After the price of DDR2 1Gb eTT chip rallied over 10% since February 20th, it remained short term stable and dropped again last week (02/24-03/02), the DDR2 1Gb eTT chip price was down more than 10% from US$ 0.95 to US$ 0.85, once again fell through the mid February low US$ 0.86. About the original chips, the price drop and volatility was not as harsh as the DDR2 1Gb eTT, the DDR2 667 Mhz 1Gb chip price dropped from US$ 0.89 to US$ 0.83 with a 7% price decline. If observed from the market side, the spot market eTT chips were mainly dominant by the Taiwanese DRAM vendors. The recent integration related news was all over the media every day and the DDR2 eTT spot chip price also fluctuated mysteriously. Underpinned by the upcoming announcement of the DRAM industry integration directions by the Taiwanese government this week, the DRAM chip price may go back to its fundamental track.

The DRAM industry has faced two years of downturn, the global financial crisis started from 2H08, and the freezing demand. The 4Q08 DRAM chip price dropped more than 54% and the DRAM vendors were dealing with severe pressure of cash outflow. The immediate adoption of a newly developed process technology by DRAM vendors is no longer being witnessed. Even though the new technology can be immediately induced to mass production, the DRAM vendors now do observe the global DRAM market supply and demand situation and evaluate their own CAPEX, they also do consider the cost of equipment investment and future amortization issues.

The development of 5x nm process represents a more rational investment by DRAM vendors. According to the roadmaps of DRAM technology providing vendors, the 50 nm induction time table seems to be on the delaying side with the length of one to two quarters. Based on our understanding, the development progress of the 5x nm process by the varying DRAM vendors are proceeding quite well. Comparing to the 6x nm process, the gross die of DDR3 1Gb chip is 40% or even 50% more than the DDR2. The cost per chip could decrease 20% to 30% (including packaging and testing) but the investment amount needed for equipments is higher than the 70 nm or 60 nm process. This is the main reason that drags the schedule of 5x nm process migration. The 50 nm process requires immersion tools is because the line width of 50 nm process is narrower than the 60 nm, the current yellow light equipment can not meet the needs of the 50 nm process. But the immersion tool is pricing at one to two billion USD, the current 100K 12 inch capacity requires five to seven equipments. The cost of the equipments is quite big a burden of the DRAM vendors who are currently short of cash. How to use the current equipments to shrink the die size and appropriately invest in 50 nm process equipments instead of full throttle switching to immersion tools are tests and a matter of life or death to the DRAM vendors.

By: DocMemory
Copyright © 2023 CST, Inc. All Rights Reserved