It's difficult to guess at the long-term consequences of the ATI acquisition by AMD, but there are some things we can speculate on and other results are readily apparent. I, Jakub, will start with a company-by-company breakdown of what the deal means to everyone in my opinion below, Brandon will then follow up with his rebuttal at the end of this article. It’s a bit of a long read but hopefully one that you’ll enjoy…
The biggest winner in all of this is ATI. ATI now gets access to the financial resources and relative stability of AMD's revenue. ATI by itself is in a market with a brutally fast product cycle, where a 1-month delay is seen as a disaster, and being 3 months behind schedule actually is a disaster. The competition is cut-throat; a 10% difference in performance means deep cuts into prices or the prospect of sharply reduced sales. A company can go from champ to chump in a day - or vice-versa, as NVIDIA did with the 6800 launch after the disastrous 5X00-series.
In short, with the relatively constant and reliable revenue coming from AMD, ATI is in a better position. It will have first access to AMD's technology and thus almost certainly provide the best chipsets and video solutions for AMD processors. Though for the purposes of the article we call the company ATI, the fact is that it no longer exists and is now the graphics division of AMD. The obvious result of that is that Intel may attempt to stop ATI from making chipsets for Intel processors. Intel cannot stop people from putting in ATI/AMD discrete GPUs into Intel-based systems, of course.
In the short run, ATI can expect to benefit from AMD's marketing and maybe money - but AMD revenues should go down with the release of Conroe. Both companies lag behind their respective competitors in marketing, but AMD's size and influence can - properly utilized - give them a substantial advantage over NVIDIA in mindshare. The idea that AMD fabs can be used to make ATI chips is laughable, at least for the foreseeable future.
The gains in the long run all depend on how well AMD and ATI mix, how the cultures work together and if proper communication can be developed between engineers, marketing, and executives to take full advantage of their mutual capabilities. The possibility of AMD fabs making graphics chips becomes more probable in about five years, provided cash can be found to update or create a new foundry. However, it is what AMD does with ATI's expertise in graphics and chipsets which is key - can they really create a new, better x86 platform? The future is hazy, but there are doubtless some very exciting possibilities for the new company.
ATI execs and stockholders should be quite happy with their position in the future company as well. AMD has a market cap of about $8.5B, versus $6B for ATI. Though technically an acquisition, it gives ATI a lot more power than it would have under normal acquisition circumstances, such as if Intel was buying.
NVIDIA is the most likely loser in all this. About the only thing going well right now for NVIDIA is the timing - if this deal had happened two years before, when AMD was just starting its rise to the top of the enthusiast market, NVIDIA would have been in serious pain. As it is, with Conroe out so recently, NVIDIA's position is easier than it would have been.
What NVIDIA faces is the single most distasteful prospect a company can have: facing a direct competitor (the ATI division of AMD) who until recently was also the best market for NVIDIA products - both nForce and GeForce, since AMD has been on top of the enthusiast market very recently. Imagine if AMD had acquired ATI two years ago. nForce was already a big chip in NVIDIA's bin, and AMD was just beginning its domination of the high-end segment which drove high-end video sales. Could NVIDIA afford to cut itself off from the AMD market then? Not really. Even with Conroe around now, it can't now. So it ends up providing support for, and thus maintaining the popularity of, its direct competitor. AMD platforms are more attractive because they have nForce and NVIDIA graphics. The simple fact of having a choice is pleasing to consumers. While NVIDIA would no doubt love to pull the rug from under AMD, it can't afford to.
Conroe simply means that NVIDIA is less reliant on AMD doing well for the market to continue growing, but unless Conroe ends up burying AMD completely, NVIDIA is in an unstable position. As long as AMD can afford to make top-flight graphics adapters, NVIDIA's position is vulnerable. At the least, the company's independence is threatened as Intel becomes its preferred platform and NVIDIA finds itself more dependent on them, or worse, if Intel starts eyeing NVIDIA as a potential acquisition. NVIDIA has always been more aggressive than ATI, and it is doubtful that the board there wants to lose control, certainly not to Intel, where they'd be relatively insignificant players.
NVIDIA's position isn't hopeless by any means. The new company they face has a combined value only twice that of NVIDIA and revenues only about three times as high. Brandon believes, in fact, that NVIDIA's position may be no worse with the acquisition of ATI, but he'll detail his explanation later.