The Big Orange One has a nice map and post showing the amount of backsliding that could be tolerated with new census figures and the same relative political strength from state to state in order for Obama to win in 2012 (about eight percentage points). Obama could win while not securing Florida, Ohio, Virginia, North Carolina and Indiana, but would need to capture Iowa, Wisconsin, Minnesota, and New Hampshire (as well as Colorado, New Mexico and Nevada) to do so.
In theory, Colorado is the swing state, by margin of victory, although, like the author, I don't really think it will turn out that way. I worry more about states like Iowa and Wisconsin. The analysis of a potential shift in Virginia, in light of its increasingly Northern political and cultural climate is also appropriate.
Notably, Wisconsin's GOP Chair was just elected to replace Michael Steele as the chair of the RNC, so they're on to this plan. But, ultimately it comes down to who runs on the Republican side and how the economy is doing in the fall of 2012. If the economy improved in the next year and a half, and the Republicans nominate an idiot as they did in the Colorado Governor's race in 2010, then life is good and President Obama is re-elected. If the economy continues to stall and the Republicans nominate a solid candidate (the prospects of which seem rather dim at this point), then President Obama loses.