Your right, at least for the next 6 months or so. The first gen K7 is on the same lousy .25u process as the K6-3 (but a much larger die size) and AMD will not be able to get any volume out the door. INTC is holding back the 600Mhz Katmai waiting for AMD to once again fall on their faces on deliveries (expect to see 50-100k units total shipped this year - max). K7 is also not pin compatable with P3. How many board makers are going to invest in K7 design if they can't get any volume. This may change next year IF AMD can get the K7 shrink out on .18um, but that's assuming they can even stay in the business that long (and that they can get yields). Expect INTC to announce 600mhz Katmai after AMD drops their pants to the world and proves once again they can't deliver. INTC will get a premium price for the Katmai when that happens.
taps into its production agreement with IBM, watch out. hopefully they will do this as the ASP for the K7 can absorb revenue splitting with IBM. who knows, maybe IBM has already started to pump out K7's for the K7 launch. top k7's are expected to sell for in the neighborhood of $1,000.
10 years ago IBM had an awsume manufacturing capability in chips. I'm not sure they still do. They do tend to be among the leaders in announcing advanced R&D, but what are the costs behind this "new capability" (IE Cu interconnects), that they like to tout. IMHO, Intel is just as advanced but does not announce when the first chip functions, but rather when the process is both economical and the scaling requires it.
Since IBM shut down all most of it's real manufacturing fabs (Fishkill etc.), does it really have the capability to support AMD in a cost effective way? I really don't know if they can do it even if the AMD process is robust (and I have serious doubt's about that).
They used to be the leader, but they may end up being doomed if they jump on the AMD bandwagon. Only time will tell, hope you're wrong, I'm hangin' way out there on Intel.