As bad as 2010 was for House Democrats, 2012 could be even worse. Republicans don't have a lot of exposure, since most of their gains were in red territory. More importantly, Republicans will control more seats in redistricting than they have since the states began regular decennial redistricting in 1972.
Using census estimates of where population is growing and falling within states, as well as Dave Bradlee's handy redistricting application, I offer my thoughts on how redistricting will most likely shape things in 2012. If I don't discuss a state, it just means that I don't see any meaningful changes occurring. To determine which states will gain/lose seats, I use Polidata.org, which sorts through the complex redistricting formula. Please note that these are based off of census estimates, not final numbers; in 2000, most observers were surprised when North Carolina gained a seat and Utah did not.
Obviously, legislatures are strange beasts and can do any number of things with redistricting. Some have estimated that Republicans will gain as many as 30 seats from redistricting. I see the most likely scenario being Republicans picking up a dozen or so seats from Democrats as a result of redistricting, although the upside on that prediction is probably with the Republicans. Remember, candidate recruitment and the national environment can cause a gerrymander to boomerang on a party that spreads itself too thin: Pennsylvania legislators thought they had enshrined a 13-6 Republican delegation; by 2008 it was 12-7 for the Democrats.