Dr. Wessner introduced the third panel to “what I think has so far been a remarkably rich discussion of some of the challenges facing partnerships around the world.” He welcomed the European colleagues who like those from Asia had made substantial efforts to attend and introduced moderator Michael Borrus, who was a member of the steering committee under which this study was being carried out.
The Petkevich Group, LLC
Mr. Borrus said that the panel might prove to be “a bit of a surprise to some of you” in light of the rapid strengthening of the European presence in semiconductors in recent years. A decade ago, he said, many people were focused on the United States and Japan in assessing the struggle for leadership in semiconductors and did not anticipate a significant European presence. Recent developments, he said, had proven this belief to be misguided. Indeed, he said, Europe has reemerged as a very significant player in information technology and a dominant one in some sectors of communications services and equipment, notably wireless, embedded industrial electronics, and certain parts of the semiconductor industry.
He described a “complex constellation of reasons behind Europe’s resurgence in information technologies,” including the slow but persistent progress to a truly common market, the development of the euro, and an increasingly more integrated European financial services marketplace. The latter has meant greater liquidity and a wider range of financial institutions, including venture capital and growth-issue stock exchanges like the Neue Markt in Germany. Europe’s financial marketplace can now fund innovation in technology industries, he said, in ways it could not a decade ago. Another reason for Europe’s emergence is the choice of common standards in wireless communications, which has helped to boost European manufacturers to the leading edge. And finally, the European cooperative programs, such as ESPRIT and RACE, which were formed more than a decade previously, had been important in fostering collaboration among European companies and between Europe and the rest of the world. To discuss the first of these programs he introduced Dr. Jürgen Knorr, chairman of MEDEA, who had spent many years at Siemens in a variety of senior management roles.
THE MEDEA PROGRAM
Micro-Electronics Development for European Applications (MEDEA)
Dr. Knorr, who began by characterizing himself as “one of those European dwarfs who would like to play a small role in semiconductors,” confessed that it is “very strange” to be leading a program to support the multinational semiconductor industry. The tradition had always been to support one’s own industry and nation, but now his organization was saying “no.” The semiconductor industry has become global and MEDEA’s objectives have to be shaped accordingly.
Some Special Difficulties for Europe
He then said that the national features within Europe do cause some special difficulties. Europe, he noted, is neither a nation nor a republic. He observed the political economy of this federative system to be the “analog of a semicustom integrated circuit (IC), a non-optimized, coordinated conglomerate of different functions which try to shoot for one target—a target, though, that is interpreted in different ways.” French English is different from German English, he said, and both differ from what the Italians may understand when using the same words. As a result this reality caused the Europeans to think together about the best way to close the gap between Japan and Europe in manufacturing, about intellectual property creation in the United States, and about what is possible politically and financially.
He also noted that Europe’s background is different from that of the United States or Japan, Korea or Taiwan. The differences show up at the World Semiconductor Council, where long-term perspectives vary and are worth reviewing for a moment.
The Context of Competition
Entering a New Cycle
The development of industries forming our society, he said, is relatively well described by the so-called Kondratiev cycle.22 Earlier in this century, resources such as steel, oil, and electricity defined engineering technologies and their application in railways, bridges, skyscrapers, and automobiles. Companies focused mainly on the competition in the context of specific regions and of companies in those regions. These different regions used different standards and protected their companies and their employment.
After the 1970s, he said, the so-called fifth Kondratiev cycle began. The resources being used were now silicon technology, semiconductors, and design software. The applications were to the military, consumer electronics, data processing, data communication, and especially mobile communication. The competition was increasingly determined by intellectual property, not by location or hardware, as it was in the past. IP (intellectual property) includes the experience needed to make something using the knowledge of engineering, physics, and chemistry. This is transferred through brains, through human resources. Human brains in today’s economy are the most flexible resource and they can be applied globally. The forces of globalization, along with World Trade Organization rules, have pushed away, first, the political barriers and, second, the economic barriers.
In this way, said Dr. Knorr, policy entered the electronics industry, especially the microelectronics industry. Global competition developed between nations, such as the United States and Japan, and led to the famous U.S.-Japan trade agreement.23 The Europeans entered global competition not as a single European nation but as a patchwork of nations.
