"But I'm skeptical as to whether or not this can allow AMD to reach 30% given Intel's fab expansions "
"Okay, the deal with UMC can potentially increase AMD's unit production by as much as 25%."
Actually, the figure is 20%. From your own transcript of the the Analyst's conference for 2001
"<Jerry>Relative to our agreement with Intel, the strictest interpretation is that 20% of our microprocessor can be manufactured in a foundry, and that is an arms length foundry arrangement. Relative to production in a jointly owned, joint venture, arrangement, it is in proportion to our ownership. ".
jc-news.com
As I have stated before, 2002 is going to be the bloodiest pricewar we have seen in CPUs unless the market picks up a lot more than anyone is projecting. Both AMD and Intel will have much more capacity than the market can absorb. So either they will build up inventories, or they will be slashing prices like no ones business. All indications so far is for price slashing. Having said that, I suspect that AMD can still grow their market share, at least they managed to do that in 2001 in the face of very stiff competition from Intel. Ok, Intel will be able to cut their cost by both 0.13 micron and moving to 300mm. But the problem is that most of Intel's costs are not production related, but structural related. AMD has much less structural overhead. At a guess, even with flash being at zero, AMD can make a profit if their ASP for processors is at $80. Intel, OTOH, needs probably closer to $150, with the same rules. If there is a pricewar in 2002 as it appears like ther will be, the ASPs for both companies will be closer to $80 than $150.
So the question will be, how low will they go? Flash is recovering, but that benefits AMD more than Intel. With MirroBit, Intel will no longer have the lowest cost per effective bit, even with the smaller geometries, unless they release the StrataFlash variant with 4 bits per cell. Can Intel make enough money off of the areas where AMD doesn't compete to finance a pricewar in the areas where AMD does compete? That is the only question for 2002. Intel doesn't have a lot left, they don't have much left on the desktop, where Intel will likely have the performance crown through most of 2002, but only at the very highest speed grades. But the 1 and 2 processor workstation and servers are vulnerable, as are mobiles. Now true, more than 2 processor SMP stuff will be theirs through out 2002, but there isn't all that much volume there. The brutal mathematics is that unless Intel is willing to run a huge loss per quarter, there is a limit to how much they will be willing to push a pricewar. Unless AMD screws up the transition to 0.13 micron, that is.
So AMD is much more likely to be production limited than Intel is. And that means AMD can reach 30% or more of the market unless Intel is willing to bleed great rivers of red ink. If Intel is willing to run those sort of losses, then both companies will be in serious trouble. |
