Technusatsu,
Do you think that the cancellation of Itanium is a strong possibility? If not, then what's your point?
If McKinley does not deliver, AND is not delivered shortly, I think it is a realistic possibility. If it happens, it would happen as part of house cleaning that would go together with departure of Craig Barrett.
Another would be for a major vendor, like Compaq pulling the plug on the flow of red ink, which could start an avalanche of cancelations, and changing the attitude from actively working with Intel, funding the Itanium project, to something along the lines of wait and see.
I think there is problem in the fact that for Intel, Itanium (Merced) has been downgraded from a potential money maker, to just a test platform for the software development. The customers know it and are not buying any system. The hardware vendors are stuck with no revenue, and all costs - not an attractive proposition, since I assume that the differences from Merced to McKinley will be non-trivial, and only a fraction of "learning experience" with Merced is transferrable. The rest of the costs are money lost forever, if McKinley makes it.
I am *not* saying that Itanium will get canceled. All I am saying that there is a non-zero probability that it will happen. Suppose the probability that Itanium will be canned is 25%. If the cost to settle the litigation if Itanium does get canceled is very modest $10 billion, the expected value of the settlement is $2.5 billion.
Every day that goes by, while Itanium is MIA, the probability of cancelation goes up, and cost to settle the litigation goes up as well as more money is spent on Itanium projects. So multiply the increasing probability by increasing cost of settlement, and the expected value of this liability starts to grow fast.
Joe |
