>>I recognize that getting this number [capacity utilization] accurately is probably fraught with all the problems of attempting to calculate ASPs. But I'd find it quite interesting, and perhaps, quite useful. <<
Wouldn't we all...
Tracking utilization is hard, among other reasons because the fab itself may not know with any accuracy. It's tough to deconvolute the effects of yield (poor process control requires more monitor wafers), product mix (more recipe changes requires more downtime, more monitor wafers), and so forth. Utilization can also change very suddenly, especially at foundries, because it's driven by consumer demand for the product. Say, for example, that a foundry expects to sell its entire production for September to a prominent toy maker, and therefore turns away other orders. If the big Christmas toy flops, utilization goes from 100% to 50% overnight. The space might be filled again next week, but the analyst who calls right after the order is cancelled is going to see a pretty gloomy picture.
ASPs are relatively easy--revenue divided by number of units shipped.
The analysts all throw utilization anecdotes around, but don't generally make quantitative predictions based on the numbers. That tells me that they don't have confidence in the data.
Katherine |