Jay - <<Wall Street is currently modeling flat equipment growth in 1998. This is too high. (DRAM) Fabs MUST reduce production to survive! They are currently losing big bucks.>>
Well, there are three things I think are wrong with this:
1) Many of them are losing money because they aren't using modern technology (as you posted in another article, Japan is using 0.5 micron). They either upgrade to 0.35 or better or they die. They put it off all last year, and it is killing them now. MU, on the other hand, is not making money, but it isn't losing it either. Having said that, they aren't likely to expand production until prices stabalize.
2) Not all semiconductor manufacture is DRAM, and in fact as DSP's take off I would expect DRAM's to drop as a percent of total semiconductor output. In any case, currently many of the Taiwanese logic foundries are overbooked, and need to add capacity.
3) As for DRAM's, the current industry estimate is that DRAM manufacturers will reach capacity sometime in late '98 or early '99. This estimate hasn't moved in more in about a year. And, I would expect capacity increase orders about 6 months before then.
Thus, I expect that there will be a slight drop in semi-manufacturing equipment orders in the immediate future (3 to 6 months), as companies and countries try to assess what is happening in a macro-economic sense. But, it will only be a small drop, since many companies must upgrade or lose their shirts, or (in the case of foundries) are at capacity.
Clark |
