Bobs,
Berti sounded somewhat frantic about getting Eclipse qualified. I assume that's because a lot of SP1 capacity is being reserved for Eclipse. Once Eclipse starts into full production this q or next things should really start to heat up. Finally, Sept, Oct and Nov are the hottest sales months for SPSN so overall their goals seem realistic, but who knows?
My understanding is this:
- there is the base MirrorBit NOR, which now covers 80% of sales
- there is a specific implementation of Mirrorbit NOR called ORNAND.
- ORNAND is selling in its base implementation, still at a relatively low level (mostly in Japan). From CC, about $52M ($26M 65nm SP1 + $26M 90nm Fab 25)
- then, there is a specific implementation of ORNAND called Eclipse.
- there will be another implementation of ORNAND called EcoRAM
I would guess that the remaining 20 floating gate is in JV3 plus some trailing edge Mirrorbit in JV3.
Fab25 is mostly Mirrorbit, maybe a tiny bit of ORNAND (~26M).
SP1 is all Mirrorbit, and I guess mostly ORNAND (~26M). Spansion is betting that a lot of the future growth will be here.
The problem for Spansion has been the slow qualification by customers, resulting in slow uptake of 65nm. Spansion spent the money to be able to run 2000 wpw I believe at the end of Q1 or Q2 time frame. Now they say they will get to 2000 in Q3. So the uptake has been much slower than Spansion expected, but they are optimistic about H2, especially Q4.
From cost point of view, SP1 will produce very cost effective parts, but low utilization prevents that. As utilization approaches the target in Q3, costs go down, margins go up...
Joe |
