5
A Typology of Alliances

If the 16 alliance types described in Chapter 3 are listed by frequency of occurrence in 1990, they can be rank-ordered as illustrated in Figure 9. The order is not surprising in terms of where the U.S. and Japanese semiconductor industries stand today. One would expect fabrication contracts, sales/marketing agreements, and joint development projects to be among the most actively sought and commonplace forms of bilateral alliances, given the current commercial situation. Still, the relatively small number of technology exchanges and patent licensing agreements is somewhat surprising, because these offer the simplest and most direct routes to technology transmission. For several decades, licensing of technology through sales of rights to patents was the sole form of U.S.-Japan semiconductor alliance, and the evolution of new and more complex alliances reflects the maturation and complexity of the industry itself.

To understand the impact of various types of strategic alliances on technology transfer, we can modify and expand our original classification, using two criteria on which to construct a typology: (1) the level of commitment to repeat transactions over time (taking into account the ease or difficulty of exiting from various types of strategic alliances) and (2) the degree of closeness or joint involvement and common organizational membership (organizational fusion). The various types are listed in Table 3.

The first cluster in Table 3 is the easiest to enter and exit, and constitutes the largest category of U.S.-Japan alliances (see Figure 10) but technology transfer is limited largely to one-shot transactions. The opportuni-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy 5 A Typology of Alliances If the 16 alliance types described in Chapter 3 are listed by frequency of occurrence in 1990, they can be rank-ordered as illustrated in Figure 9. The order is not surprising in terms of where the U.S. and Japanese semiconductor industries stand today. One would expect fabrication contracts, sales/marketing agreements, and joint development projects to be among the most actively sought and commonplace forms of bilateral alliances, given the current commercial situation. Still, the relatively small number of technology exchanges and patent licensing agreements is somewhat surprising, because these offer the simplest and most direct routes to technology transmission. For several decades, licensing of technology through sales of rights to patents was the sole form of U.S.-Japan semiconductor alliance, and the evolution of new and more complex alliances reflects the maturation and complexity of the industry itself. To understand the impact of various types of strategic alliances on technology transfer, we can modify and expand our original classification, using two criteria on which to construct a typology: (1) the level of commitment to repeat transactions over time (taking into account the ease or difficulty of exiting from various types of strategic alliances) and (2) the degree of closeness or joint involvement and common organizational membership (organizational fusion). The various types are listed in Table 3. The first cluster in Table 3 is the easiest to enter and exit, and constitutes the largest category of U.S.-Japan alliances (see Figure 10) but technology transfer is limited largely to one-shot transactions. The opportuni-

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy FIGURE 9 American-Japanese semiconductor alliances by agreement type: 1990. Source: Bruce Kogut and Dong-Jae Kim, "Strategic Alliances of Semiconductor Firms," unpublished report to Dataquest, January 1991. ties for more sustained technology transfer are built into the second cluster of alliances, repeated transactions, but this represents the smallest category. The optimal framework for continuous transfer is embodied in the third cluster, the organizational fusion category, featuring some form of common organizational membership. Despite high entry barriers, a surprising number of alliances fall into this category. It should be pointed out, however, that alliances in the third cluster require the highest levels of up-front investment and risks, and are the hardest to organize and manage (but once organized, the most apt to be sustained). Still, movement toward the third cluster increases the likelihood of long-term interaction and a two-way flow of technology. What separates cluster I alliances from those in clusters II and III is that the latter two give rise to quasi-formal or formal structures of continuous technology transmission over time. The alliance types in cluster II (especially joint development projects) offer limited opportunities for continuing or two-way technology transmission. Because of the seminal and growing importance of software, the formation of strategic alliances based on standards coordination is becoming increasingly common and important in terms of giving shape to the rapidly changing marketplace. The third cluster of alliance types, including joint ventures, product/technology-seeking investments, mergers, and acquisitions, encompasses the most enduring

