AAA-level games are about production values - and you're not going to see small teams pull that off very often.
If, one day, production values stop being about money - and become common place - then it will no longer be considered AAA, because what's common is not very interesting.
By such time, developers will need other ways to stand out - and that would be a welcome change indeed.
I guess we need to define the markers of a AAA game. Obviously, that's going to be a little fuzzy and have some exceptions, but we can probably agree on some generalities. This is one time when the drag from console hardware limitations might actually help PC gaming, even for games that are PC exclusive.
- Graphics have to be at least very good. This is actually where I'm looking for better development tools. We're already seeing "rental" engines popping up. If those get better and cheaper, they'll bring the indie guys up a tier from where they are now. The corporate-backed developers will always have an advantage engine-wise (either "renting" the best, or having the manpower to make their own), but the indies have to get to a point where their graphic limitations are minor.
- UI has to be user-friendly. As this is more an "engineering" problem than anything, the small guys should be fine, particularly if their graphics engine doesn't tie their hands.
- AI has to be decent. This goes back to coding, so it might be the hardest hurdle for the small guys to clear. I honestly have no idea how hard it is to develop and code a good AI. With the number of games that fail dismally at this, maybe it really is difficult to do. Doubt there are "rentals" out there for this, so this will certainly be tough for a mom-n-pop to keep pace with a corporate army.
- Game can't be 5 hours long. The good news is that the big boys are rapidly coming backwards, so a small team that can generate 40 honest hours of content might even go to the head of the class.
- Story can't be completely suckish. Creativity has never been the problem for small guys; if anything they've already got the advantage here. Mostly added to the list for completeness.
- Sound can't be completely suckish. Sound tools are already in place and there are lots of rental composers running around, so this shouldn't really be a problem. Still gotta have labor, but it's 1-2 guys for either team me thinks.
Am I missing anything?