Like I said in my original post, AMD would know doubt be able to supply enough chips to Apple since its needs aren't that high. But we are talking about stability or perceive stability for an Apple base that has repeatedly not had it. To that effect news like this isn't reassuring for customers:
http://www.xbitlabs.com/news/cpu/display/20060126222840.html
AMD claims that the problem is lack of packaging materials which although it may be true, sounds kind of funny. Lack of packaging materials will effect Apple even if they don't buy retail box, because even trays need packaging. The article saids OEMs get priority, but like I said it doesn't bode well for perceived stability and marketing purposes.
Personally, the use of "because they can" as the reason for the AM2 delay doesn't seem reasonable. The problem with the reason is the fact that the AM2 delay puts its launch right on top of the Merom family launch. Even if we assume that AM2 will still maintain the performance lead over Intel, we can generally agree that the performance lead won't be as great as it currently is with the P4. AMD's power and heat lead will be similarly reduced or eliminated. This would mean that AMD will have to out market Intel. Now Intel may be a poor chip designed, but they certainly are a good marketer either due to good strategy, which can be debated, or by sheer amounts of money that they can throw out. Why would AMD put themselves in a position where they have to try to fight for the same TV time or newspaper space?
Marketing isn't the only issue. Intel can also outprice AMD if need be. AM2 will still be using the 90nm process while Intel will be using 65nm. Merom is also the 2nd generation on the 65nm and by Q3 the yields would have improved enough to surpass Intel's current 90nm process. It isn't just the processor either. By Q3, Intel will be double transitioning their chipsets from 130nm to 90nm and from 200mm wafers to 300mm wafers. The wafer transition alone increases die production by 240% and reduces energy costs by 40%. AMD may have better performance, but if Intel cuts their prices and the chipsets can easily sustain a 50% cut, AMD will have a hard time competing.
http://www.xbitlabs.com/news/chipsets/display/20051124214142.html
It would make much more sense for AMD to launch as originally planned in April. That way they can do it on their own terms and market at will. Even if they don't have the initial volumes at launch, but the time Intel launches a few months later AM2 would have been mature, the market would have accepted it, and the prices would have dropped and stabilized. AMD could then easily evaluated Merom's performance and release faster processors as needed on their already proven platform.
If there was a technical glitch or other hard reason to postpon the AM2 launch then of course AMD was justified. However, the "they did it because they can" argument just doesn't fly with me. AMD has the advantage right now. Why would they sit around, if they have the product ready, when they can just as easily lead the market?
Did you know that dual core Turions can be released at 2.66GHz and only consume 25-30W?
That's very nice and all, but could AMD have provided it to Apple sufficiently ahead of time last year so they could have launched in January instead of Yonah? The dual core Turion requires the new Socket S1, which we've heard the least of compared to AM2 or Socket 1207.
Please, I'm dying to see those benchmarks
In this case I was refering to Conroe vs. AM2 or Woodcrest vs. Opteron. The G5 tower won't be replaced until those roll out so I wasn't referring to the current Xeons which we know will lose. I've posted Dempsey scores before that compare nicely to the Opteron 280 so the Opteron isn't invincible.