To: dougSF30 who wrote (197406 ) 5/18/2006 2:59:10 AM From: Petz Read Replies (2) | Respond to of 275872 me: "AMD was selling at capacity for both Q4 and Q1" you: "Nope, explicitly contradicted in the Q1 CC." I am sorry, but you misread what Dirk Meyer said. SELLING AT CAPACITY == what I said == selling all the chips that could be produced, given limitations on all factors, internal and external. Note the "selling AT capacity" does NOT say that AMD did not have ENOUGH capacity, but that capacity and demand were balanced. CAPACITY CONSTRAINED == what you think I said == unable to produce as many chips as demanded due to internal constraints. Dirk:With respect to the first question, you know, in aggregate we were not capacity constrained for the quarter. There were occasions where demand for, as an example, leading-edge product exceeded our forecast. In our efforts to meet that forecast we were occasionally supply constrained, for example, on package substrates. In aggregate across the quarter we were not capacity constrained. The way I interpret "in aggregate we were not capacity constrained for the quarter" is that 1. Demand was equal to supply, it did not exceed supply. Hey, that's exactly what I said, "selling at capacity!" 2. But AMD could have sold more chips if there were no supply constraints. (Otherwise, a "supply constraint" would not be a "constraint!") However the amount of this possible increased (no supply constraints) production was probably small. So why were units down in Q1 from Q4? Because demand (in units) was down in Q1 from Q4, and AMD correctly anticipated this! To make the most of that situation, AMD decided to set their wafer starts in Q4 for vastly increased Opteron volume, especially DC. Of course this decision resulted in lower total units being produced, and AMD had to lower the ASP of Opterons to sell the increased production. Even so, even the cheapest Opteron is much higher ASP than the A64's and Semprons that were displaced, so overall ASP rose significantly and total unit volume dropped. No mystery here! BTW, it appears that the "supply constraint" Dirk mentioned affected Opterons and/or X2's ("leading edge parts"). Not sure what substrates he was referring to. As a consequence of the Opteron decision, unit share of desktop and notebooks was flat and server unit share was up by a huge amount. So there is no indication that volume was light because of an unexpected lack of demand. The lower total units and (considerably) higher ASP is a direct consequence of the decision to grab more server share while the grabbing is good. Intel units dropped even more, especially the high-ASP Xeons, with predictable results. And clearly overall CPU demand in Q1 was not quite as good as what AMD expected in January. Nevertheless, without the "supply constraint" Dirk referred to, CPG revenue probably would have increased anyway. The fact that days of inventory dropped 10% during Q1 is further evidence that demand was healthy as AMD gained huge $ market share. NOTE TO "DA THWED": Something Doug said ("is this faith-based investing") struck a chord in me. I believe in faith-based living and that does not allow for 20% of my life being spent reading and posting on this thread, consumed with making money. So I will limit my participation mostly to weekends for a while and occasional lurking. If I'm missing something important, send me a PM, I'll probably miss a direct post. Petz