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy TABLE 3 A Classification of Strategic Alliances Types I. Free Exit/Arms Distance Purely profit-seeking investments Sales and marketing arrangements Procurement Assembly and testing Second sourcing Fabrication agreement Contract research Consulting services Patent acquisition Technology exchange Cross-licensing II. Repeated Transaction/Quasi-Formal Bonds Regular exchange of technical personnel Joint development III Costly Exit/Organizational Fusion Joint ventures Participation in research consortia Product/technology-seeking investment Mergers/acquisitions FIGURE 10 Alliances by cluster grouping: 1990. Source: Compiled by NRC Semiconductor Working Group using Dataquest data.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy organizational forms for ongoing technology transfer (even though the scope and flow of transfer depend on a variety of exogenous factors). As the semiconductor industry matures, more alliances of the cluster III type are forming. From the standpoint of alliance stability and the potential for greater mutuality of benefits, the direction is encouraging. Whether it actually leads to a greater inflow of technology from Japan that enhances America's technological infrastructure remains to be seen. What the U.S. semiconductor industry as a whole needs (a stronger manufacturing infrastructure) may not be transmitted effectively through the mechanism of private company tie-ups. Alliance agreements may fulfill individual company needs but not meet collective needs. Here again, there may be a disconnection among corporate objectives, industry-wide needs, and national interests. It may take an industrial association (such as the SIA), a business federation, or the government to call attention to the broader, longer-term needs of the semiconductor industry as a whole and to encourage companies to structure alliances to ensure that the country's infrastructure is strengthened. In Japan industrial associations and MITI play that kind of role. They remind private companies of broad, collective interests and goals in seeking to harmonize activities at the three levels of company-industry-nation. The fate of a key industry such as semiconductors is far too important to leave completely to the market mechanism (no matter what laissez-faire economists say). In the United States too, there may be a need for "extramarket" consensus building, harmonization, and encouragement to contribute to the collective good, although the mechanisms will necessarily be far less formal than in Japan. DIRECTION OF TECHNOLOGY FLOW: NATIONAL INTERESTS Outside the realm of formal alliances, a variety of informal channels exist through which technology transfer occurs. Foreign companies seeking to obtain technology have direct and open access to knowledge in the public domain—university-based research, university-industry research consortia, graduate training, contract research, academic consultation, scholarly and business journals, professional conferences, patent application disclosures, personal relationships, and so forth. Although measuring the volume of one-way technology outflow through such public channels is impossible, the fragmentary and unsystematic evidence (such as the number of foreign nationals receiving graduate training in the United States) suggests that the volume of flow is heavy—perhaps heavier than the one-way transfer through strategic alliances. Contract research and consultants commissioned by Japanese companies also have the effect of moving knowledge into Japanese hands (unfortunately, data for such transactions are unavailable, and there is no

