Yeah: everyone gets an egg in the face now and then, and I asked for this one. I need to stay away from the Inq. When you're addicted to information, though, it's like a drug: you get it where you can. :?
Anyway; I agree that it was probably in ATi's best interest to merge with Intel rather than AMD, however I'm not sure that I agree it would have been a better merger overall.
Before C2D, AMD's only soft-spot was mobile computing. AMD was gobbling up market share everywhere else. Intel was able to maintain such a stranglehold because (1) their mobile processors were able to hold their ground against AMD's lineup and (2) Centrino made integration simple, inexpensive (relatively), and very effective. AMD didn't have a chipset for mobile computing, and ATi did. A great way for AMD to truly compete against intel was to offer a fully-integrated solution; similar to Centrino.
Add the 'discovery' of the GPU's Floating-point (i.e.stream) processing power to the equation, and it almost seemed like a no-brainer for AMD to acquire ATi. nVidia was far from AMD's grasp in terms of price, so they didn't have many options.
Enter C2D. I'm sure this shook things up a bit, but at some point someone at AMD made the call to keep the acquisition moving despite the impending disaster (for lack of a better word). This is really the decision that comes into question.
AMD has finite resources: were they better allocated towards stream/mobile computing, or should AMD have held off on the merger and focused on maintaining the market share for which they so vigerously fought? Was it wise to dilute their resources right after their competitor came out with a product superior to their own?
My personal opinion is that the ATi acquisition, simply for the sake of mobile computing, would have been unwise. However, if AMD can use other ATi assets and IP and apply them to the mainstream CPU space, then it could have been worth the investment.
I do believe that HT does support the aforementioned theory that ATi technologies can be brought to the mainstream space (I'm still hoping for a GPU on HT). I guess only time will tell, though, since we won't know the reasons for the acquisition until it bears fruit.
At any rate, to go back to my original point, I don't believe that an Intel/ATi merger would have had equal worth. The Assets and IP that ATi held weren't nearly as valuable to Intel as they were to AMD (though in guerilla
tactics, Intel could have bought ATi just to prevent AMD from buying them
)