Competition Is Now Between Companies
That, he said, is the background to why we are here today. It is important to understand that the real competition is no longer among nations but among multinational companies, including Siemens, Philips, Motorola, and others. Globally competitive companies need to form transborder joint ventures, corporations, and mergers. These activities, he said, will determine the future structure of the industry more than programs such as MEDEA.
However, the necessary infrastructure in education, human resources, research, and technology can and should be driven by global competition, which in turn is stimulated and supported by national programs. So competitive structures are an amalgamation of both, and sometimes contradictory because governments try to spend taxpayer money only on activities that benefit their nation.
Responding to the Value of Information Technology
An important force behind MEDEA is that in all the countries of Europe the politicians and parliamentarians who write budgets have learned that the driving force of industry in the next 15 years will be electrical engineering and electronics. He noted data from the U.S. Semiconductor Industry Association illustrating that semiconductors add more value to the U.S. economy than any other manufacturing industry. He also showed data from Germany showing that information technology creates more jobs than any other employment sector. Europe has a special need to cooperate, he said, because most of its countries are too small to compete alone against regions such as the United States and the Asian Pacific. Moreover, he noted, they must cooperate specifically in IT, which will offer the most jobs and industrial strength—as long as the countries remain globally competitive.
This shows up in the regional trade balances. During 1992-1997 Europe and the United States consistently had a negative balance of trade with Japan, while Japan and other Southeast Asian nations individually experienced positive trade balances with Europe and the United States. One of Europe’s disadvantages, he said, is that it is a patchwork of nations with different cultures, different languages, and different national policies. To develop synergies through cooperation is more difficult than it is in the United States. Even different national behaviors add problems that should be solved before claiming that Europe is a unified market.
The Benefits of Cooperation
The development of the common wireless standard GSM is a success, he said, not because it is the most advanced technical system but because at least a few big countries have agreed upon it. Cooperation can help create a critical mass, share risks, reduce the costs per partner, share know-how, and shorten lead
times. If countries work together, even if they have spillovers, sharing is the most effective way for those who, for whatever reasons, cannot do it alone . For these reasons and to stimulate employment the European Commission and the national governments stimulate transborder, cooperative R&D.
He reviewed the microelectronics support programs in all the major manufacturing regions, including the United States, Western Europe, Japan, Taiwan, and Korea. He then turned to MEDEA, an industry-initiated and industry-driven program supported by the national governments of 12 participating countries.24 Its objectives are to stimulate transborder R&D cooperation and to strengthen the global competitiveness of the European microelectronics industry in technology and in future applications.
Focusing on Core Competences
In semiconductors, he said, there are three core competences for technology: design techniques and libraries, CMOS-based technology platforms, and manufacturing technologies. There are also three core competences for applications: multimedia technologies, communications technologies, and automobiles and traffic applications, or automotive electronics. The outcomes of these six core competences are system-on-a-chip domain and IC technologies domain capabilities. These are intended for the globally competitive industry segments: ICT companies, semiconductor companies, and equipment and materials companies.
MEDEA was a 4-year program that started in 1997 and terminated at the end of December 2000. The total cost was about 2 billion euros for 45 projects. These projects were distributed among 25 major partners and a total of 105 smaller partners, including SMEs, universities, and institutes. The 12 participating companies were Austria, Belgium, Switzerland, Germany, Ireland, France, Finland, Greece, Italy, the Netherlands, the United Kingdom, and Sweden. The United Kingdom made the smallest contribution, having had its own policy.
Criteria for Inclusion in MEDEA
The main criteria for project selection, as defined by industry, include the following:
Innovation in fundamental research, industrial research, and pre-competitive development;
Existing European R&D capacity—a condition of the national governments for using taxpayers’ money;
Competences of the partners;
Complementary strengths of the consortia;
World market potential; and
A leveraging effect on employment.
A board manages MEDEA, and a support group oversees steering groups for both technology and applications. The technologies include CMOS technology and manufacturing technology. The applications include multimedia technologies and communications, both of which share a design function, and automotive applications.
Four Principal Results
After its 4 years of supporting cooperative R&D between European microelectronics companies, summarized Dr. Knorr, MEDEA has achieved four principal results.
