Also, there are lots of 2d RPGs out there, though it's true they're mainly in the indie scene, and those projects you linked aren't exactly indie studios. There's different levels of expectations for indies.
I was thinking down the lines of, can C2 be used for truly high-quality, large-scale projects? Right now, C2 has no flagship game which really highlights its full capabilities, so as someone who's still quite a newbie when it comes to development, I probably have a decent idea what might theoretically be done with C2, but on the other hand, without that shining example I might be underestimating it.
I can't think of a reason why C2 couldn't be used for hugely successful games like SpaceChem or Braid, but the truth of our current state of affairs is that virtually all of even C2's best games would fall into the above-average hobbyist range, like No One Has to Die, AirScape, and Super Ubi Land. That's not a knock on those games, because they're all legitimately strong games, and they were made by small teams or just one person (like Braid). (And hey I just ran across Magi, which is really delightful.)
From what I can tell, C2 is ideally suited to these sorts of games. No one's yet made an RPG with C2 that anyone would consider buying, while I can easily see an expanded version of AirScape making silly amounts of money in independent sales or on Steam or XBox/etc.
T:ToN and PE both use pre-rendered backgrounds with true 3-D characters and creatures, which C2 might be able to duplicate some of, but I'd just have to see a playable demo (or have Ashley or Tom promise it's so) to believe it's possible or practical to make a game on par with even Planescape: Torment, Fallout 1&2, or Neverwinter Nights. C2 doesn't have to be the perfect tool for all games, just the best one for some.