|Scott - I disagree with your analysis.|
AMD only has 65nm capacity. They aren't going to waste that on chipset silicon. The chipsets and graphics come from foundries which have to make their profit too. I don't see much of a market for highend graphics in lowend systems, which is the niche AMD is forcing themselves into. Intel's chipsets and graphics are done in house on older depreciated process technology. Intel brought that inhouse because foundries are problematical, to put it nicely.
As for market share, AMD does not have a compelling product in any market segment with significant profit potential. All they can do at this point is try to under price Intel in the low to midlevel market, a strategy which has had them losing many 100s of millions each quarter and there is no end in sight. Even if AMD could manufacture their quad cores, they underperform even Intel's 65nm products and if the recent rumors of the barcelona B3 step still having the TLB bug are true then no one knows when it may be fixed. Even if fixed it is still a huge very expensive die on a poor yielding process. Speaking of process, 45nm is now apparently slipping to 2009. If that's not enough, without HiK-metal gate they will fall further and further behind on both performance and power consumption. The much heralded "asset lite" strategy fails to mention that AMD is restricted in the x86 volume they can contract out. That means AMD has to maintain their own fabs and develop their own process with help from IBM. Their loan guarantees for the German fabs have full employment clauses as well. That's enormously expensive but no alternative is legally available. Bottom line is that without a compelling product portfolio and a competitive process/manufacturing infrastructure, there is little to support the view that AMD will be gaining market share any time soon and those loan payments have to be made from profits that just aren't there. It would be to your benefit to look at the whole picture.