Very well stated dude, my contribution:

- so besides applying butter on ati's stair handles -- not the steps though -- nvidia may absorb uli's tech which is more polished; good!
- slamming the door on ati makes nv feel big and have laughs w/cigars but decreases their revenue by a tiny bit; maybe good!
- ati is pushed into developing their i/o shit; bad for nvidia!
- ati may end up paying more for via/sis subtitution bits and even if not, MB builders are forced to redesign and make new releases on existing boards due to diferent components; nvidia laughs their lungs off!

Kaz,

You know as well as I do, that perception is more important than reality. While the differences between the ATI and competing chipsets may not be very important in actual use, if it is perceived as lacking "important" features like SATA-II, it is doomed. Does this matter in the real world? I am inclined to agree it does not. Will people demand it? Of course, that is why you do not see this part used in a lot of motherboards, and instead the use devices from other makers.

It will put ATI in disadvantageous situation for the short term, but ATI would not even consider leaving the market for this reason. They will simply make it a greater priority and close the gap. It is not something so complex like designing a competitor to the Merom, it is much more straightforward and they will accomplish it.

There are very few instances where not selling to your competitor, or supporting their products, works well in the long term. It's normally a short sighted policy that costs a lot of money. Let's use Intel as an example. They decided not to support AMD platforms, and made it impossible for AMD to develop for their platforms (it was part of the legal settlement). In doing so, they were successful in forcing AMD to use less stable chipsets and be a less stable platform, so it was worth the loss in chipset and motherboard revenue. But, how about now? How much money could Intel be making motherboards and chipsets for the AMD platform? Not a huge amount by their standards, but still a decent chunk of change.

I never liked it when the corporate hierarchy dictated what each branch should do, to avoid competing with other branches. If your company does not, someone else will, and they will end up with the money. IBM is a perfect example of failure, both in the hardware they created which was overpriced and sometimes artificially slowed down so as not to compete with the more expensive parts. OS/2 was initially made for the 286 for that reason and the fact IBM was doing much better in the 286 segment of the business than the 386 (where Compaq was, at that time, actually selling more machines). If OS/2 had been made 32-bit from the start, it would have been what we use today. If IBM had aggressively gone after the PC market rather than tip-toe so as not to compete with its other branches, it would certainly have done a lot better than it did.

This type of corporate behavior that Nvidia is doing is strangely common and predictably negative for the company. Cooperating with competitors is usually good business, because if you do not, someone else will or they will do it themselves and shut you out of revenue.

Is there anyone that has any doubts ATI will solve this problem? It will just end up costing Nvidia sales, and at the same time get them ill-will from the industry. It is just a bad move.