Memory Microchips to See New Challenges
November 12, 2001 (TOKYO) -- Vendors of memory microchips have begun to take up new challenges involving the cultivation of the third huge market following those of personal computers and cellular phones.
Their main points of focus are on three sectors: the mobile terminal sector, the network equipment sector, and that of digital home appliances. Waiting for a chance to enter these target sectors, each vendor has been accelerating development of new types of memory LSIs, while shoring up existing memory LSIs including DRAMs and flash memories.
Their targets are for early launches of the new markets around 2005, thus pushing ahead in parallel with three major tasks of developing memory technology, cultivating a new market and building up the business structure (refer to the accompanying figure).
Flagging Markets Shift Emphasis to Exploring New Sectors
A downward trend is evident in conventional memory LSIs in the background of actions taken by memory LSI vendors toward the new market. DRAM businesses have stopped growing in terms of value with low-end PCs prevailing at an increasing tempo since 1995. The PC market began to indicate a clear sign of decline in 2000. Another type of conventional memory, flash memory, also faces demands for lower costs, because the unit prices of end-use products are low.
By necessity, each memory LSI vendor has started to develop new types of memory LSIs. So, there has been an emergence of various types of memories. These include a phase-change memory and a polymer memory by Intel Japan K.K. and memory LSIs based on multi-tunnel junction (MTJ) being developed separately by Hitachi Ltd. and Samsung Electronics Co., Ltd. as well as FeRAMs and MRAMs already under development by each vendor (refer to the table).
Pushing Ahead in Parallel with Technologies, Applications and Business
Those new types of memory LSIs, however, cannot meet all the requirements in terms of read/write speed, levels of integration and power consumption. For example, the FeRAM has an issue of higher density that is yet to be addressed. Also, MRAM has a weak point, because it is not free from a tendency to increase power consumption and the difficulty in realizing smaller chip dimensions. Therefore, vendors have embarked on memory-technology development as well as market cultivation and construction of business structures, taking aim at the early launching, in five years, of the market for new types of memory LSIs.
In the LSI business, application technology development requires tremendous burdens on companies. Thus it is not easy for a single company to carry out development of basic technology as well as such an application technology. To cope with this situation, each memory LSI vendor has begun sorting out all the development matters, including those for in-house work and those for joint development or outsourcing.
For example, a vendor may choose joint development of a process technology or other basic technology with its competitor, while pursuing some differentiation by concentrating on application technology for itself. In accordance with this principle, Infineon Technologies AG of Germany has a business tie up with Toshiba Corp. for developing FeRAMs, and with IBM Corp. for developing MRAMs, respectively. In terms of application technology, vendors are applying the "design-in" method to the memory LSI sector, which is now widely used in the logic LSI sector. IBM is the one that has already introduced the design-in method.
On the other hand, running a business of existing memory LSIs requires a business structure where a vendor can focus on new markets while maintaining business in conventional markets. This is because new markets have more growth potential while old markets have a larger scale in terms of absolute value. In order to maintain conventional markets, it is necessary to leverage joint development as well as external resources, such as silicon foundries, in the direction of this memory business.
Heading for New Markets, Exploiting Strengths
New application markets are expected to open in 2005 for either new memory LSIs or existing memory LSIs. Taking aim at this target, vendors will leverage performance required in the new market.
The new types of memory LSIs have a different principle of cell operations. Therefore, it is highly possible to achieve the level of performance the new market will require. Among others, the FeRAM excels in read/write speeds and the number of writing times, and it features low power consumption. Thus, a memory device with these features may easily meet requirements of mobile equipment. The MRAM device has potential for outrunning DRAMs regarding read/write speed and level of integration. According to the parties concerned, if these features are highlighted, such memory devices can be optimal memory-LSIs for network equipment requiring higher performance levels and large capacities.
Existing memory LSIs are heading for their new market based on business results so far achieved with actual product shipments. As they have their basic technology already completed, they can be improved upon quickly. Thus, major DRAM vendors have started to ship modified products, changing a package size to a smaller one suitable for implementation into mobile equipment as well as loading a power-down mode.
Some major flash memory vendors have set to work on expanding application fields of cellular phones. According to Fujitsu Ltd., one of top-ranked vendors of flash memories, its shipment ratio of the memory for cellular phones has fallen below 50 percent. Adhering to this policy, each memory LSI vendor will drastically improve on memory cell structures as well as circuits.
FeRAM Strategy Clear-Cut
A scenario for launching the market for new types of memory LSIs is outlined as follows: starting business in fields with a less intense rivalry with existing memories and showing good results; then taking the offensive into the three true targets of network equipment, mobile terminals and digital appliances. In sum, it will take a step-by-step approach. If they try to directly break into those three sectors, their lack of experience in DRAM and flash memory might work against them when establishing an alliance with others for design-in.
In this regard, FeRAM has a clear scenario for a market start-up. Related vendors will first concentrate on the IC card sector, which is not seriously challenged by DRAMs or flash memories, through making the best use of features such as high-speed writing, nonvolatile characteristics and low power. After showing good results, they will try to enter the mobile-terminal market. Already, Fujitsu and NEC Corp. have built this strategic model.
An MRAM scenario might not look so clear as that of the FeRAM. But vendors take a hopeful view on launching a market in the network infrastructure sector where very high-speed operation is needed, leveraging MRAM's features. In addition, the three types of memory, namely a phase-change memory, a polymer memory and a memory using MTJ, will be commercialized around 2005, according to their plans. Therefore, it is necessary to draw up scenarios for these memory devices, as concerns how they should go into products from the very beginning.
Heading for Third Huge Market for Memory LSIs with Development of Basic and Application Technologies Together
Three Sectors should develop altogether: memory technology development, new market development and business structure development
Concept chart created by Nikkei Microdevices
Performance Comparisons of Various Types of Memory LSIs
DRAM NOR FeRAM MRAM Phase change memory Polymer memory STTM
Operation speed fast slow fast very fast fast slow fast
Integrated density high high currently low high in principle high in principle high in principle high in principle
Power consumption high low very low low low low very low
volatility volatile nonvolatile nonvolatile nonvolatile nonvolatile nonvolatile nonvolatile
Production level mass production mass production small quantity under development under development under experiment under experiment
(Yasuaki Nagahiro, Staff Editor, Nikkei Microdevices) |
