- The worldwide market for semiconductor memory will grow at an average annual growth rate (AAGR) of 13.9% for the period through 2007.
- Volatile memory, growing at an AAGR of 14.3% will outpace nonvolatile memory.
- New applications of flash memory will not be able to keep pace with the overall growth in PC, workstation and network equipment memory requirements.
- In 2007, nonvolatile MRAM technology will begin to reach mass-scale production.
- Beyond 2007, as the price of MRAM falls, MRAM will become the commonly used memory chip for the main memory of PCs and main memory network server chips.
VERTICAL MARKET OVER INDEPENDENCE
Japan has traditionally dominated the computer memory chip market. Japan's multi-nationals Fujitsu, Hitachi, Nippon Electric Corp (NEC), Mitsubishi, Sharp, and Toshiba have for the most part controlled the production and supply of memory chips. However, the major focus for Japan's multinationals hasn't been on either memory chips or semiconductors. The majority of revenue for these companies has come from the production of computers, computer peripherals, telecommunications equipment, and consumer electronic systems. Semiconductor and memory chip production has been part of a vertical integration strategy for Japan. The majority of memory chips produced by these companies have ended up in the electronic systems they produce. In the past, this strategy provided a consistent supply of memory chips at low prices. This permitted Japan to offer low-priced consumer electronics and to capture the largest share of the worldwide consumer electronics market. Furthermore, the Japanese have been able to make a profit on their computer memory chips by selling excess production on the world market.
This vertical integration strategy in a more globalized economy has not been successful in recent years. As competitors from Korea and Taiwan have increased market share and capital investment costs for the production of memory chips, the price of computer memory has gyrated. Therefore, it has been less expensive for Japan's electronics companies to purchase computer memory chips from outside vendors rather than from their own internal operations. This has resulted in significant losses for Japan's memory chip operations. In response, Japan's memory chip and consumer electronic markets have restructured. Taiwan now manufactures a large percentage of Japan's commodity memory chips such as DRAMs and ROMs. In 2002 Japan has refocused its development efforts on advanced memory chip technology and the production of higher margin chips such as flash memory chips and embedded memory for system-on-a-chip products.
For the most part, American companies abandoned the DRAM memory market. Micron Technology is now the only major American manufacturer in the DRAM market. IBM and Texas Instruments both left the mainstream DRAM market to focus on products with more market stability. Advanced Micro Devices, another American company, competes in the nonvolatile flash memory market through its joint venture manufacturing facility, Fujitsu AMD Semiconductor Ltd. (FASL). European companies such as STMicroelectronics have remained in the flash memory market, but like Texas Instruments they abandoned DRAM operations. Infineon Technologies is the last major European manufacturer that competes with Asia for a share of the DRAM market.
Unlike most European and American memory chip companies, Japanese memory companies were vertically integrated into their parent's consumer electronics operations. Japan's memory companies are part of multi-conglomerates that have major operations in other areas of the electronics market: PCs, digital cameras, supercomputers, automobiles, telecommunications and wireless phones. Because of this, Japan's memory companies were guaranteed their parent company as a primary customer. Korea's Samsung memory chip operations are also part of a multi-conglomerate, and like the Japanese, they offer a multitude of varied products. Samsung, the world's leader in memory chip sales, has operations in cellular phones, PCs, set-top boxes and televisions.
It can be concluded that Japan initially had an advantage in the memory chip market. This was because Japan's computer memory chip companies had their parent companies as guaranteed customers and as providers of the large amounts of capital needed to build computer memory chip factories. Furthermore, Japan could produce lower cost consumer electronic systems because they could obtain computer memory chips at a lower cost from their own computer memory chip operations. This situation, however, worked against them when more competitors entered the market and when the computer memory markets suffered downturns. During these times, the price of obtaining computer memory chips on the open market was less than the cost of obtaining computer memory from their own computer memory chip factories. This not only resulted in losses from Japan's computer memory operations, but also Japan was not able to compete effectively in its traditional lead market-consumer electronics.
The question facing Japan was: It is more effective for computer memory chip companies to compete as part of conglomerates or as sole entities?. From developments in the memory chip market, the industry has concluded that an independent memory chip company is the better strategy. Computer memory chip companies as independent companies have more flexibility. For example, independent memory chip companies can form joint ventures with other computer memory chip companies, which gives them access to the large amounts of capital needed to build state of the art chip factories and low cost chips. This also frees their parent companies from devoting large amounts of capital away from their primary operations.