Above-critical-mass collaborative projects. Each project involved a minimum of two countries and two companies. He also cited collaboration with national programs in France, Germany, Italy, Belgium, and the Netherlands. He noted that in the future there may be a 300-mm program in France with Philips and STMicroelectronics. There was also a German program with Infineon and Motorola that was first in manufacturing 300 mm. He called this effort a parallel, not overlapping, program.
Better understanding between semiconductor suppliers and system houses. This was expedited by choosing English as the common language.
Better understanding about where to focus resources.
Closer cooperation along the whole “food chain.” Dr. Knorr remarked that this result was acknowledged by both industry and governments, even though the consortium had not created quantitative criteria to evaluate outcomes. The main objective, especially between different countries in a particular market, was to learn to cooperate insofar as competitors can do so. The second important result was that the whole society gained awareness that collaboration is desirable and a productive use of taxpayer money, and that it also tends to increase employment.
Moving to MEDEA Plus
As a consequence of these results the participants decided to initiate a MEDEA Plus program in 2001 under the guideline, “system innovation on silicon.” The objective of this sequel program goes beyond system-on-a-chip to include system-in-a-package. The overall mission of MEDEA Plus includes the following goals:
Address the main microelectronics challenges in silicon application platforms and enabling technologies following the international technology roadmap. He expressed hope that the consortium would be effective in drawing together the large amount of decentralized, sub-critical expertise needed in many institutes.
Focus on key priorities and build on European industrial strengths to improve Europe’s competitiveness in system innovation.
Adapt to the fast-moving technical and marketing environment and expedite global cooperation. This objective is not normally in line with national government support, he said, but he expressed satisfaction at being able to forge international cooperation with IBM and Motorola.
Offer opportunities for participation by SMEs, institutes, universities, and large groups.
He closed by noting that the CEOs of international companies might have to withstand pressure emanating from their own national governments for joining MEDEA Plus. “Otherwise,” he said, “life would be so easy.”
Dr. Wessner asked about the level of funding from the MEDEA program for actual work and about contributions from the larger member nations. Dr. Knorr answered that cost of the four-year program for companies and institutions participating in projects was roughly 2 billion euros, or 500 million euros per year. National governments supported the parts of the project that were done in their countries. The amount of funding for near-to-market research was less than 30 percent; funding for more basic research was about 50 percent. In the national programs, he said, it is sometimes harder to differentiate semiconductor research within ICT. He estimated that in France, Germany, Italy, and the Netherlands the public support ranges between 20 and 50 million euros per year. In other countries the amounts would be smaller and harder to estimate.
GOVERNMENT-INDUSTRY PARTNERSHIPS IN EUROPE I
Submicron Semiconductor Technologies
European Strategies in the Global Semiconductor Industry
Dr. Draheim, CEO of Philips Semiconductor, said he would explain the reason his company chose to join the submicron consortium, “even though cooperation in Europe is not easy.” He began by illustrating some of the differences between the world’s regions with respect to technology.
In Asia memory drives microelectronic technology.
In the United States processors and memory drive microelectronic technology.
In Europe application-oriented circuits drive semiconductor technology.
Europe Has Diversified Its Semiconductor Business
Dr. Draheim said Europe cannot focus on one particular area. Instead, Europe in recent years has diversified and been able to take a leading position in several areas: communications, automotive electronics, smart cards, and multimedia. For applications in these areas, system innovations involving silicon are the driving force, requiring embedded technologies. As examples he said that communications and cellular radio need RF (radio frequency) integration, portable systems need flash integration, smart cards need non-volatile memory, multimedia need DRAM integration, and automobiles need high voltage. Specifically, automotive applications need more than 60 volts and high temperatures of about 200°C—a difficult challenge but not essential just yet. For telecommunications, RF integration is needed at high gigahertz and low power so that a device can run for weeks instead of just hours. For smart cards one needs non-volatile embedded memory, while multimedia demands chips for real-time processing.