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy way of knowing how widespread these practices have been). Second sourcing and fabrication agreements have served to transfer knowledge from small U.S. start-up firms to large Japanese semiconductor manufacturers. Similarly, although to a much lesser extent, the same can be said for assembly and testing agreements. At bottom, the decisive factor driving such one-way transfers may be comparative weaknesses in U.S. manufacturing infrastructure and processing technology. The Japanese and Asian edge in manufacturing across so many different sectors may mean that product niche technologies will continue to move out of the United States into Japan and Asia. Industry leaders note that U.S. industry has greatly increased its emphasis on quality and manufacturing productivity in the past few years.25 In 1990, cluster I alliance types accounted for 35 percent of all bilateral activities in semiconductors. Cluster III also encompasses alliance forms that tend to push technology in one direction. Acquisitions, mergers, and equity investments have the explicit objective of obtaining new products and technology. The abundance of capital in giant Japanese companies, their willingness to invest in the United States, the small U.S. start-up companies' desperate need for capital to survive, the striking differences in industrial structure, and the system of permanent employment (which helps to insulate Japanese companies from outside take overs) provide incentives for such Japanese investment. Over time, these factors may also erode the capacity of U.S. firms to invest as extensively in Japan as the Japanese invest in the United States. Japanese corporations have accounted for nearly all of the equity-related transactions in semiconductors over the past five years. By their very nature, certain types of strategic alliances generate a two-way flow of technology. Included here are technology swaps, cross-licensing agreements, personnel exchanges, joint development projects, and joint ventures. Such positive-sum alliances grow out of a commonality or convergence of corporate interests and goals. Technology swaps, cross-licensing (cluster I), and regular personnel exchanges (cluster II) are common and relatively simple alliances to conclude; the costs and risks tend to be low but so too, by the same logic, are the advantages and benefits. Joint development projects and joint ventures represent far more challenging and complicated types of strategic alliances. They usually grow out of a history of mutually satisfactory interactions or the powerful pull of a common, overriding threat or a common customer base. 25   Some leaders believe that the SEMATECH consortium, with 14 U.S. member companies sharing the $200 million annual cost with the U.S. government, has been a key factor in helping to narrow the quality and manufacturing gap with Asian producers. Demonstrable evidence will come as U.S. manufacturers in larger numbers offer competitive manufacturing and second-sourcing services for U.S. Firms.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy The number of joint ventures is still relatively small. In 1990, joint ventures represented only 14 percent of all U.S.-Japan strategic alliances. In view of the formidable difficulties of organizing and sustaining successful joint ventures (such as deep-seated differences in corporate cultures), the paucity of joint ventures is hardly surprising. Indeed, considering the very small number of joint ventures that have managed to survive the passage of time, we can understand why more joint ventures fail to get organized. The mere fact that a strategic alliance takes the form of a joint venture does not necessarily mean that technology will be exchanged; sometimes the direction of flow is lopsided. Nevertheless, in those few cases where joint ventures have succeeded, the mutual benefits to both sides tend to be substantial. Joint development is another form of strategic alliance that can have the positive-sum effect of enlarging the overall pie. If organized properly, joint projects function to diffuse technology in two or more directions. Like joint ventures, joint development alliances tend to grow out of a history of mutually satisfactory transactions (e.g., iterative procurements and second sourcing). Most joint development projects in 1990 were based on central processing units (CPU) and microprocessing units (MPU), with relatively little emphasis on application-specific integrated circuits (ASIC) and memory chips (see Figure 11). By contrast, memories and ASICs dominated manufacturing alliances. Thus, by their nature, joint development alliances FIGURE 11 Agreement type and product technology: 1990. Source: Bruce Kogut and Dong-Jae Kim, ''Strategic Alliances of Semiconductor Firms,'' unpublished report to Dataquest, January 1991.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy tend to focus on the more complex technologies of the future, whereas manufacturing alliances—second sourcing, fabrication, assembly, and testing—concentrate more on commodity chips. What all this implies is that strategic alliances are not random occurrences. Cluster I types are the simplest alliances to enter into; these are the alliances that companies with no previous contact are most inclined to conclude. The two-way types of alliance in clusters II and III tend to follow a linear pattern, building on the foundations laid by cluster I agreements. There is a logical, step-by-step progression, as shown in Table 4. The scope of interaction, mix of risks and rewards, and level of commitment and organizational complexity increase as companies move forward from one phase to the next. Leap frogging sometimes takes place, especially when small U.S. start-up firms (with no previous experience with Japanese companies) are involved but the modal pattern is one of steady progression based on an unfolding process of interaction and mutual trust. At any time, the progression can be halted or reversed but the costs of exit increase as one advances along the continuum. The progression from joint development to joint venturing is a monotonic jump, one that sharply ratchets upward the costs and risks of joint action (as well as the potential benefits). For this reason, joint development represents a kind of natural "resting" or stopping point along the alliance continuum. Relatively few companies cross this threshold. Because joint development projects can be finely targeted and do not require permanent union but can be disbanded, the costs, risks, and difficulties tend to be significantly lower than they are for joint ventures. American and Japanese companies, bringing complementary technological and marketing strengths to the table, frequently find that they have strong incentives to launch joint development projects, particularly if they share a common customer base or operate on the same technical standards. In 1990, joint development projects represented the most common form of strategic alliance, accounting for nearly one-quarter of the total. Without longitudinal data it is impossible to discern trends, but this type of alliance TABLE 4 Alliance Progression Early Contact Regular Transactions Common Endeavors Joint Ventures Patent licensing Procurements Joint development   Cross-licensing Fabrication     Marketing/distribution Second sourcing    

