AMD's Hector gazes into IBM's crystal balls

A FEW WEEKS ago, we told you about NVidia, Xylinx and a bunch of others running for the hills from the vast land of East Fishkill. It seems that they were all quite suddenly dissatisfied with IBM, and went off in search of greener pastures. The Taiwanese fab companies must have been very pleased.

Rumous swirled about low yields, miserable transitions, and other assorted nastinesses. IBM went from SOI innovator to whipping boy in about 10 minutes. It is curious that a company with so much R&D, so many talented engineers, and so much to gain would screw up so badly.

I mean, they did pretty much commercialize SOI, Copper and other techniques, and they make a mean Power chip, 4, 5 and PC. Apple seems to be doing a lot better since IBM decided to win back their business. So what went wrong for NVidia and Xilinx?

In a word, either nothing or AMD, depending on how you look at the issue. It seems that AMD decided that after the bang up work IBM did on the 90nm chips, it was time to cement the releationship. Ever wonder where the fully functional 90nm silicon came from at Comdex?

Yup, it seems AMD just monopolized IBM capacity, and NVidia and company were shown the door, politely most likely. Another factoid disguised as a quiz, guess how big the wafers are that the 90nm A64s are cut from? Hint, it is not eight inch, but you only get 7 more guesses.

So this raises a lot of interesting questions about marketing. First, what does Hector see in his crystal ball that requires so much silicon? Next, what does this say about AMD cutting down the die size of the Hammers? Both are far more complex than where the damn things are made, and I for one don't know. µ