I would like the suggest the Itanium was a huge business success. At the start of the 90s, there were many decent workstation CPUs: SPARC, MIPS, PowerPC, Alpha, and PA-RISC. Intel CPUs were not considered in the same class. Big engineering work was done on workstations powered by these chips. But as the decade went on, the chips became vastly more complex and expensive to design. Near the middle of the decade, it was becoming clear what the trajectory was and companies started looking around trying to figure out how to navigate this. On top of this problem, Intel/HP teamed up to design a processor that was quite beyond (in terms of complexity) anything being designed at the time and it became clear competing was just not economically feasible. So basically the industry folded. In one move, Intel took out entire industry of companies that could compete. PowerPC is still there, but is marginal. SPARC, PA-RISC, MIPS, and Alpha design teams are gone.

Interesting technical tidbit - I remember an Intel engineer giving a talk at Stanford about the architecture and Hennessy came to listen. At the end of the talk he asked why they thought moving so much complexity to the compiler would work when other (including John) found it to be problematic? They didn't have an answer and I could see the look on John's face and at that point realized they might not understand what they are getting into.