Carl, re: By 50%, I mean 50% of the entire memory market for all DRAM of all types.
CY2000 % DRAM units by market. (64Mb equivalents. Source is Semico, but I think they can get history right).<G>
56% Computer. (DT PCs, W/S', Mobile, Servers, PC graphics, etc.)
7% Peripherals (Printers, copiers, monitors)
11% Consumer (Camcorders, Digital cameras, DVD's, Set top boxes, TV's, etc).
14% Communications (Switches, Routers, LAN, Hubs, Cell Phones, Modems, etc.)
12% Industrial/Other (Robots, Test Equip., Med. Equip., Auto equip., Diagnostic equip., GPS, and Military equip.)
While the Computer segment is projected to decline to ~50% in CY2002, I'll use these numbers.
Examining your definition of Mainstream =>50% share of the total DRAM market:
Using your prediction of 35% PCs with DDR in CY2002, the ~140M PCs, (I'll throw in Servers to help with the 35%), represents about 20% of the total DRAM market.
(While I think your prediction of 35% is off, I won't argue it).
Do you see 30% DDR share coming from the other market segments?
But don't worry about definitions, arguing about them is a waste of my and your time.
If you're going to analyze the PC market, understanding product segmentation is important, particularly when it comes to CPUs and Chip sets. And since Intel definitions are fairly standard in the industry, why not use them?
(Here is where I get a silly reply from someone comparing prices of RDRAM in the present with SDRAM in the past, LOL.)
I think I've been comparing DRDRAM prices in the future with SDRAM in the future.
DRDRAM is still on a learning curve, SDRAM can't drop much further. I have no doubt that even in a rising market RDRAM will be more expensive than SDRAM/DDR. The ratio of price at the systems level is important in terms of product segmentation. Currently Intel and the Box makers have driven P4 prices well down into the "Mainstream" segment. But most important is understanding Intel's marketing strategy for the future. (Which, BTW, I don't). <G>
JMO's |
