The other challenge for MMOs, of course, is moving lots of PCs around at once, which makes all this stuff a lot harder to do at high-quality than in single player titles.
This is a huge problem, but not on the quality level. Quality is on the client side, and even average game rigs in the 2010s will be able to support this level of detail for 100s of avatars. The problem is the N^2 interprocess communication problems that stress the servers and the network connections. But those are also problems that are being solved by Moore's law and by faster broadband rollouts.
In the past few years launching an MMO has seemed like nothing more than a good way to lose a lot of money, with only WoW and your own EVE being successful launches in years now.
Actually EVE predates WoW. But there has been a noticeable change in the MMO market in the past 24 months. We now live in an era where there are between 1 and 2 million "MMO Hobbyists" - people with allegiance to no game, but to the genre. They switch from AAA to AAA MMO as they ship (virtually all of which are Theme Parks), burn through the content faster than the developers thought possible, and then quit with complaints that the game is "boring" or "too short". This creates an "s-curve" of player participation which is actually really bad for MMOs - they have variable costs that scale with the size of the player base (servers, bandwidth, customer service, GMs, billing, etc.), but they take much longer to ramp up and ramp down than the length of the initial spike.
Star Trek, Conan, Warhammer, and Aion all suffered from this problem spectacularly.
One of the things the next generation of MMOs is going to address is finding ways to moderate this effect.
On the other hand, there are games that are doing quite well. Runescape has 5 million players (1 million paying $5/mo). Aion still has about 200K Western full price subscribers, and a much much larger business in Asia. Lord of the Rings and Dungeons & Dragons online have had a huge renewal once they went to a hybrid free-to-play/microtransaction model. The next AAA MMO likely to ship will be Star Wars: The Old Republic (but it is barely an MMO in my opinion and should probably be classified into its own special category).
The market for subscription based Western MMOs is between 5 and 10 million people (its very hard to narrow that down because many people play more than one MMO at a time). The market for some kind of F2P/MTX model is probably 50 million. So there is a lot of upside yet to come from this space - it's barely been tapped in some ways.
RyanD