The problem with your idea is that I can come up with various reasons why it would fail, and from the vague intro, I can't find a single reason why I would want to play it.
A 'good' idea should be the other way around: it should make me curious about the mechanics so much that I'm hardly looking at the risks, and more importantly, it should focus on a scope that is realistic.
Think of it this way, what are you truly attempting to accomplish to make this experience worth playing? Can you paint a particular moment in gameplay that would be 'awesome' to experience? List 2-3 of them, and tell me why this is awesome.
Then, look back at the idea up there, and start to think at what you can do without.
One obvious culprit jumping at me here would be the AAA highly polished graphics. It's clear from the above you don't have the experience and financial means to achieve that kind of production value, but that doesn't mean you can't make a good game.
So, can you make a good game with 'decent' graphics? And by decent, I don't mean ASCII graphics mind you. There's a wide layer of quality levels between top-notch AAA and crap and you need to figure out what's good enough to support your game.
One interesting part of your project is the episodic element. It requires some work upfront so that you can push additional content, but if done well, it allows you to segment your production costs and post-pone them to after your release your first few episodes which means you won't need to put up the full budget upfront and can always gauge whether you want to invest in a new episode based off sales you've made, etc.
Another element jumping at me which is currently vague and would require a lot of attention: what is a low-cost but highly engaging way for you to convey your story? It seems to me that the focus of your game is narrative content delivery. The above post does not explain very thoroughly how that would happen, and I would suggest to avoid cutscenes as they are likely to end up costing you the bulk of your budget. At the other end of the spectrum, wall of texts would be boring.
If you can find an original and efficient way to pass along your content to the player, then it could make your game.
Take a quick look at 'Bastion' and that narrator voice-over. There's no denying the game is famous for it, and yet, it was almost an afterthought from the developer, but it was original enough because it reacted to the player's actions and granted them control over the narrative of the game (smash a few crates in-game and you will see what I mean). If you can come up with your own 'twist' to storytelling, you could have something more interesting, less deja-vu, and cheaper to implement.
Also, realistic physics: that's crap. Everyone has an idea of what they'd like that to be, but if you actually input that correctly within an engine, you'll see that this is very boring to play with and responds poorly to most control schemes. It's best to make fun physics which 'feel right' (as opposed to over-the-top) than to try to make them as realistic as possible in most situations. I've seen very few (if any) games that felt better with realistic physics, and mind you, I'm including actual serious simulation games in the lot.
This is directly tied to one of the 3 sacred Cs of game development, so it's an art of it's own and requires a lot of polish and there's no easy way to make it happen. What "feels right" will always thrump what "is right" in video games.
AAA Multiplayer. Is this a requirement to make this happen? Couldn't this be added later? What kind of experience do your colleagues have? And does that experience alleviate the costs of the time involved to develop a backend infrastructure, code and actually host production servers and most likely maintenance ops to scale up/down based on player flow, etc.? I could dedicate an entire article to this one alone: multiplayer is not "just another feature"; if your game does not revolved around this, don't include it. It will cost a lot more than you might imagine (and even more than 10 times what you might think after reading this!)
From a logistical standpoint, what you have to remember is that in order to build this kind of game, you would need a large cross-functional team, which unlike an indie 'basement-based' team requires a lot more logistics to function.
You will need offices (which I assume you already considered), computers, a server for source control, at least one stack-machine server if you intend to test multiplayer features during development, an external test server to plan for content delivery (which would most likely have a CDN, loadbalancer, several app servers and one db to start with), a team of at least 40, but most likely 80 if you intend on releasing this within a reasonable timeline, a dedicated producer, 1 or more assistant producer(s), in-house audio development (or someone to handle outsourcing for you, which could fall onto your 2nd assistant producer), a minimum of 5 programmers (1 of which focusing exclusively on the engine assuming you are using a pre-existing engine, 1 on systems, 1 on UI, 1 on gameplay, and one as a generalist which is likely going to hit audio, etc.), a bunch of artists (3d modelers, a rigger, skinner, texturer, animators, fx, UI, level artists), a game designer, a bunch of level designers (assuming many areas), and given the nature of the game one or more writers (depending how text-heady the story will end up being).
If you plan on having voice overs, things can get a bit more messy, and I'd recommend outsourcing to an audio professional with a lot of experience so he can do the casting, and sit-in on the recording sessions.
And QA...
There's a lot more to this, but I don't see any point to go more in-depth at such an early stage.
