid a cent to a plaintiff. Are Cypress's lawyers that good or are we being harassed with nuisance litigation?

Earlier this year semiconductor litigation took a new and bizarre twist. Sematech actually sued the government of Travis County, Texas, in an effort to exempt itself from the local school tax by claiming that it was a "charity." Sematech hired one of the leading law firms in Houston and paid hundreds of dollars per hour for its legal ammunition. Ultimately, Sematech withdrew the suit just as it was about to go to trial and agreed to pay its fair share of the schools and roads its employees use. The situation is absurd: The federal government paying to sue a local government to reduce funding for education. Is this really how Congress intended to "assist" the semiconductor industry?

I would like to offer two suggestions on litigation. First, Congress should be wary of high-minded talk about "protecting intellectual property." Such talk is usually a codeword for more harassing lawsuits. Second, Congress should pass legislation that requires plaintiffs to pay the legal fees of defendants in failed intellectual-property lawsuits. It may not be possible to end the litigation wars, but it is possible to make frivolous skirmishes less appealing to the aggressors' profit-and-loss statement.


The United States is the best country in the world in which to start a semiconductor company. There are problems, to be sure, but they can be addressed by common-sense improvements to our system, a few of which have been suggested here. The U.S. semiconductor industry's period of market-share loss ended years ago. We gained back share last year. Every failing CEO who uses America as an excuse supplies his or her entire organization a clear leadership message that failure is tolerable or even expected. No American company can afford to hear that message.