you...
since it always makes sense to run fabs at full capacity and since amd was not "capacity constrained", what prevented them from fully utilizing capacity with their "cost advantage" to take more market share.
me...
I take it your referring to the comment Dirk made in the CC? I'm not sure, but I expect that any lost capacity was probably due to planning problems such as AMD having too many server chips when they could have sold more Turions. Since revenues were flat and ASPs were up that would seem to indicate that AMD did sell fewer chips. This would make sense since it appears AMD concentrated solely on the high-end markets, and left the low-end to INTC. On the other side of things AMD did draw down inventory by $50m q/q so who knows? Anyway, I tend to believe, as I've said previously, that this was intended as a message to INTC that AMD didn't want to get into a price war, at least right now.
Speaking of cost advantages again; one thing that doesn't get talked about much is revenue per employee. From the respective company employee counts it looks like it must take 4x to 5x the number of INTC employees to create a processor as AMD needs. Of course all this is relatively unimportant since INTC ends up selling 5x as many processors as AMD making the labor costs much less significant for INTC. Yet INTC still needs to sell those processors for far more than AMD charges, on average, and still makes less per share than AMD does. I don't know but working for INTC must be like working in a rendering factory with fat hanging everywhere. Less costs yet INTC has to charge more and makes less? Something doesn't compute.
This is why it's important for AMD to take the low-end volume, despite it not being all that profitable. Every product that AMD can sell profitably is one less product INTC sells which means one less product INTC has to spread the costs over and one more product AMD can spread its' costs over. Multiply that by millions and now your talking some money.
