Last attempt at capacity, for now.
Ok, let's take another look at this issue from a different perspective. We know that up until a couple of years ago, or so, AMD was perennially stuck at about 7M processors a Q. Today AMD is producing somewhere around 11+M processors a Q. So in the last 2+ years AMD has managed to increase production by much greater than 50%.
A couple of things. First, the period roughly correlates with the introduction of Opteron, after a somewhat slow start. Second during that period there were at least two inflections points in productive capacity, at 90nm and to a lesser extent the Rev "E" move. Third, at no time during the increase in capacity was there a corresponding increase in inventory even though the scaling of the capacity additions were highly irregular.
Well, what does all this mean? To me it indicates that the best indicator of processor sales during the last 2+ years has been AMD's capacity to produce processors. Further, there is every indication that had AMD the capacity to produce say 13M processors today, all those processors would be sold. So from that point of view, yes, AMD has been capacity constrained.
Does this mean that with fab36 coming online we can expect an immediate tripling of production? Obviously not, but I bet there are a lot of people at AMD trying to figure out just how many processors AMD can make and what the ramp should be to maximize revenues without hurting ASPs. Because AMD has been successfully selling everything it has been producing for the last 2+ years this is uncharted territory that AMD will more than likely move aggressively, but with caution into.
OK, a couple of things concerning the market today and that of a couple of years ago. 2+ years ago AMD had virtually no presence in either the server or laptop markets and the presence it had in the workstation, desktop market was marginal at best. So the problems AMD had back then mainly had to do with getting OEMs to adopt Athlons/Opterons and create lines of products. This proved to be a daunting problem that resulted in a very slow uptake despite the obvious advantages that AMD processors presented. Today, the situation is much different. Even though there is still a great need for the OEMs to expand their lines and add new AMD based products, that is not where the battle will be fought. Today AMD has the name recognition and presence to attack INTC on a product by product basis. This means that the next battle will be won by taking sales from INTC and not by introducing new products.
So far there is nothing that indicates that AMD is anywhere near meeting market demand for AMD processors. So what I expect is for AMD to push its' new capacity into the market as fast as it can, based on OEM acceptance and INTC counter measures. Next year should be very interesting. Next year at this time we should have a much better idea just how capacity constrained AMD is right now.
AI
