To: Softechie who wrote (43409 ) 3/23/2002 3:03:45 PM From: mishedlo Read Replies (1) | Respond to of 99280 MU - comments from the Rambus board Here is a shocking comparison of the #1 DRAM manufacture vs the #2 DRAM manufacture in regards to wafer fab process "production" capability.koreatimes.co.kr “Moreover, SEC plans to raise the production of chips with microcircuit processes of below 0.12 microns to an 80 percent level by the end of the year.” ebnews.com ”Micron continues to accelerate its die shrinks, projecting that by the end of the year 65% of production lines will be at 0.15-micron capability and 5% will be at the next 0.13-micron node." Personal Speculation based on reading these two articles - This implies that Samsung is one year ahead of Micron in terms of “production” process capability. - This implies that Samsung's die sizes are smaller than Micron's, regardless of DDR or RDRAM. - Samsung's production .12u RDRAM die size is probably smaller than Micron's .15u DDR. - Samsung's yield on DDR333 and DDR400 is likely much higher than Micron's due to the smaller die sizes, which means Samsung enjoys a majority of DDR premiums. - Samsung's conversion to 12”(300mm) wafers allows significant growth without building new wafer fabs. Combine this with their conversion to .10u in 2003 and Samsung has a lot of growth potential. - Likewise, Micron can grow capacity by converting to 300mm so why do they want to buy Hynix? Does Hynix have .13u capacity or is it simply to help control supply/price of DDR? It appears to me that Samsung is in a very commanding position right now. When times were good, it appears Samsung invested heavily for the future which is now paying off. Due to their lead in process technology, Samsung is in a better position to win any type of DDR price war. I see no reason why Samsung would not want to continue to help push the market toward RDRAM because it helps avoid competitive pricing issues. RDRAM is simply another extension of Samsung's long term investment in advanced technology which is now paying off. With PC1066 coming on line, Samsung will be able to further fragment the “profitable” DRAM market in their direction, without competition from Micron. It is also becoming clear why DDR2 is a 2004 product. My guess is that DDR2 requires process technologies equal to, or smaller than .13u process technology to make economic sense (just like RDRAM). It appears Micron's .13u process technology will not be “production” ready until at least late 2003. If anything, Micron probably wants to stall Samsung on DD2, via JEDEC spec issues, until they get in a position to compete effectively against Samsung. This is just another reason why it is to Samsung's advantage to help push the market toward RDRAM. Is it any wonder Samsung would want to work with SIS on a PC1066 chipset?