you:
It would be a pleasure to see, but its highly unlikely, when
you ask me. Your right, much will depend on the cost side,
but this takes time, as you know as well.
me:
Well, were we may differ is that I think fab25 with the switch from AL to Cu last year is making great headway and can produce enough to completely offset TSMC.
Yes, I expect SP1 is producing at capacity, but like any new process that will continue to improve, hopefully a lot and soon. However, I do expect demand for Eclipse, both 90nm and 65nm, to outstrip capacity by the end of the year. Another thing is that currently SP1 is producing NOR and ORNAND, no production of 65nm Eclipse yet(samples yes). I expect margins on Eclipse to show very big gains as these will not only have the SP1 cost advantages but will also include higher density single chip packages. These chips will allow SPSN to play in higher-end wireless markets that have been dominated by NAND/DRAM combos. Also I expect that CSID demand for higher density MirrorBit products will be considerable.
Yes, BIST only applies to SP1 and as I understand it, currently it is only a partial implementation, not the full package. So testing costs should continue to fall as SP1 ramps and BIST continues to improve.
Yeah, demand is the real question. As I said I expect a real sea change in demand with Eclipse. Unless the entire market goes to hell SPSN should do ok. The other guys are going to have lots of problems though.
In the q4 CC management said that ASPs would be down about 10% in q1, all of which and possibly more, would be made up by cost reductions. Most of those cost reductions should be coming from fab25.
Numonyx is a big question mark, but in any case I can't see them being the competition they've been in the past. Unless INTC/STM decide to throw an ever increasing amount of good money after bad there's little they can do except try to contain their losses by reducing their exposure in markets where SPSN has a clearly dominant position. |
