Thursday, November 30, 2000
It was presumed that DRAM supply and demand in 2000 would be balanced-achieved by an oversupply the first half of the year and a deficiency during the second half. However, the market did not live up to expectations.
We witnessed a slight oversupply in the first quarter and severe shortages in the second. Prices escalated in the second quarter and continued an upward trend through August. The catalyst for these market conditions was the belief among DRAM buyers that a shortage was imminent. This prompted buying of DRAMs at relatively low prices-between $5 and $6.
In many cases, OEMs and brokers used the low pricing as an opportunity to build inventories. Indications were that by the fourth quarter, prices would more than double to approximately $15 for a 64-Mbit DRAM.
By itself, this heightened buying activity was little cause for concern. When the market is saturated, PC OEMs typically pare their inventory due to daily DRAM price fluctuations. They request that their supplier stock DRAMs in their hubs, and when demand is created for product, OEMs simply pull parts from those hubs. But when a shortage exists, DRAM suppliers limit the quantity of product in the hubs, and PC OEMs and other DRAM consumers take additional steps to manage their DRAM requirements.
One of these steps is to build internal product inventories rather than stockpile components at supplier hub locations, purchasing parts to meet current demand as well as warehousing additional product for anticipated price increases and allocation. At this point, other DRAM consumers, such as brokers, also begin buying DRAMs believing it to be a good investment.
As August approached, prices reached $9 for 64-Mbit DRAM parts. Market players anticipated the back-to-school PC demand would be strong. In reality, demand was relatively good, but not as strong as expected. From that point on, prices began to decline.
The reaction of DRAM brokers and other investors anticipating strong prices was panic, resulting in product inventories being sold into the open market. The flood of product on the market coupled with fragile demand established the foundation for decreasing DRAM prices, which spiraled down from a $9 contract level to the $6 range.
As the channel emptied early in the fourth quarter and DRAM inventories were depleted, PC market demand for DRAM again began to escalate. This year, PC market growth is expected to be 16% higher than in 1999. This equals total sales of approximately 136 million PCs this year. However, the importance of the PC on overall DRAM demand is diminishing.
DRAM growth in the non-PC market is expected to rise from 19% in 1999 to more than 47% in 2004, according to Semico Research Corp., Scottsdale, Ariz. In 1999, computers accounted for approximately 61% of the total DRAM market demand, followed by industrial at 12%, communications at 10%, consumer at 9%, and peripherals at 8%.
Over the next three years the computer segment of DRAM demand will be reduced to less than 40%, of which less than 25% will consist of PC applications, and the remaining 15% representing server applications.
Semico further projects that by 2004, the networking and communications market segments will account for 24% of DRAM consumption. Consumer electronics applications will increase dramatically, growing to 23%, while peripherals and industrial market sectors will decline slightly, with each taking about 7%.
If you believe the digital revolution is over, think again. My perspective is we are in the initial stage of the digital revolution, which will continue to drive demand and average memory bit growth of more than 75% from year to year. This has been the historical record for the last 10 years.
We are at the tip of the iceberg, and for the first time we are able to see multiple applications that will drive DRAM demand. It is not only PCs that represent the dominant market factor, but also telecom, networking, consumer electronics, and server-related products for Internet connections.
The demand for DRAM will be sustained at more than 70%. The supply side will be more restricted than in the past, as there are now only four major suppliers producing 75% of the worldwide DRAM output. Their combined manufacturing forecast is expected to produce less than the projected demand for DRAM bits.
Furthermore, funding for new capacity is limited, and financial houses are not yet attracted to the cyclical nature of the DRAM business. Some DRAM manufacturers have a restricted cash flow and lack the means to invest in the newer 12-in.-wafer fabs and have put these plans on hold. In addition, these manufacturers are not planning to add 8-in.-wafer capacity, which could adversely affect DRAM supply.
These factors will create constraints on the supply of DRAM over the next year. Purchasing executives should take advantage of this opportunity to strengthen alliances with their key DRAM manufacturer to ensure they have established a long-term commitment and can access product availability to continue to grow their respective businesses.
Farhad Tabrizi is vice president of marketing at Hyundai Electronics America in San Jose.
By: Farhad Tabrizi, vice president of marketing, HEI
Copyright © 2000 CST, Inc. All Rights Reserved