Bob, my wording was, indeed, poor. What I should say (and meant) was that AMD had probably received an EARLY sample of the CURRENT generation of DUV tools (throughput of 80wph). My point though was not that they are DUV pioneers, but that they'd probably got the tools at least a year prior to the ramp up.
Returning to your 64mb ramp up trigger for DUV: as James had pointed out, repetitive structure of DRAMs could lend itself to reasonably competitive non-laser solutions, thus slowing things down a bit in the DRAM world. My impression (that I will try to check out tomorrow) was that MU is going to produce 64mb DRAMs with .3mu technology. They are generally considered savvy, low-cost producers (of course, that may be because their .35mu 16mb line is technologically superior to older .5mu processes).
I guess the 64mb die area diminishes quadratically with the reduction in line width. If .3mu 64mb dies can be manufactured with I-line equipment, per-wafer yield will be around 70% of the .25mu yield, but i-line maintenance costs are lower, consumables are cheaper; somewhat longer delays for DUV tools may delay market entry, etc.
Even with these caveats, 64mb crossover is undoubtedly important for increasing DUV demand.
CYMI did well yesterday and SOX did recover after early selloff - good signs for this morning.
Regards,
Y. |
