I don't think IA-64 can be dead...because if IA-64 is dead, then INTC is dead, assuming they can't pull some magic alternative out of their butt (like buying everybody else in the industry that builds chips).
Intel simply has too much riding on it. This business of pushing out 32-bit out-of-order multi-pipeline processors that can still execute the 1976 "lahf" instruction is just not quite going to hack it for much longer.
Intel has to get back into the game with IA-64. It's just a question of when and how well, and how far behind the curve they'll be when they finally get something on the market that works well enough for OEM's to base competitive machines on. As I've said before, you can never count Intel out...unlike Microsoft, they're actually technically proficient. Plus they're well-managed, well-connected to developers and box-builders, and flush with cash (for the moment...I personally don't see how their revenues can possibly hold up selling Celerons and Pentiums). But IA-64 is much harder than anything they've ever tried before, and it's holding true to form: it's in the league where you have to throw one or two away before you get it anywhere near right, and with that baby, throwing one away takes years. Meanwhile, SUNW moves forward; late with its own UltraSparc III, it will nonetheless have a giant lead on Intel.
The faithful (i.e., those companies whose brilliant managers have made them totally dependent on Intel and Microsoft while making themselves comfortable on shareholders' money) will start pumping out IA-64 boxes long before McKinley (or whatever it actually is) works well enough for the laughter to die down. Intel will be given plenty of rope.
I don't pretend to know: will the Profusion chipset and pumping up the volume on razor-thin commodity margins keep Intel from bleeding out before IA-64 becomes competitive?
This will be one of the more interesting business stories of the early millenium. I guess it's no secret where my money is.
Regards,
--QwikSand |
