From what I've read, the 9600 PRO is going to be not quite as fast as the 9500 PRO, though I don't know how anyone can know that for certain.
Why does ATI struggle financially despite having great products? Well, maybe GreatGrapeApe is right, and they are actually profitable, though I always take a company's "adjusted net income" figures with a grain of salt, since there is no standardized methodology for computing adjusted net income. I have had more success using the bottom-line figure of "net income," since this is a standardized, auditable figure, and accurately reflects a company's profits (or lack thereof) after deducting expenses from revenues. Ape is right--it is possible to be profitable through other means while showing a net loss, but it still is not indicative of a company on sound financial ground--especially if it is repetitive.
Now, that being said, ATI HAS been doing much better profit-wise of late, and this can be directly attributed to the success of the 9X00 product line.
Why does nVidia make more net profit on revenues? There are several reasons, and I don't know all of them. I do know that nVidia is a CHIP and CHIPSET manufacturer--they do not manufacture graphics cards or motherboards. That's advantageous for a few reasons--nVidia does not have to spend capital on facilities for mass-producing completed graphics cards. They create the GPU and a reference design, a few hundred samples, produce chips by the tens-of-thousands through leased foundries, ship those chips to companies like BFG, Asus, MSI, and Gainward, and those companies absorb the production costs, including staff, facilities, quality control, the vast majority of warranty claims, and a large share of the marketing costs.
ATI, on the other hand, does still produce complete graphics cards. Now, obviously, that is not helping their bottom-line, since ATI is trying to move its operations towards the nVidia model. Believe it or not, despite the animosity between fanboys, the two companies actually harbor some mutual respect. ATI would prefer to be a chip manufacturer, which is why we are seeing more and more third-party manufacturers of "Powered by ATI" motherboards, like Visiontek, Sapphire, and others. ATI understands that chip production is the most efficient and profitable portion of their operations. They even talk about it in their 2002 financial report.
nVidia also has continued to do well in the much-maligned budget and mainstream markets. The Ti4200 is one of the most successful GPU's in the history of 3D graphics. The mainstream is called "the mainstream" for a reason--because that's where the majority of a company's revenues derive.
Add to that the fact that nVidia OWNS the AMD chipset market so completely that even VIA's brand-new and much bally-hooed KT400A chipset STILL does not beat the nForce2--not even in single-channel mode! VIA is not going to be able to bring KT400A to the market in quantity for at least another month or six weeks, and by that time, we should be hearing about the nForce2-S, or nForce3 (or whatever nVidia calls it) which will feature further enhancements to improve performance, and will also feature integrated SATA on the southbridge. (At least, that's the rumor.) So, nVidia will own the Athlon XP--and likely the A-64--crown for at least another six months.
That being said, AMD must continue to be competitive for nVidia to have continued success in the chipset market. So that is ONE marker that nVidia owns, but can't really control.
And don't forget nVidia's $200 million advance from Microsoft for building XBOX core logic and integrated graphics and audio. nVidia also claimed a pretty healthy royalty share from XBOX sales. I can pretty much guarantee that nVidia has made as much if not more money off XBOX than Microsoft has.
Anyway, factor everything in, and what it amounts to is that nVidia has just been more efficiently run than ATI, and has effectively taken advantage of its reputation as the 3D chipset king.
Right now, however, nVidia is losing revenues on NV30--guaranteed. The enthusiast market may not be the biggest, but it does generally have higher margins, and it can be the difference between profitability and merely breaking even. nVidia did the smart thing by scrapping FX-Ultra and moving directly to NV35. The deal with IBM is also interesting. A lot hinges on how good NV35 is and how soon it can be brought to market (I'm hearing maybe as soon as late May-early June. NV35 has already taped out, so it isn't unreasonable, assuming IBM's foundry can produce higher yields...)
ATI has made a big impact in graphics cards over the last few months. In terms of investing, however, nVidia seems like the safer bet--for now. If ATI can bring R350 and R400 to market on-time and with sufficient yields, they have an opportunity to own the graphics performance crown for at least another year. If that's the case, ATI will show a real, bottom-line, on-the-financials, SEC approved profit this year--and then all bets are off. ATI could become a juggernaut.
However, if the rumors are true, and they are running into some of the same .13 issues regarding yields, then it changes the equation considerably. Don't underestimate the advantage nVidia still holds in the marketplace based only on their reputation. We're talking pure economics here--nVidia is a viable company and I expect them to challenge to regain the performance crown.
If I had to buy a graphics card today, I would buy an ATI.
If I had to buy stock today (thank goodness I don't) I would probably buy nVidia, and hope NV35 is a great chip.
In reality, though, I'm not real high on either stock, considering how soft the tech-stocks are right now.
I'm probably going to wait until Q4 to decide which way to go.
<-----Insert witty sig line here.