Two problems. First, when the term "cost addition" was used, it was used in comparison to the "cost addition" of 300%-500% for RDRAM, which presumably refers to the purchasers' cost, not the manufacturers' cost. There's no way that the manufacturers' cost of RDRAM is 300%-500% more than SDRAM. And, if they were referring to the purchasers' cost of DDR, there's no way it's going to be only 3% more, because the laws of supply/demand and opportunity cost, will drive DDR prices through the roof in this environment of tight supply of flash, SDRAM, and RDRAM.
Second, with respect to margins and yields, by the time DDR is needed in high volumes (if ever), RDRAM will have had the benefit of a years' rampup and still strong demand. This means that yields will be improving and that margins will be excellent. Though some of the DRAM dodos would cut off their noses to spite their faces, for the most part these guys operate rational businesses which will continue to ramp up RDRAM, drive up yields, and make money.
RDRAM is a superior product in the long run, and has the support of the gorillas. With a one year lead, it will be difficult to catch even if DDR motherboards that work soon become widely available.
Prophet |