Drawing on Particular Strengths
Therefore, he said, even though Europe and the United States are following the same roadmap, Europe has some particular strengths in the applications-oriented fields. In particular, Philips expects major breakthroughs in four fields: portable infotainment, third-generation mobile communication, home networks, and enhanced digital TV. He elaborated on one example, the home network. To develop such systems requires many cooperating partners: research institutes, equipment suppliers, semiconductor manufacturers, system houses, SMEs, business groups, and technical centers. For home networks one needs microelectronic components and software modules. The particular components include high-band-width internal networks, a high-bandwidth access network to the Internet, storage systems, and enhanced image sensors. For software the needs include personalized services, multimedia Internet services, and easy information management. All of these platforms have to be concurrently developed. He pointed out how rapidly the placement of microcontrollers is growing, especially in home applications, including coffee makers, telephones, garage-door openers, microwaves, sewing machines, and cameras. The total number of ICs in the home is expected to grow from about 60 in 1990 to 300 in 2005.25 In the office the total number of ICs for 2005 is expected to reach 110 and, in the automobile, about 95.
Why in the home? He made the point that all appliances are not only going digital, but they are also being connected. “The home will become so intelligent,” he said, “that I fear I will not be intelligent enough to live in my own home.”
The Necessity of Cooperation
As systems become more complex, we need more cooperation. He said that Philips participates very actively in MEDEA, which is oriented toward hardware development, and with ITEA (Information Technology for European Applications). These are complementary programs that are “absolutely important for our future development.” He listed two primary reasons why cooperation is necessary.
For complex systems, the increasing cost of technology development calls for horizontal cooperation to share costs, reduce risks, and shorten time to market.
The increasing costs of complex-systems engineering calls for re-use of intellectual property and for more vertical and horizontal integration with the customer.
A company cannot, he concluded, do everything on its own. As an example of the cooperative development of a technology, he cited SiGe-BiCMOS (silicon germanium Bi-polar Complementary Metal Oxide Semiconductor) for RF applications. Cooperation has begun with nine partners, including telecom companies, semiconductor companies, institutes, and universities. He said that the next big cooperative challenge will be to develop systems-on-a-chip with the goal of achieving the same functionality in one-fiftieth the space. To meet this challenge, he said, Philips plans to cooperate in both MEDEA Plus and ITEA.
Funding for ITEA
A questioner asked if ITEA is financed by the European Union or by the private sector. Dr. Draheim answered that ITEA has the same financing structure as MEDEA, which is supported by national governments. The difference is that MEDEA is oriented toward hardware and ITEA is oriented toward software. The budget is not based on specific requirements from each company. Money is raised on a per-project basis from the different countries. He did not know ITEA’s budget for its first, startup year, but said it will grow to more or less the same size as MEDEA for its 4-year planned lifetime.
Dr. Wessner asked if there is any linkage with the European Union framework program. Dr. Draheim said there is no official linkage but rather close cooperation and information exchange with the IST program of the European Union.
GOVERNMENT-INDUSTRY PARTNERSHIPS IN EUROPE II
Dr. Beinvogl began by noting that with the internationalization of SEMATECH the three major information technology players in Europe are all now members of SEMATECH. He said the three are not only financial contributors to International SEMATECH but also significant technical contributors. He gave two examples. The first was 300-mm technology, which is a key agenda point in International SEMATECH. He suggested that Europe has achieved a world leadership position in 300 mm that in turn benefits International SEMATECH and its members. Another example, he said, was that the IMEC institute in Belgium, a world leader in cooperative research, has already achieved very close cooperation with International SEMATECH on one major project.
Strategic Challenges in the Semiconductor Industry
He then turned to some aspects of government-industry R&D partnerships in Europe. He showed a chart illustrating the time required for various technologies to reach one million users, ranging from 20 years for black-and-white TV to fewer than four years for the analog cellular phone. “The message is,” he said, “that everything in this world seems to be accelerating.” He said that the ITRS (International Technology Roadmap for Semiconductors) has been moved ahead by nearly a year over the last couple of years.26 In discussing government-industry partnerships, he said, we have to take this speed into account and be prepared to accelerate.
He emphasized the outstanding importance of semiconductors, which have “huge leverage” on all kinds of electronic equipment, creating more value added and contributing to faster growth and many new jobs—especially highly skilled jobs. The downside is the scarcity of highly skilled specialists. Based on these facts, he said, microelectronics is considered to be a major strategic industry by national governments worldwide.
