THE NEW YORK TIMES, TUESDAY, SEPTEMBER 16, 1980 The Fight Over Computer Chips U.S., Japanese By PETER J. SCHUYTEN It is being billed as the fiercest battle in semiconductors since circuit designers first learned to store computer data on silicon chips more than a decade ago. After two years of promises and false starts, the 64,000-bit random access memory is finally emerging from the laboratory and entering production, with a host of American and Japanese manufacturers vying for supremacy in this next generation of semiconductor. computer memories. In the unrelenting drive to reduce computing costs by packing ever more performance into semiconductor chips, the 64,000-bit memory represents the latest in the state of the art in computer circuits. As such it eventually will take its place in everything from large mainframe business computers to data terminals and even personal comput ers. The silicon chips that form the central core of a computer's memory store the binary I's and 0's, or bits, of computer information. Typically, eight bits form a byte, or an alphanumeric character-a letter or a number. Testing Japanese Challenge The arrival of the 64k RAM, as it is called, signifies more than just the addition of another in a series of increasingly sophisticated and capacious computer memories that date back to 1970 when the 1,000-bit RAM was introduced. It is expected to play a pivotal role in determining the strength of the Japanese challenge in integrated circuits, currently the subject of Congressional inquiry. Indeed, the Japanese, even with a late start and only two companies fullyparticipating, were successful in carving out a 40 percent market share in 16,000-bit memories, the predecessor generation to the 64,000-bit chip. By contrast, five manufacturers, Fujitsu, Hitachi, Nippon Electric, Toshiba and Mitsubishi, will be out in force at the beginning of this latest generation of memory. Among the American manufacturers who have already unveiled their versions of the 64k RAM are Texas Instruments, Motorola and the National Semiconductor Corporation, while in weeks the Intel Corporation and perhaps the Mostek subsidiary of United Technologies, the current industry leaders, are expected to introduce their products. "It is clear that the Japanese have a very big push on in 64k RAM's," said A.J. Stein, vice president and general manager of the integrated-circuits Continued on Page D24 U.S., Japanese Vie Continued From First Business Page division of Motorola's semiconductor group. "It's a very key device in their future-and in ours as well." The total market for semiconductor computer memories is around $900 million, and growing at 20 to 25 percent a year, estimated Daniel L. Kleskin, a vice president of Dataquest Inc., an electronics industry market research firm. Factor in Reducing Costs More than any other component inside a computer, semiconductor memories have been responsible for the astonishing drop in data processing costs. In 1972, for example, a 1,000-bit memory sold for around $10, or about a penny per bit of storage capacity. By 1984, by contrast, the 64k RAM is expected to offer storage for about six one-thousandths of a penny, according to Benjamin M. Rosen, president of Rosen Research, an electronics industry research firm. Moreover, by incorporating on a single chip the storage capacity that previously had been mounted on four circuits, the 64k RAM greatly diminishes the number of wiring interconnections required within a computer, while reducing power requirements, size and weight, important considerations in, say, a mainframe computer that typically offers 4 million bits of internal memory storage. "Giving More, Costing Less' "As with previous memory generations, the 64k RAM violates that economic law of nature by giving more but costing less," Mr. Rosen said. Despite these advantages, however, the 16k RAM, which at present sells for around $3 in volume quantities, is expected to hold sway in the marketplace for at least another two years, as the price of the 64k RAM drops from around $50 to between $15 and $20. "I don't see the crossover point in terms of volume occurring until at least mid-1982," said Pierre R. Lamond, vice president and technical director of the semiconductor division at National Semiconductor. Next 12 Months Critical' Nevertheless, most manufacturers are vying with each other right now to "qualify" their parts, as people in the industry refer to the process by which users select their suppliers. "The next 12 months are critical," said Dean Tombs, a vice president of Texas Instruments. "That's when relationships are formed that are likely to be maintained for some time," he added. But designing a competitive chip is one thing, industry observers say, while succeeding in the marketplace is another, and that what really counts, from a user's point of view, is the ability to deliver reliable parts in quantity. 