But is the pendulum about to swing back again? New developments in chip manufacturing suggest that Japan has reduced the initial capital investment for the production of memory chips. Recent developments at Trecenti Technologies, the joint venture between Hitachi and United Microelectronics (of Taiwan), may prove the beginning of a trend. In February 2002, Hitachi purchased United Microelectronics' interest in the venture and is now the sole investor. Trecenti was formed in March 2000 as Japan's first 12-inch wafer fabrication facility for the mass production of memory and advanced system-on-a-chip products.
Below are the emerging memory technologies:
Ferroelectric RAM (FERAM)
High Density Memory
Magnetoresistive RAM (MRAM)
Multichip Stacked Packaged Memory
Magnetoresistive RAM is expected to replace flash nonvolatile technology in the long term. By 2004 if not sooner, flash memory chips will face increased competition from ferroelectric memory for cellular phone applications. In 2002, PSRAM (an embedded DRAM technology) will replace SRAM chips for cellular phone applications. By 2004, PSRAM also will be replaced with both ferroelectric and magnetoresistive RAM.
Embedded memory technology remains an impediment to the growth of low-density discrete memory chips. This is because logic chips can now incorporate memory blocks in the megabits range. Microprocessor manufacturers, such as Intel, have developed microprocessors that have 2 Mbytes of SRAM memory within the microprocessors. This has resulted in the market for discrete SRAM chips for motherboard applications virtually disappearing and the average selling price of advanced microprocessors increasing.
Stacked package memory, where multiple types of memory die are bonded together and stacked vertically on top of each other, is another market that will grow fast in 2002. This is because of the reduced board surface area requirements needed with stacked technology. Stacked package memory will become one of the primary types of memory chips for portable applications such as cellular phones and laptops.
Content Addressable Memory (CAM) has emerged because it can now be produced at cost levels that are manageable for high-end applications. The technology, which is especially suited for network equipment, also offers promise in traditional applications such as PCs.
STUDY GOAL AND OBJECTIVES
This report has market, technical and objectives. The market objective is to breakdown worldwide semiconductor memory product and company market shares, determine specific growth segments of the computer memory market, pinpoint potential new computer memory technology market application areas, evaluate companies that have valuable technology portfolios and examine competitive factors that the computer memory industry must face.
REASONS FOR DOING THE STUDY
This study has been undertaken for several reasons. One is to determine the important trends and directions in the worldwide computer memory market. This includes both technology and trends. Another reason for this study is to provide a model to analyze the supply and demand for memory chips, which is important to both producers and purchasers of computer memory chips. Producers, in order to maximize profits and minimize losses, must plan their capital investments such that they correspond to worldwide memory chip supply and demand cycles. Buyers must also plan their purchases such that they minimize excess inventory and have ample supplies to meet demand for their end products.
Another reason for this report is to determine both present and future applications for computer memory. Applications are one of the primary growth forces in the computer memory market. Personal computers have been the traditional growth force in the memory market, but new products such as cellular phones and digital cameras have allowed the computer memory market to further sustain itself. Future applications will be vital to both growth and profitability of the memory chip market. Without new memory chip applications, memory chip companies will not be able to grow and advance the technology.
The final reason for this report is to determine the effects new computer memory technology will have on the industry. In this regard, this study examines recent computer memory technology developments and compares and evaluates the technology for future mass production.
CONTRIBUTIONS OF THE STUDY AND FOR WHOM
This study makes several important contributions. These include a methodology to determine computer memory supply and demand. In this regard, the report breaks down the various application end-user segments that contribute to demand. The report also breaks down the production capacity base of memory chip companies. A production capacity formula is presented that allows calculation of memory chip production capacity based on the feature length of the manufacturing process, wafer size and the number of wafers produced per month. The report also contributes technologically. It does so by explaining the basic technological and manufacturing components of computer memory. Furthermore the report examines technology that has the potential to significantly reduce the costs and improve the performance of electronic systems that utilize computer memory.
The audience for this report includes professionals who must have extensive and technology information about computer memory. This includes professionals from most segments of the electronics industry. Specifically, cost and technology driven companies such as PC, digital camera, cellular phone and electronics contract manufacturers should study this report. Advances in memory technology and developments in memory market are expected to have a profound influence on both PC and cellular phone companies operations.
The report is also important for companies that are involved in the semiconductor market. This would include chip manufacturers, fabless semiconductor companies, design houses, semiconductor packaging and assemble companies, and semiconductor equipment manufacturers. Disk drive, smart and memory card, module companies, and producers of memory subsystems also will benefit from the report.
