It's become pretty clear that 1.5x is too low a multiple for 300mm equipment. It's also pretty likely that more 200mm fabs will be built before the first 300mm fabs open up.
More fundamentally, though, I think Dr. Castellano has his cause and effect backwards. He says:
"But with the boom and bust overcapacity cycles that
have been occurring in the IC industry every three years or so, the 2.25 times more
chips on a 300mm wafer will negatively impact fab construction to the point that
only 10 fabs will be needed each year through the first half of the next decade."
Implying that 300mm wafers will exacerbate overcapacity problems. In fact, it's the other way around. Overcapacity has delayed the 300mm transition, and the transition will not occur until chip demand catches up.
In other words, if demand for chips keeps growing at historic rates, 300mm wafers will be needed to stay on the price/performance curve, lots and lots of 300mm fabs will be built, and the equipment makers will be happy. If demand for chips stalls, the equipment makers will be unhappy, regardless of the wafer size.
FWIW, my conversations with Dr. Castellano suggest that he is a thoughtful analyst, but that he also enjoys being provocative. I would read the entire report before investing money based on his conclusions.
Katherine |
