RE: "...as tool costs increase, the fixed cost component rises, making it more difficult to employ capital in smaller increments."
A new fab cost between $2-3B. At the AMAT conference call, it was stated that tool costs have moved up from 50% of capital expenditure to 70%. Consequently, while 30% of the capital cost, land, buildings, automation, material delivery systems, etc., may require a "big" increment, this increment doesn't increase capacity and hence doesn't have any potential to affect cycles. The 70% for tools, $1.4-2.1B, can be implemented one production line at a time. My estimate is that the most expensive tool, the lithography scanner, costs $10-15M, and a single production line costs about $200M (this total figure is top of the head and needs considerable refining). If I am correct, the fab can be filled out in 7-10 increments which are closely tied to end user demand.
RE: "That is also a very interesting and important distinction. I think it would be quite valuable to contact Luenberger and get his reaction to your theory."
I am not certain how "theory" applies to my statement. Most cyclical companies are influenced by the macro economic cycle which affects product demand. Unit demand is cyclical. Falling unit demand is exacerbated by falling prices which can fall to variable costs before ceasing to be economically rational. Semi cycles have occurred while unit demand continued to be monotonically increasing. Companies can't control unit demand, but they can try to control capacity to keep supply consistent with rising demand. |
