IMHO, Intel just doesn't want to do the same error they did with Pentium 4, eg. release a processor with an extended instruction set when no application has been built to use it. That's what allowed AMD to grow on the market.
Furthermore, Intel Itanium has very poor compatibility with 32-bit applications, whereas AMD Athlon64 supports them natively. So releasing Itanium too early would once again mean poor performance compared to AMD, and potentially reproduce the P4 problem.
Was it about OpenBSD?
Oh sorry, the improbability of having a hole in the base system wasn't so high after all.
IMHO, Intel just doesn't want to do the same error they did with Pentium 4, eg. release a processor with an extended instruction set when no application has been built to use it. That's what allowed AMD to grow on the market.
Furthermore, Intel Itanium has very poor compatibility with 32-bit applications, whereas AMD Athlon64 supports them natively. So releasing Itanium too early would once again mean poor performance compared to AMD, and potentially reproduce the P4 problem.