Partnerships as a Response to Challenges
Despite continuing competition between multinational companies, regions, and even countries, there is an increasing tendency to collaborate and to form partnerships. These partnerships, especially in Europe and particularly in Germany, can be divided into three major domains.
For a description of the ITRS and its goals see its Web site at <http://public.itrs.net/>.
Industrial cooperation with publicly funded R&D institutions, such as a university (e.g., IMEC and the Fraunhofer).
Public funding of industrial research, which can go up to 50 percent in Europe generally. A complex set of rules defines the actual percentage, which can be considerably less than 50.
In most cases public funding is combined with funding from other sources, such as institutes or companies from different countries.
In Europe there is no agency comparable to the Defense Advanced Research Projects Agency for microelectronics—meaning no non-civilian support from governments. In addition, there is no indirect public funding for R&D, only fiscal incentives. He said that the network of contracts of the corporate research group is complex, that research conflicts are common, and that it is hard to maintain an overview of so many activities and relationships.
A “Mental Gap” in Technology Transfer
He showed a depiction of the various technology transfer mechanisms in Europe, the United States, and Japan. In Europe he noted a gap between non-industrial research and industrial research and labeled this a “mental gap” in order to indicate that it was more imagined than real. He said that IMEC, GRESSI (Grenoble Sub-micron Silicon Initiative) and FhGmbh are good examples of how the mental gap can hamper cooperation with industry.
Dr. Beinvogl said that the overall funding structure in Europe is basically top-down, from the commission level through framework programs, which defines the contents of funding for individual projects and funding. By contrast, the national authorities, which provide most of the actual funding, cooperate at the European level through a formal certification process and by labeling national, bi-national, or multinational products with the so-called EUREKA label (European Network for Industrial R&D). This label helps to obtain funding.
Successes and Lessons
A Major Success Story
He emphasized as one “full-blown success story” the joint venture between Infineon and Motorola, called the SC300, for semiconductor 300, which had ended about three weeks before the workshop. This joint venture was formed by two companies whose ambition was to be poised at the leading edge in the transition to the next wafer size. The company, set up in Dresden, had met its goals without encountering any major obstacles. The first fab was already fully constructed and was due to start operation in the second quarter of 2001, with a
capacity of 5,000 wafer stocks per meter at 300 mm, starting with DRAM manufacturing.
He summarized another major collaboration between Infineon and Motorola, which had cumulatively spent $300 million in the previous 2 1/2 years. Both the German federal government and the state government of Saxony agreed that the project had excellent potential to create jobs and they provided substantial funding.
For comparison he mentioned a decade-long collaboration with IBM, and to a large extent also with Toshiba, to develop DRAM technology in New York state. They have expanded this partnership to add a logic program. The cumulative cost to all partners has been $1.1 billion.
Partnerships Have to Be Global
Finally, he said that these kinds of partnerships now have to be global. He mentioned Infineon partnerships with Intel, which was a strategic investor in Infineon at the initial public offering. There is also a three-way partnership now among Infineon, IBM, and UMC (United Microelectronics Co. of Taiwan) to develop and manufacture advanced logic processes. The company also participates in a newly founded advanced DRAM consortium which brings together the major companies in the field, along with Intel.
Dr. Beinvogl closed by saying that “we are very deeply convinced that there is no way around globalization in our business today.” Everyone has to find and fund their own best partnerships—those that are consistent with the globalization of their development efforts and of their business.
A Decrease in the Number of Engineers
In response to a question about the sufficiency of human resources in Germany, Dr. Beinvogl described a major decrease in the supply of engineering graduates since the early 1990s. “This is absolutely dramatic,” he said, “it’s not just a little effect.” He said Infineon is now forced to look all over the world for skilled people. He said that about eight years ago young people were discouraged by predictions that electrical engineers would not be able to find jobs. He conjectured further that career “fashions,” which may dictate that it is more desirable to be a lawyer than an engineer, have had an inhibitive effect.
Bill Long of Business Performance Research asked whether the relatively long history of multinational cooperation in Europe gives it an advantage over other countries in this symposium. Dr. Beinvogl said that they had indeed learned to cooperate across borders, which is an advantage. He said the cooperation was necessary, moreover, to achieve a critical mass to compete with a country as large as the United States.