Zeev, regarding your points, I agree that I have not accounted for the exact extent of the overcapacity. This is because I do not know how much excess capacity there is out there. I know that as capital spending decreases there should be some absorption of capacity as fab utilization rates improve, but I really don't know how long it will take to equalize supply and demand. Do you have any estimates of how much excess supply there is, and how long it will take to absorb it?
Regarding your second point, however, if you re-read my post you will find that I only refer to a steady increase in demand for transistors, not to chips, raw silicon, nor end sale of semi in dollars, none of which will necessarily increase in any given year. I am well aware that the total number of chips could decrease as more functions or more ram is placed on the same chip. As you cite, a 64MB RAM chip could replace multiple smaller chips for some applications. Similarly a PC-on-a-chip could eliminate driver chips, SRAM chips, BIOS chips, cache controller chips, and VGA chips. I have no way of forecasting what will happen to the total chip count, or even to the amount of wafers needed in any given year. My point was that if the semis decrease spending of capital equipment, then presumably the rate at which they are able to reduce feature size will be decreased, and that will cause them to use more wafer real estate than they otherwise would have.
I am also aware that shifts in demand to epi or 300mm could create spot shortages that will improve margins, but these kind of shifts would typically require new capital investment by the semis, and presumably would tend to make the surplus situation in the 200mm wafers worse. Due to reductions in spending I would expect the shift to these newer products to be slower than expected. If this is the engine that will drive their future profits, then logic would dictate that their recovery will come after the equipment sector, not before it. Thus this is an important distinction. If the engine to improved profits is increased demand for 200mm due to higher fab utilization rates caused by postponing capital purchases, WFR should recover before the semi equipment companies, but if the engine to improved profits is new products, then the recovery should be after the semi-equipment group. So which is it? Or is it both?
Thanks,
Carl
|
