wbmw -
That's a pretty common occurrence in the technology industry. Better to go tried and true than to experiment with something different and fail. AMD took the safe approach and they succeeded. And the people who laud technology "innovation" are usually the ones who applaud AMD's mediocrity.
You raise an interesting point. Assuming that IPF is the innovative architecture for which you speak then one has to differentiate between experimenting with innovative technology vs. betting a good chunk of the company on unproven innovative technology.
Very little true innovation exists in this world. I agree that IPF is innovative, just not very practical for most things. I also agree that AMD64 is not innovative, rather incremental improvement on existing technology. (That is always how it has been positioned, too.) It is, however, the solution which works better for most people.
That's a key difference between science and business. From a scientific technologists viewpoint, IPF is much more interesting. Just wasn't very good business.
BTW - that does not make AMD's solution "mediocre". It is a good solution for the market. Transmeta was mediocre (even while innovative), IPF was bleeding edge, and AMD64 is that good solution for the market.
Market Data 
Markets 
AI