'We can't conclude a big innovative difference between one part and another," said Richard W. Anderson, the general manager of Hewlett-Packard's Data Systems division, a large user of semiconductor memories. "In fact, at any point in time, we will probably be buying parts from as many as six vendors, with typically two or three dominating," he added. Intensifying the struggle is the fact that Japanese manufacturers will be concentrating most of their efforts on the United States market. Nippon Electric, which is expected to announce formally its entry in the 64k RAM sweepstakes later this month, is said to be aiming 60 to 80 percent of its production capacity for the United States. Whatever the outcome, the semiconductor companies realize only too well that the effective life-cycle for a product such as this is around four years. In fact, most of them are already at work designing the industry's next generation of RAM. This will be a device capable of storing up to 256,000 bits of computer data on a chip. Mr. SKORNIA. I would like to turn to Mr. Willet, who, on behalf of Chase financial policy, has completed for us a study on comparative cost of capital and rates of return which we think is our central problem as an industry. He has a report on that and some charts and then I think we can have a fruitful discussion of that issue. STATEMENT OF JOSEPH WILLETT, VICE PRESIDENT, CHASE FINANCIAL POLICY GROUP, CHASE MANHATTAN BANK Mr. WILLETT. I would like to summarize very briefly the results of the study undertaken in behalf of the Semiconductor Industry Association by the Chase Manhattan Bank which looked at differentials in the cost of capital between Japanese and United States firms. The purpose of the study was first to determine the extent of the financial advantage enjoyed by the Japanese companies. [A portion of the study referred to follows:] UNITED STATES and Japanese SEMICONDUCTOR INDUSTRIES: A FINANCIAL COMPARISON (Prepared for Semiconductor Industry Association by Chase Financial Policy, a Division of the Chase Manhattan Bank, N.A., June 9, 1980) EXECUTIVE SUMMARY This study was undertaken at the request of the Semiconductor Industry Association for the purpose of ascertaining (1) the ability of the Japanese semiconductor industry to employ markedly higher debt ratios than the U.S. semiconductor industry, and (2) the extent to which the U.S. companies are adversely affected by their Japanese competitors' use of higher leverage. While the U.S. is currently the world's leader in semiconductor technology and production, the U.S. share of the world semiconductor market has dropped from 75 percent in 1969 to approximately 61 percent by 1979. The U.S. semiconductor industry is concerned that it may lack the ability to raise sufficient capital to maintain its share of the world market. Specifically, the industry believes the ability of the Japanese semiconductor companies to continuously employ considerably higher debt ratios regardless of business and economic conditions places U.S. semiconductor companies at a substantial disadvantage in raising capital. Based on the amount of capital required by the typical U.S. semiconductor company to produce a dollar of sales, we estimate $3 billion of capital was required to support U.S. industry shipments of approximately $6.5 billion in 1979. In light of the increasing capital intensiveness needed to maintain technological leadership, we estimate the U.S. semiconductor industry will require about $15 billion of additional capital during the current decade in order to keep pace with an 18 percent annual growth rate expected in worldwide semiconductor sales. The ability of the U.S. semiconductor companies to raise the capital they will require in the debt and/or equity markets and the degree of leverage they will be able to command will largely be a function of the individual company's profitability, management strength, outlook, and market position. Those companies which consistently earn a superior rate of return on capital are able to command higher leverage ratios and a more attractive price for their shares. Conversely, those companies that consistently fail to earn an adequate rate of return, i.e., fail to earn a return at least equal to their cost of capital, will encounter greater difficulty in raising capital. For this reason, we analyzed the rates of return on capital of a number of U.S. and Japanese semiconductor companies in relation to their estimated cost of capital. In addition, we quantified the financial advantages obtained by Japanese semiconductor companies from employing higher leverage and estimated the benefits to the U.S. semiconductor industry of increasing its leverage to a comparable level. While the major differences between the U.S. and Japanese financial markets are discussed, an in-depth analysis of the impact on the competitive position of U.S. semiconductor companies stemming from differences in U.S. and Japanese economic, tax and trade policies was beyond the scope of this study.1 A summary of our major findings and conclusions follows: The Japanese semiconductor companies' size and diversity provide them with a financial advantage over most U.S. semiconductor companies 1 The reader interested in an in-depth discussion of Japanese economic and trade policies is referred to the bibliography attached. |