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy In 1989 Hitachi and TI announced they would expand their alliance in which 1. They would jointly develop 16M DRAMs, 2. Hitachi would license TI's 64K and 256K SRAM designs and  provide manufacturing technology for the chips, and 3. Both companies would relabel each other's chips for sales. Trade-offs for TI   Benefits R&D cost savings Access to Hitachi's process know-how Credibility of Hitachi as a partner Costs Sacrifice older SRAM designs Potential loss of sales to Japanese rivals FIGURE 12 Joint R&D—Texas Instruments and Hitachi. NOTE: In 1991 Hitachi and TI announced an intention to extend the alliance to 64M DRAMs using a common design. Source: Sheridan Tatsuno, NeoConcepts, 1990. appears to be increasing steadily over time and is likely to continue to rise. The calculus of costs, risks, and benefits can be very favorable, relative to the other forms of strategic alliances, if companies bring complementary strengths to the table. The Texas Instruments-Hitachi alliance (1989) is a good example (see Figure 12). The impact of certain forms of strategic alliances on the one-or two-way flow of technology is shown in Table 5. Whether, and to what extent, the various alliance mechanisms facilitate one-way or two-way technology transfer depends, of course, on the specific provisions of the agreements. In certain cases, joint ventures will promote two-way technology transfer; in other cases, technology will be captured by only one side.26 The mechanism itself is sometimes not the decisive factor, but various types of strategic 26   Robert B. Reich and Eric D. Mankin, "Joint Ventures with Japan Give Away Our Future," Harvard Business Review, March–April 1986, pp. 78–85.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy TABLE 5 Alliance Impact on Technology Flow One-Way Flow Neutral Two-Way Flow Patent licensing Strictly financial Technology exchange Assembly/testing investment Cross-licensing Second sourcing Sales/marketing Personnel exchange Fabrication Procurement Joint ventures Contract research   Research consortium Consulting services   Joint development Standards coordination     Product/technology-seeking equity investment     Merger/acquisition     alliances have built-in propensities. What ultimately determines the direction of technology transfer are the specific terms upon which alliance partners agree. ASYMMETRICAL PAIRINGS: LARGE AND SMALL COMPANIES By disaggregating the 1990 data on strategic alliances according to size of companies, it becomes clear that the vast majority of U.S.-Japan alliances are between small-or medium-sized U.S. companies (many of them young start-ups) and large, vertically integrated, diversified Japanese corporations. Alliances have been struck between large U.S. and large Japanese corporations, but the number is much smaller than those involving small and large pairings (see Figure 13). A few alliances have been concluded between small U.S. and small Japanese companies, but these are of scant significance from the standpoint of technology transfers. Finally, an increasing number of alliances involve large U.S. firms such as IBM, DEC, AT&T, and Hewlett-Packard linking up with small U.S. firms. It is the alliance of small U.S. firms and large Japanese corporations, the most prevalent form of U.S.-Japan linkage since 1980, that is thought to transfer the bulk of American know-how to Japan. Looking at the question from the microlevel perspective of small U.S. companies, we can glean reasons for what is sometimes myopic and self-defeating behavior on the part of these firms. Perhaps the primary explanation for the propensity of U.S. companies to "give away" technology lies in the glaring asymmetry in the size and staying power of small U.S. firms, on the one hand, and of large, diversified Japanese kaisha, on the other. In negotiating deals with Japanese giants,

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy FIGURE 13 Alliances by company size: 1990. Source: Compiled by NRC Semiconductor Working Group using Dataquest data. small U.S. firms usually bargain from a position of overwhelming weakness, given their fragility and lack of independence. To get what they need, they have often felt that they had to give up their only valuable asset—marketing rights for new niche products or leading-edge know-how (especially in software). If they were unwilling to share their know-how, potential Japanese partners would not have the slightest interest in striking a deal. By leveraging technology, small U.S. start-up firms sometimes try to survive from month to month. On the list of most urgent needs are the following: obtaining large infusions of capital; finding outlets to market, distribute, and service their products; and advancing their R&D activities to compete in the next generation of products. For small U.S. firms, efforts to strike a balance between conserving negotiable assets and meeting immediate needs have often been futile. Simply staying afloat requires that they bend all their resources and energies to meet their urgent needs (see Table 6). Even if a small company happens to be cash rich or have access to low-cost capital—which is definitely not the typical pattern—and even if it possesses a "hot" product or seminal new technology that attracts eager suitors, the need to come up with the best manufacturing arrangement and to find marketing, distribution, and servicing networks in Japan's challenging market often makes it imperative that small U.S. firms connect with one or more large Japanese corporate partners. The up-front costs, downside risks, and enormous difficulties associated with independent attempts to break into the lucrative Japanese market are simply too daunting to ignore. In many cases, the bargaining power of small U.S. firms is also limited by such factors as (1) the press of time and a strong sense of urgency occasioned by short-term time horizons; (2) a lack of experience and knowledge in dealing with Japanese companies; (3) the difficulty of playing one poten-

