JC,
I don't know NEWP. But I believe Zeev Hed is, or has been, a fan.
Re SIA Forecast for 2002 and capacity utilization: (Different Post and poster)
There are far too many moving parts to guesstimate capacity utilization knowing only the Total annual $ sales volume.
For e.g., prices fall 30% / annum on average. That means that $100 of chip sales at yearend 2000, would be priced at $49 at yearend 2002 on average all else equal. And nothing else is equal. So the leadin statement is absolutely wrong if looking at any specific chip.
During periods of overcapacity, prices fall faster than trend. That is certainly the case this year. Thus more chips must be sold to make up the $ volume. This tends to use up excess capacity quickly.
During periods of rapidly falling prices, Fabs tend to reduce their costs / chips. One key way to do this is through shrinks or other process improvements giving them more good chips / wafer. Getting more working chips / wafer is the same as getting more capacity.
More complex chips are made combining the function of 2 or more chips on a single chip, whether that's additional function, or just doubling the number of bits on a chip as is done with DRAM; or changing the mix of chips produced.
In a nutshell, seeing more $ forecast is encouraging; but one would need to look at the forecast details before forming any capacity conclusion. And that presumes one also knows how much new, leading edge capacity is coming online; and how much old, trailing edge capacity is being mothballed.
FWIW,
Ian. |
