Is there any real use to looking into the bitcoin system by those other then the deep pocketed able to afford the pricey GPUs required to perform decently?
Even if i did have 10 machines all running 5970's overclocked running 24/7, what type of return can be expected, and how can the GPU processing power of that type of setup be less than a single Nvidia 4x4 Tesla server rack running 24/7?
I don't consider myself "deep pocketed", though I will admit I had never spent more than the bare minimum on graphics cards until I discovered bitcoin. Back in February I did some sums, and worked out that I could afford to buy a second-hand 5870 for £130 (about US $215). That included tax, and postage was another £5 or so. I figured that the worst case scenario was that I'd end up with a really good (by my standards!) graphics card. Best case scenario was that I'd make enough bitcoins to pay for the card. I've been lucky, both with bitcoin generation and the exchange rate, and I managed to pay for the card within a month or so.
Even with recent difficulty changes, I've tended to generate around three blocks per "difficulty", i.e. I reset my mental clock each time the difficulty changes (roughly once a fortnight). This "fortnight" (the one we're in now) I've not generated anything yet. Last fortnight I got 3 blocks - 2 solo mining, one mining as part of a pool (though I didn't get 50 bitcoins for pool mining). The fortnight before that I only mined in a pool, and I found 3 blocks (and made around 100 bitcoins).
So, to answer your first question, it's not necessary to have lots and lots of expensive GPUs to mine successfully. However, you might consider spending your money just buying bitcoins, rather than buying mining hardware. It depends on your level of interest and your needs. I felt I could justify buying one graphics card, I'm not so sure I could have justified buying several (though, in hindsight, it would have been very profitable).
Turning to your second question:
this list and
this calculator are very useful. use them to try different scenarios. For example, you mention 10 machines all with 5970s. You can fit more than one 5970 in each machine, but lets ignore than and just consider ten 5970s. Each 5970 has a Mhash/s rate of between 530 and 645: lets pick 560 as a conservative value. Ten 5970s will yield 5600 Mhash/s (10 x 560, obviously!) which will, according to the calculator, provide an average of one block (50 bitcoins) every 17 hours 32 minutes at current difficulty.
...but that's not enough information on its own. You also need to consider when the difficulty will change, and what it will change to. We don't know, but
bitcoincharts.com has an estimate. It also has various exchange rates, which will help you work out what your 50 bitcoins every 17-18 hours are worth in your usual currency (the currency you probably use to pay for GPUs and other things!)
Hope this helps. Only you can really answer the questions you've asked, and even then you'll be making guesses based on what may or may not happen in the future. The best advice I can give is to not spend what you can't afford to lose: if you can afford to spend money on one new graphics card, and you're not concerned about that card paying for itself (and most people outside this forum probably never consider the prospect that a graphics card could pay for itself!) then a small investment in a GPU could be an excellent idea. If you're in a position to make a large investment, that could potentially be very profitable but equally could result in you losing the entire investment, then consider either buying bitcoins directly, or investing in several GPUs and the accompanying hardward (motherboards, power supplies etc).