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy tial Japanese partner off against others; (4) the reactive posture of most small U.S. companies, which typically wait passively for Japanese companies to approach them; (5) the propensity of small U.S. firms to lock themselves into alliances with the first Japanese suitors to make a proposal (without checking out other potential partners); and (6) the difficulties of breaking into established networks in Japan. Although many of these disadvantages can be counteracted if the small firm has good advice from those familiar with other U.S.-Japan alliances, including provisions for dealing with intellectual property rights (IPR), many small U.S. firms have not received such advice. American companies (large and small alike) often enter into alliances with only limited, short-term goals in mind. Nor have small U.S. firms routinely entered alliances with the conscious intent of leveraging what they learn from the Japanese partner to enhance and consolidate their strategic position down the road. By failing to go through a process of organizational learning, U.S. firms incur major opportunity costs and eventually find that they have not benefited as much as their Japanese partners. Their shortcomings stem from both structure and strategy. In stark contrast to small U.S. start-up firms, established kaisha possess an impressive array of negotiable assets, their needs are not nearly as pressing, and their whole approach to foreign alliances is long-term and strategic. To kaisha, strategic alliances represent an indispensable tool for gaining immediate commercial advantage and moving down a longer-term learning curve. Alliances are not mandatory in the sense that their short-run survival depends on them, as is often the case for small U.S. start-up companies. In this respect, the kaisha's balance sheet of assets versus needs is often the mirror image of that for small U.S. start-ups: assets far outweigh liabilities and needs (see Table 7). TABLE 6  Characteristics of Small American Companies Assets Limited and time bound   Leading-edge or niche product and/or technology   A few high-powered researchers Needs Urgent and multiple   Large amounts of capital   Timely evidence of profitability (or prospect of future profitability)   Manufacturing capabilities or access   Marketing/distribution/service (especially overseas)   Continual development of next-generation products Time horizon Short term and not bent on organizational learning

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy TABLE 7  Characteristics of Large Japanese Corporations Assets Deep and multifaceted   Abundant capital   Strong bank backing   Intercorporate shareholding pattern   Relative insulation against pressures for short-term profitability   Manufacturing excellence (costs/quality/flexibility/turnaround time)   Strengths in process technology   Extensive networks for marketing, distribution, and servicing (especially in Japan and Asia)   Vertical integration   Horizontal diversification   Strong subcontracting networks   Keiretsu organizations   Collectivization of risks, costs, and profits   Power of demand-pull procurements   Large patent holdings   Close working relations with bureaucracies Needs Contained, not desperate   Promising new products/technologies   Coverage for in-house gaps/deficiencies   Perception as a good international corporate citizen, not a predatory trading partner Time horizon Long-term and bent on organizational learning As latecomers, the kaisha have always shown a remarkable receptivity to outside ideas, but the reasons go deeper—to structural factors having to do with disparities in levels of vertical integration and scope of diversification, as well as to differences in management and corporate organization. Japanese companies are larger, more vertically integrated and diversified, have more financial resources, and are more aggressive in their pursuit of alliance opportunities abroad than U.S. firms. They also have a wealth of experience in dealing with small-to medium-sized companies, because many kaisha have extensive subcontracting networks with small-and medium-sized Japanese companies. Some of this experience of asymmetric interaction is relevant to their relationships with small U.S. firms. Almost as much as structural factors, however, the attitude and intent of Japanese companies go a long way toward explaining imbalances in the distribution of benefits. Japanese firms view strategic alliances from a longer-term time horizon and enter the relationship with the intention of

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy learning as much as possible in order to strengthen their competitive position several years down the road. Their capacity for organizational learning is matched by only a small handful of American companies. Whereas U.S. firms may gain access to low-cost fabrication facilities or realize short-term increases in sales from strategic alliances, their Japanese partners often have bigger objectives in mind, such as applying what they have learned to the development of whole families of new products. Many of the kaisha have established within their organizations new product divisions, which have no corresponding units in U.S. companies. These new product divisions have the responsibility of ferreting out all possible opportunities to add products and technologies to the company's extant portfolio. Not only do these divisions assess concrete opportunities in-house, looking at the range of commercial applications of technology in their company's R&D pipeline, they also thoroughly explore and aggressively pursue all commercial opportunities overseas, including those offered by small U.S. start-ups. It is, in part, because of such organizational differences (reflecting broader differences in corporate strategy) that small U.S. companies wind up wedded to Japanese corporations. By comparing the assets and needs of small U.S. firms and large Japanese corporations, we can understand why fledgling U.S. companies find themselves at a marked disadvantage in negotiating deals with kaisha . There are signs, however, that small U.S. firms are beginning to develop more savvy approaches to structuring their alliances with large Japanese kaisha. Small U.S. firms developing technologies that have application to products designed in the United States or, at least, are controlled in the market by U.S. standards, such as operating systems standards, networking standards, interface standards, and communications standards, may be in a relatively good position to build alliances with Japanese companies. Such U.S. semiconductor companies tend to focus on chip design. They have the skill required to design a competitive set of chips that serves a focused customer need, and they can protect their intellectual property with strong patents. These companies do not push the frontiers of process technology; they use what is available from foundry vendors. Until recently, most of these foundry relationships have been with Japanese companies with large fabrication ("fab") plants. Some have asserted that "fabless" companies have not contributed to the strength of the U.S. industry, but they have added value at one stage of the strata. Further, their success in the market has encouraged several U.S.-based wafer suppliers (such as AT&T and Philips/Signetics) to open their fabrication facilities to these firms, leading to increased U.S. content. Another group of U.S. semiconductor companies include those that make a mark by pushing forward the state of process art. Cypress Semiconductor, which used its initial public offering to build its own state-of-the-art fabrication plants, successfully utilized this strategy. In the last five years there

