Usually I agree with Shamus. Speaking as a game development student myself, though, I've done my research on this, and...
#1: Game budgets haven't ballooned by a factor of ten.
The average game budget still sits around $30,million, with even the most expensive games of the current generation being just upwards of $50,million and rarely climbing over $60,million, unless you include the extensive and lavish marketing campaigns that publishers like Activision and Microsoft use to practically brainwash the public into purchasing their games. Those few cases in the top ten tend to include games like Too Human, which have ludicrously long development times due to indecisive design and bad direction more than the cost of the graphics in the game. People make assets, get forced to re-make them when the platform specs change halfway through production, the entire script changes about three or four times due to some whim the director has about adding a new character, the whole game gets re-built from the ground up because they didn't prototype it properly and decide that it sucks, et cetera, et cetera, et cetera. Final Fantasy 12 in particular suffered from a few of these problems, with Vaan and Penello being added in at the very last minute of preproduction and huge changes in the team's upper staff in mid-stream, and I imagine 13 didn't benefit from having cross-platforming foisted on it midway through production either.
#2: These environments aren't that impractical.
The average environment in Final Fantasy 7 is no more complex than any given Unreal Tournament 3 deathmatch map, many of which take place in extremely detailed, highly urbanized locales that have a way of faking a larger scope than they actually represent; something environment artists are or should be well-practiced at. It's not all that unreasonable to expect Square to be able to produce these locations, especially since viable blueprints exist in the artwork for the original game--not useable assets, mind, but blueprints. Speaking from personal experience working on environment art myself, I can tell you that the blueprint really is the hardest part of the work, even when you're working on HD models. What they don't have the blueprint for is all the stuff you don't see on the top-down perspective--the cityscape in the distance, that sort of thing, but as I say, if UT3 has taught us anything it's that you can make even a tiny map feel like it's part of a big city with some clever fake-jobs.
Meanwhile the overworld's an antiquated means of exploration at this point, Square would probably have a rough time finding a way to make that look good, but most folks probably wouldn't care and it's kind of their jobs to figure out how to solve problems like this. Frankly, on the scale of design problems, "how do we make the overworld not look like ass?" is pretty low on the totem pole, somewhere beneath "what's a good alternative to this crappy planet-scanning minigame?"
A lot of Square's problems developing environments, I think, stem from their overdependence on unique assets. If you haven't heard, basically they make environments the same way they do characters--they just model it from the ground-up to be completely unique. This worked well on the PS2, giving games like FFX and Kingdom Hearts a lot of memorable setpieces, but it's a terrible pipeline for making any volume of content. Other developers use "whiteboxing" to block in the layout of an environment first, balancing it in sheer level design terms while environment artists work on the pretty parts. Usually these consist of a lot of reuseable assets: cliffsides you can stick up against walls, rocks that you can shrink and grow and rearrange to create pebbles and boulders alike, smokestacks you can stick through rooftops, chunks of houses, segments of skyscrapers, repeatable textures that tile seamlessly, et cetera. Using this scheme you could easily re-create a lot of the environments in FF7 in no time at all; I can literally pull out my strategy guide and pick apart all the modular pieces I'd have to build in less than an hour. Building them, texturing them, and assembling them into a single one of those environments would take me a few weeks, but probably no longer than it took them to produce the original environments for FF7 with the junk tools they had at the time.
You can see where Square tried to pick this technique up in some places with FF13, but they didn't get creative enough with how they did it, opting to repeat huge chunks of their environments and often having them just floating in the middle of skyboxes without any reference points. It's as lazy an environment design as I've seen--and I've seen student environment artists at work. The fact is Square hasn't updated their pipeline in many years and it crippled their production big time. If they pursued making environment art the way they did in FF13, where I suspect they had to reassess what they were capable of multiple times after beating their brains against a wall trying to force environments the way they normally do it rather than adapting a better, smarter pipeline, they probably would take 40 years.
#3: So, I'm a student. You don't have to take MY word for it...
Maybe it aint worth much since I've never been on a production this size myself, but let's play the compare/contrast game with some of the other companies in the industry and what they've been producing lately.
Square had a team of 300 people for Final Fantasy 13, and it took them 4-ish years to make it.
Bioware, with comparable resources, produced both the Mass Effect games in roughly that time. Each game was fully voiced and developed far more complex interactions than the likes of any game Square has ever produced, creating a game of equivalent scope to any Final Fantasy.
Naughty Dog, with about a third of the resources, produced both the Uncharted games--which, while modest in scope compared to the likes of a Final Fantasy game, won unprecedented accolades for technical and artistic achievement with respect to its environments.
Finally, Ubisoft created nearly all of Florence, nearly all of Venice, the entirety of Tuscany, the hills surrounding Tuscany, and more, all while making these venues veritable jungle gyms of interactivity, populating them, and furthermore bringing that population into the gameplay itself.
Each of these examples proves something about what modern game development can achieve that Square didn't come close to with its latest entry into the Final Fantasy series--something I think you'd have to admit even if you enjoyed the game. I'm not saying "I can do better," I'm saying these three companies and others did in fact do better--each twice over, in fact--in the time it took Square to make FF13. With all due respect they have proven themselves decidedly inefficient at producing games compared with any of their competition, who've actually increased their scope in many ways as opposed to scaling back. In the case of Insomniac, the Ratchet and Clank Future series actually has more levels and content than its PS2-based predecessors, if not then just as much. In the case of Naughty Dog, Uncharted may not have the sprawling exploration-based platforming environments of the Jak series, but it has easily stronger content where it's lost a bit of volume. As much as we wax nostalgic about the games of yesteryear, I'd say we have a lot of good examples here as to how developers have managed to top themselves. Developers other than Square, I mean.
#4: Who the hell said an FF7 remake would have to be on the PS3?
What's wrong with putting it on PSP? Or, hell, even PS2? As long as fans can play a version of the game that doesn't have block-men running around I don't think they'll be all that bothered. Shoot, Final Fantasy 8 and 9 pretty well approach the PSP/PS2 standard (just with a lot fewer polys and a lot lower-res models) and have the same scope, I don't see why this wouldn't be reasonable. If you need to sacrifice stuff, sacrifice stuff. Ditch voice acting. I don't think anybody liked Cloud's voice actor anyhow.
Honestly, though, we have every reason to expect better out of Square and we have every reason to think that an FF7 remake is completely doable. The pipeline's there, they just don't use it because they don't want to spend the effort to teach their artists a new way of working. The precedent's there, they just can't live up to it. The budget's there, but they're making projections from hugely flawed productions and ignorance of their competition's techniques. We have every reason to expect this re-make can happen.
Just not by Square's hands.