Doug, there's an awful lot to be said about the profit value of niches, ask IBM :-). If you go and try and configure a HP Server, like I have, you find a base 2K 1.5 GHz Madison soon becomes an 8K option in your Server which is another nice markup for HP. Now the 64 billion dollar question is whether Itanium can stay niche full enough wrt Opteron to continue commanding these premiums. With Sun in the fray I have my doubts long-term but the combined might of Intel's,HP's and Dec's Architects do have a strong a chance of any of keeping Itanium ahead of Opteron in RAS, Server CPU count, FP and any other distinguishing features that will allow Itanium to establish, like the venerable IBM Mainframe, a long-term niche. An Optimist like Chipguy would go further and argue that architecturally Itanium is strong enough to start taking it to x86 in the desktop space eventually. Interesting times ahead but the chances of Intel canceling Itanium are very slim I would say, the alternative is only sharing x86 with AMD forever which would gall me if I was in charge of Intel and while I had billions in the bank I wouldn't countenance it. AMD has it tougher, it not only has to beat Itanium and Power but it also has to beat Intel in x86 as well first to survive. Definitely survival of the fittest in AMD's case which is probably why Opteron is such a potent server cpu, it's borne out of decades of ruthless desktop competition.
News 
Market Data 
Discover 
