I've produced some long detailed posts about my view on AMD in the past about their current situation but to sum it up I would say they have, and still are kicking themselves in the foot and therefore keep on suffering from a dead leg.
They are slow to react, pointlessly over hype something yet manage to virtually ruin their share price, have no clear vision how to get out of this mess, are bleeding money, have no means to produce a better performing chip for at least 6 months and their confidence seems to be yet unchanged which leads to this dire performance.
It's basically what you get if you throw a bunch of ego maniacs all in one room all thinking they're kings and queens of this world but the fact of the money is they are, but only in theory.
Nothing to wait for unless you like a long ride. 2008 will be a poor year for AMD, there just isn't anything they can or seem to want to do. They have been slapped in the face and act as if the wound is still sour so can't move on and setup a true fight by getting their act really together. 45nm won't rescue AMD over night. They will just add more cache and leave the northbridge/L3 cache at 2GHz so the net result will be 5% performance boost per clock overall/generally speaking.
It's both striking and amusing how AMD thought they could pull the Phenom off at 65nm, I could laugh over a few beers with the big buys at Intel all night as it really makes your eyes pop out in amazement. For those unaware what I'm talking about. AMD always had problems with their 65nm, X2's till this day top out at 2.6GHz for this very reason. AMD thought they could do magic and considered making Phenoms will be easier. How? Perhaps they had a bit too much to drink on the night they made this decision. Phenom is a massive 281mm2 die size. If you have problems making a die 50% smaller than this at 65nm, there is no way in hell you'll have less problems making one that's 200% the size. AMD thinking they're a magician decided to go ahead with the monolithic idea anyway and bingo, yields are average at best. On one side you get a bunch of dies that have at least 1 core damaged and on the other, you get dies that won't clock above 2.4GHz unless you feed them a lot of voltage (a clear indicator of bad silicon).
AMD never wanted to make tri-cores, they just don't want to be left with all these foul dies that's lost money for them. They need to recover some of that at least, even if only to on the surface try and impress investors of their smart thinking and devotion to making good business.
Folks who decide these things are paid handsomely plus get great stock options and yet still prove they're not worth the money. No true MSc + PhD + MBA guy would make such a bad move unless he doesn't care for his career, at all. If whoever made this move at AMD isn't already fired, then it just shows how cocked up AMD management is.
If there is a time frame worth waiting for in terms of AMD, it's 2009 however I fear AMD won't pull out of this itself. It's going to be a buyout for them I reckon. AMD may be left with 51% control, time will tell.
Business has no sweethearts, they did wrong and now they're paying for it and it's only their fault.
The only reason AMD did well with the K7/K8 is that they bought a lot of knowledge with their buyout of Nexgen back in the 1990's, which the then K6 was living proof, a first in a series of CPU's we all know very well. The problem is they got over-confident in the process and forgot they're a small company.
Buying ATI, ahh well, too much too early is how to sum this up. They could have just brought in the knowledge to design chipsets. They don't need GPU's really to be successful at selling CPU's, unless you could the IGP's. I guess Hector Ruiz wanted to go out with a bang, and he surely will, time is ticking
Facebook
Twitter