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy have been examples of fabless companies whose raison d'être has been improvement in some process area. One example is Power Integrations of Mountain View, California. The founders figured out a way to make very high-voltage transistors that consumed very little chip area by using the same processes commonly used to manufacture normal, low-voltage integrated circuits. Power Integrations has developed a way to integrate power supply and logic. The market for power supplies is very large because every electronic system from a complex mainframe computer to an electric toothbrush needs a power supply. This technology is also useful in producing "smart" battery charger products—an important requirement for the increasing number of battery-powered electronic computing and communications products. Power Integrations aimed to build a large semiconductor company on the venture capital it raised. It needed a large supply of wafers, and success would depend on its ability to control its proprietary technology. First, Power Integrations proved its process by building devices at Orbit Semiconductor, a small-run fabrication house in Silicon Valley that produces wafers on a quick-turn basis for fabless semiconductor firms. Next, it shopped for a large fab partner. Realizing that its market was global, it did not limit its search to the United States and did not want to grant a broad license to its technology. It found Matsushita of Japan. Matsushita would be a large customer for its products in the consumer electronics and computer divisions. A deal was struck in which Matsushita will build wafers for Power Integrations at a competitive price and will be licensed to use the technology internally, for which it will pay Power Integrations a fee and royalty. Power Integrations will maintain the right to sell products to other Japanese customers. An alliance has been formed. Power Integrations gets capital in the form of a license fee, loses the market of one large potential customer, and maintains the rights to its technology and patents. The company has since entered into a similar arrangement with AT&T. The need for volume supply drove Power Integrations into this type of deal, whereas the strength of its patents and the present commitment of the courts to respect IPR made larger companies willing to enter agreements that limited their use of the technology and authenticated Power Integrations' position by paying a substantial fee to it. TriQuint Semiconductor, a gallium arsenide (GaAs) semiconductor company based in Beaverton, Oregon, is another example of a small U.S. company developing technology-based strategies for global marketing. The company was formed through the merger of the TriQuint subsidiary of Tektronix and two venture backed companies—Gazelle Microcircuits and Gigabit Logic. The GaAs business is small now, but its potential is large because of the growing interest in high-performance computing and mobile communications. The Japanese have been active in GaAs for many years because it is

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy an important technology for some display applications. In the United States, the early backing for GaAs came from the Defense Advanced Research Projects Agency (DARPA), which was interested in the communications capability of the technology. For TriQuint, Japan is a natural market of interest, although all the major Japanese computer and communications companies have their own GaAs efforts. In order to agree to acquire the products, potential partners have insisted on an extensive license, so no alliances have been formed to date. Even though TriQuint is in need of capital, it has chosen to regard the price as being too high if technology is to be granted in exchange for capital. American interest in the field is growing, and another GaAs company, Vitesse Semiconductor, recently completed a successful public offering to meet its capital needs. Although 15 years ago, U.S. companies would have given up the Japanese market and their technology to get Japanese capital, that is not the case today. A final example is nCHIP, a company engaged in the new field of high-performance multichip modules. The founders were originally associated with Lawrence Livermore National Laboratory and the company has a technology with good patent protection for building silicon subsystems that operate at more than 100 MHz. nCHIP's substrate for the packaging technology is silicon, and its success depends on the low-cost supply of a large volume of silicon substrates. Although users are excited, it has been difficult to fund this effort because high-volume demand is still several years away. The company went to Japan in search of an alliance that would solve its supply and capital requirements, and structured an agreement with Sumitomo Metal and Mining (SMM) in which SMM is granted internal use of the technology for its subsystem products and makes an equity investment in nCHIP. The license in this case is more extensive than in the case of Power Integrations because the capital needs are greater and thus more had to be bartered. "Pure" equity with a longer time window would have been preferable, but the market will be large and the agreement maintains nCHIP's rights to market directly in Japan. Capital is available in the United States to start new semiconductor firms if their contribution is considered focused on a specific issue in the strata of semiconductor technologies. Since capital is not available to build fully integrated firms, alliances have become a way of life. Large U.S. companies increasingly recognize that alliances with start-up firms may be beneficial to their full product lines, and they are reviewing proposals made by these start-ups. Meanwhile, Japanese venture investments are not increasing at the pace they did a few years ago. Small U.S. firms must continue to pursue markets in Japan because of their size. These examples illustrate that small U.S. semiconductor companies may be moving up the learning curve in their interactions with Japan. Companies based in other Asian countries are increasing their efforts to

