@Phantasm to be fair, Intel tried to kill x86 at least three times. First, with iAPX432 in 1982, then with i860/i960 5 years later, then a decade later with Itanium. With the exception for iAPX432 (which was totally and completely insane), other attempts were sound. i860 was a good RISC chip and Itanium wasn't so stupid at the time. Compiler sentience and hindsight 20/20 aside, back in the late 90s it wasn't even known that CPUs would peak at 5GHz frequency. Intel could've been very well planing for Itaniums to hit 10GHz and above.
Check the graph attached. Before CPUs hit around 2-2.5GHz mark, it was thought the power requirements and therefore heat production were scaling linearly with the frequency, not exponentially. You can blame this fact for the failure of both Itanium and Pentium 4 cores. Itanium was initially designed with physically larger cores in mind running at much higher frequency than what they ended up with.
