Just because a chip company has a chip to show the world that runs stable at 3GHz doesn't mean they can mass produce it in enough quantity to make it a real on the shelves so to speak product. This is true for the blue team and green team, fanboyism doesn't come anywhere near this.
They made another mistake yet again, they should have shown benchmarks if they claim it to be so good because analysts will not change their minds with being fed just hot air and quite frankly analysts determine in a big way AMD's share price. If AMD wants to spark interest from investors giving it the cash it just needs, then being quite is not the way to go.
For all one knows, that there could have just been a fake example of a 3GHz chip. All AMD had to do was modify ala hack the DLL"s responsible for the Vista Performance Index application to misreport the CPU speed. Sure this would be a very risky and stupid way to do marketing and try and get attention but hey if one's desperate then it's a viable option, so long as it never comes out in the public domain. Not very a professional means to get the world's eyes on your product to say the least.
The bottom line remains quite simple really, AMD will not be launching Phenom or Agena, the same chip basically at anywhere near 3GHz. Think 2-2.1GHz max which means that yields just aren't good.
What looks impressive at first, when examined closer, can many times be nothing more than a hoax of some kind.
It's quite easy for some smart guys to write up some PowerPoint presentations and impress world with new buzzwords and paper grown technologies that aren't even ready yet.
What AMD needs to deliver is rock evidence and not some hot air and number trickery since it's not scientists they're trying to impress but the public domain which quite frankly wants just the plain truth and nothing but the truth in easy to understand and confirm ways, ala benchmark data.
The art of a great product is to give the public want they want at the time and price they want it.