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy enter into alliances with small U.S. companies. These overtures are viewed favorably because of the significant semiconductor capability that now exists in Korea, Taiwan, and Singapore, countries that are presently active in seeking alliances. The growing array of players in the global semiconductor industry thus offers alternatives to U.S. companies. SYMMETRICAL PAIRINGS: LARGE COMPANIES If size disparities place small U.S. start-up firms at a disadvantage, the playing field is more level when large American and Japanese companies enter into alliances. Most of the large-large alliances involving electronics companies (as opposed to companies trying to diversify into electronics) have served to transfer technology in two directions. Often, U.S. companies ask Japanese companies to second source, fabricate, sell, or jointly develop leading-edge products, such as RISC chips and microprocessing units, as in the case of the Mitsubishi-AT&T alliance. Another common form is a technology swap, with Japanese companies supplying know-how in dynamic random-access memory (DRAM) or static random-access memory (SRAM) in exchange for American know-how in MPUs or ASICs (as in the case of the Toshiba-Motorola alliance). In most cases, both sides come to the table with clear bargaining assets, looking to conclude agreements that will fill in niches, lighten cost burdens, and spread out risks. In contrast to small start-up firms, the large U.S. electronics companies do not have their backs against a wall and are capable of assimilating new products or technologies from the outside. Yet even when large U.S. corporations are involved, fundamental deficiencies in America's industrial base—especially weaknesses in its manufacturing infrastructure, the short-run imperatives of U.S. capital markets, the semiconductor industry's overreliance on the computer industry, and the absence of a substantial presence in consumer electronics—lead to difficulty in insuring that they make the most of alliances with Japanese companies. According to Robert Reich, U.S.-Japan strategic alliances, even those bringing together partners of roughly equal bargaining power, have resulted in transfers of state-of-the-art technology to Japanese competitors.27 Not many alliances have been consummated between large Japanese companies outside the electronics industry and established U.S. electronics companies. Large Japanese companies such as Kubota and Nippon Steel seeking to diversify into semiconductors invest in, acquire, or hook up with small U.S. firms. Kawasaki Steel's joint venture with LSI Logic illustrates the balance sheet of benefits and costs that accompany diversification through 27   Reich and Mankin, op. cit.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy organizational fusion (see Figure 14). These alliances represent a substantial share of the total number of U.S.-Japan semiconductor alliances. From the standpoint of Japanese corporations, it is generally easier to deal with smaller U.S. firms. Provisions to acquire and transplant semiconductor technology are more easily arranged. Japanese companies can get more benefit from their investments, stringing together a series of bilateral deals involving several small U.S. companies offering technology in different areas. Having the flexibility to pick and choose partners based on specialized products and technologies—what might be called ''boutique technology shopping''—can be a major advantage compared to being locked into substantial deals with large, established U.S. corporations. Dealing with small U.S. firms entails certain risks, in light of their sometimes precarious existence, but most Japanese investors conclude that the advantages outweigh Japan Semiconductor (NSI) is a joint venture established in 1985 by LSI Logic and Kawasaki Steel, whose first manufacturing facility in Tsukuba began operations in 1987. Kawaski Steel made an initial investment of $175 million and owns 45% of NSI, while LSI Logic transferred its wafer process from its California plant and holds 55% equity. NSI supplies ASICs and SRAMs to its parent companies and to outside vendors. NSI's $155 million second phase facility was built last year but the start of mass production was postponed until March 1993 because of the industry recession. Orders are expected to pick up, and NSI aims to reach $150 million in sales in fiscal 1993. Trade-offs for LSI Logic  Benefits Solid Japanese partner Kawasaki long-term investment&nspace Kawasaki not a major integrated circuit rival Costs Differing corporate cultures Different product goals Long start-up cycle FIGURE 14 Joint Venture—Kawasaki Steel and LSI Logic. Source: Sheridan Tatsuno, NeoConcepts, 1992.