The report is important for a wide variety of personnel involved in the electronic industry segments, including product development managers, investment analysts, venture capitalists, integrated circuit design team managers, operations managers, buyers, managers of R&D departments involved with the design of integrated circuits, CEOs and CFOs of both large and start-up design companies and those that create and sell computer memory design. Newly formed computer memory companies should have an interest in this report, as should distributors of computer memory.
SCOPE AND FORMAT
This report provides a detailed analysis of worldwide market and technology issues associated with the computer memory industry. Additionally, it provides pivotal information needed for objective investment analysis, the selection of computer memory and the acquisition of computer memory companies.
The report's primary focus is on companies that produce memory integrated circuits. The report also includes secondary information on companies that offer technology that competes with memory integrated circuits.
Comprehensive information and analysis are also included. There is a breakdown of industry sales by company and product types as well as 5-year forecasts for computer memory by product type. In order to substantiate the report's conclusions, the report examines the historical financial data, venture capital funding, acquisitions, technological trends, license agreements, and distribution agreements. Also this report substantiates its conclusions through the examination of computer memory products and associated patents all with respect to the market and sales capabilities of the computer memory companies.
Computer memory chip companies are divided into four different industry categories: Tier one IDMs, Tier two IDMs, Fabless companies, and IP companies. Tier one integrated device manufacturers (IDMs) have wafer fabrication manufacturing capabilities and annual memory chip revenues over $1 billion. Tier-two IDMs with revenues between $100 million and $1 billion are also covered. Memory chip companies without manufacturing capabilities-known as fabless companies-are also included. Fabless companies design and market memory chips and outsource the manufacture of their memory chips. Finally, Intellectual Property (IP) companies license partial or complete memory chip designs to fabless companies and IDMs.
In addition to this tier approach, the report also breaks down the market by product type. The report classifies computer memory market into three main types: nonvolatile, volatile and specialty. The main memory types are further sub-classified into the different market product subcategories. Specific memory chip product sub-segments include CAM, DRAM, EEPROM, embedded DRAM, embedded FLASH, EPROM, ferroelectric, FLASH (Both NAND and NOR), PSRAM, ROM, SRAM and Multichip Packaged Chips (MCP). Other sub-classifications include high speed, low power mobile, and high- and low-density memory chips from 1-Gbit down to 1-Kbit.
The report's primary focus is on the two memory chip categories that hold the majority market share: FLASH chips and DRAM chips. The other memory chip categories are broken down in less precise terms, which was due to limitations on the report's size. Therefore trade-offs on coverage were made. Although these segments of the memory chip market have been abbreviated, the salient points regarding the understanding of these markets, the trends within these markets, and their relevance to the other memory chip market segments are explored. PSRAM and CAMs are not usually classified as traditional memory chips, but embedded ones. They are covered with regard to their effect on the other segments of the memory chip market. (CAMs are also classified as network coprocessor chips.) Embedded memory is covered along with CAMs in the Specialty section of this report. PROM, a near non-existent segment of the memory market is not covered. FIFOs, which have been classified as a logic part as well as a memory chip, are mentioned briefly.
The report also includes extensive reference information including an address book with worldwide headquarters sites and company contacts.
METHODOLOGY AND INFORMATION SOURCES
The report follows standard research methodology. For this report over 200 companies and 200 different computer memory chip products were reviewed. Information sources for the report include standard company literature, annual, 10K and quarterly reports for public computer memory companies, a compendium of conference papers related to the computer memory market and trade press journals covering the computer memory market. Additionally, several important patents were reviewed in terms of their relevance to product and industry developments.
To determine computer memory chip sales, numerous end markets that computer memory chips are used in were analyzed. The aggregate amount of computer memory used in each end market was determined. Computer memory revenue from each end market was then calculated with these aggregate sums and the average selling price of the specific type of computer memory used. Computer memory sales projections for each type of computer memory chip were then determined by projecting the growth of the applications markets.
Where precise company product sales information is not available, the estimated product sales and company sales are determined from a plurality of and product data. This data is correlated against a model that is based on a statistical average. Specific data factored in includes:
- Acquisitions history
- Average product selling price
- Customer base
- End market applications
- Facilities size, distribution and sales
- Import and export data
- Manufacturing capacity
- Marketing expenditures
- Number of employees in a specific unit
- Number of products
- Overall market growth rate
- Product acceptance in the market place
- R&D expenditures
- Total available market size
- Trade statistics