Dresden is expected to be at production on 90nm (SOI) in Q4, according the AMD's conference call.
They're going to need to be to stay even remotely competitive.
Not all expectations are met, of course.
Given that they haven't demonstrated solid yield or bin-split on 0.13 micron SOI so far, that's a bit of an understatement.
Intel is expecting to be in production on 90nm by Q2/Q3.
Of course, Intel has the benefit of having demonstrated the technology last year. I'd therefore say that Intel's probably way ahead on ramp.
For both companies, past experience has shown that the lag from initial "production" (which is more like sampling) to high volume, is 2 to 4 quarters, depending on how the ramp goes - with Intel usually ramping more quickly, but AMD, with only one or two FABs to "ramp", sometimes faster.
Of course, with 300mm technology, Intel will be ramping no more than three fabs. It really depends on how far along the development fab is at this point as to what volume they're ready for...
AMD also said that Dresden would be able to produce 75 million hammers per year from the one FAB at the 90nm node.
Must have missed that in the CC... of course, AMD has a long history of hubris in their conference call (we'll have 30% MSS and hold it... etc.)
Then they said that that's more than they think they can sell.
I think it's way more than they can possibly produce... much less sell... especially if they continue to have problems ramping the frequency of the K8 (as they have had to date). |
