Disliking Itanium and declaring it dead are both fashionable hobbies among computer enthusiasts, but I can't help wondering how many of the naysayers are truly speaking from informed opinions based on current Itanium performance. Intel definitely bears a large chunk of responsibility for Itanium's negative reception—not only was the original Merced years behind schedule and vastly over budget, but Intel had marketed the CPU (and IA-64) as the computing standards of the future. Itanium's performance at introduction was underwhelming to say the least; the chip ran hot, sucked power, was enormously expensive, and performed like a dog in x86-emulation mode—which was, understandably, what the enthusiasts focused their attention. Itanium quickly became known as Itanic, and IA-64 hasn't gained much prestige (or respect) since.
With new, dual-core chips on the horizon, a common Xeon-Itanium platform coming soon, and the x86 industry's focus on parallelism as a means of boosting performance, I'm not sure I'd count Itanium (or an IA-64 based-design) all the way out. Itanium's current entrenchment in the niche "big iron" market might simply reflect the market Intel should've targeted the CPU at in the first place. Certainly x86-64 is here to stay, but might Intel's Itanium position reflect a long-term game with a payoff five to ten years in the future?