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy the risks. Perhaps the clearest advantage is that Japanese corporations have more say in dealing with management in small U.S. firms. Only in exceptional cases, such as the alliance between Kobe Steel and Texas Instruments (involving a joint venture in ASICs), are large, nonelectronics Japanese firms willing to tie up with large American semiconductor companies. When such large-large tie-ups do occur, they tend to take the form of joint ventures or joint development projects—cluster II and III types of strategic alliances. This means that they give rise to relatively complicated organizations requiring Japanese companies to bring some technological capability to the joint undertaking. Japanese companies wishing to cross the sectoral divide and diversify into semiconductors have a hard time doing so from core businesses that are technologically unrelated. Japanese companies in what might be called the assembly and processing sector—steel production, automobiles, machinery, and precision equipment—possess the kind of technological and manufacturing infrastructure that can be used to launch into semiconductors, whereas those in sectors such as basic materials—oil, energy, textiles, agriculture, cement, paper, and pulp—do not have the requisite infrastructure for diversification. Nearly all cases of cross-sector movement into semiconductors have involved Japanese companies from the assembly-processing sector. WHY DIVERSIFICATION? To date, Japan is one of the few countries in the advanced industrial world in which large, established companies in smokestack sectors (steel), old-line manufacturing (machinery, automobiles), and skilled assembly (precision equipment) have made the difficult transition into the high-tech world of semiconductors and electronics (see Figure 15). Among the Japanese companies that have already made the transition are Nippon Steel, Kawasaki Steel, Kobe Steel, NKK (Nippon Kokan), Sumitomo Metals, Kubota, Minebea, Toppan Printing, Toyota Motor, Hoya Glass, Yamaha, Cannon, and Seiko. Comparable examples of lateral, cross-sector mobility in the United States or Europe are hard to find. Horizontal diversification is a vehicle for technology transfer from the United States to Japan; within Japan, diversification serves as a mechanism for scattering technology widely across industrial sectors. Cross-sectoral diffusion is particularly important in terms of generating new commercial opportunities. Traditionally, the Japanese company has been seen as reluctant to move away from core businesses; when diversification occurred, it was through the creation of semiautonomous subsidiaries.28 Seiko, for 28   Rodney Clark, The Japanese Company (New Haven: Yale University Press, 1979), pp. 53–64. See also Iwao Nakatani, The Japanese Firm in Transition (Tokyo: Asian Productivity Organization, 1988), pp. 87–93.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy FIGURE 15 Investments by type of company 1986–1990. Source: Compiled by NRC Semiconductor Working Group  using Dataquest data. example, diversified into semiconductors and computers through the creation of a subsidiary, Epson. What is it about Japan that accounts for the interest in recent years in diversification? Perhaps the most obvious reason is the Japanese practice of lifetime employment. When it becomes clear that their core businesses will lose comparative advantage, there are built-in and compelling reasons for the kaisha to begin branching out into promising new fields of business activity. Diversification enables them to retain their work force, utilize sunk investments, and survive as corporate entities. Stated in causal terms, the loss of comparative advantage is the catalyst (the necessary condition), and the distinctive features of Japanese industrial organization give shape to the company's response (the sufficient condition). Some Japanese corporations may try to delay the inevitable as long as possible through government intervention (such as trade protection and subsidization). However, since MITI is disinclined to protect declining or sunset industries for extended periods, Japanese companies have little choice but to explore opportunities for survival through diversification. The combination of lifetime employment and intercorporate stockholding has the effect of closing off alternative routes of survival, such as mergers and acquisitions involving other Japanese companies. The underlying forces at work pushing Japanese firms to diversify should be kept analytically distinct from the factors that account for success or failure in implementation. Examining how and why Japanese steel companies have succeeded in producing high-quality semiconductor chips is a fascinating question, since it would be hard to imagine U.S. Steel or British Steel demonstrating the same capacity to turn out DRAMs or ASICs, but the subject is beyond the scope of this report.

OCR for page 31
U.S.-Japan Strategic Alliances in the Semiconductor Industry: Technology Transfer, Competition, and Public Policy The need to diversify has been one of the driving forces behind a noteworthy trend in U.S.-Japan strategic alliances: namely, the proliferation of tie-ups between large Japanese corporations outside of electronics and both small and large American electronics companies. Whether these alliances will be sustained during a period of downturn and whether they will result in the building of substantial technical capabilities on the part of the large Japanese corporate partners are key questions for the future.