Pravin,
There were several discussions about the mismatch of, especially Chartered start of production and demand, which I expected to be the greatest (compared to supply) in H2, when AMD has dual core and Intel does not.
On capacity concerns, I think Hector really pressed the pedal to the floor on production only when Rev E yields were proven to be good (probably beginning of Q2). You may recall a comment from either 1 or 2 CCs back that AMD makes only as much as can be sold. BTW, in the meantime, the unit output did not change much, while the avg die size went down with 90nm ramp (even with DC included in the mix).
So my (optimistic) take is that AMD can produce at least 20% more 90nm units than 2004 mix. That's 20% more than the current selling run rate of ~8.5M units.
So let's say AMD can produce at run rate of 10M units, has 100 days of inventory of CPUs. So the next 2 quarters, may have 10M from production, 2M from inventory, to end the year with 30 day inventory, and capability to sell 12M units.
I may be too high with these estimates, but I think overall, I think I am right that AMD will have capacity to bring to the market a lot more product than many expect, and the limits will be more in terms of Intel arm twisting than AMD production capacity.
But, suppose we do run into some capacity limits. Well, that just calls for higher ASPs on top of the full utilization.
So how much of an upside is there? Well, I may be too high in my estimate, but I think it is hell lot more than Q2. I think Q3 will be > $800M and Q4 > 900M. (add standard disclaimers here about some kind of disaster).
Joe |
