Petz,
Nonsense, sell one of the fabs.
That's not really an option. If AMD forced the issue, there would just be ton of one time charges.
Let's look at the fabs:
Fab 35 - is right now the main fab, produces most of the profitable stuff
JV3 - not fully equipped, but the shell of empty JV4 + JV3 are to become 300 mm SP1.
JV1/JV2 - trailing edge production, prime candidate for closure. The problem is that in Japan (and Europe), you can't just easily lay off or fire people. There are complicated rules and probably costs associated with it.
I think AMD plan is to equip JV3/SP1, and as the production starts there, JV1/JV2 will be phased out, and employees transferred. This will retain experienced emplyees and minimize potential one time charges related to down/up sizing.
Then there are packaging and test facilities in low wage countries, and wages there are not much of a problem and the hiring rules are probably flexible enough so that AMD runs these efficiently.
I think AMD's idea is to continue to go full steam ahead in flash to become the leader through the use of MirrorBit, turning AMD into volume and low cost producer, and as a side result destroying the NOR market as we know it (which is headed that way anyway). AMD mentioned they have 4 bit MirrorBit cells already working in the SDC. So the plan is to get these to production in the 300 mm fab - using ORNAND to take most of the old NOR market and attack the NAND market.
The problem I see is that, so far, MirrorBit has blown every deadline or plan, overall underperformed spectacularly from sales perspective. But then, if you look at K8, it had a slow start as well nut has turned out to be a spectacular product...
Joe |
