There are new polls out of the Democratic early states and they aren’t especially good for Hillary Clinton. ARG has Bernie Sanders at +3 in Iowa. That's probably an outlier, but the trend is pretty clear: Clinton has led by double digits in Iowa since October. Now Sanders is suddenly within single digits in four of the last six polls.
In New Hampshire it's the same story, but worse. Clinton led without interruption from August to November, but now Sanders seems to be pulling away. The most recent survey, from Monmouth doesn't just have him at +14—which is impressive enough—but the crosstabs show Sanders' strength with just about every Democratic group: She's down -8 among registered Democrats, -25 among men, -12 among voters under the age of 50, and -6 among voters over 50. Amazingly enough, Clinton is now losing women to Sanders, too, by 8 points. Her general favorability among New Hampshire likely Democratic voters is good, 73-19, but is, again, dwarfed by Sanders, who's at 90-6.
So what happens if Sanders gets a surprise win in Iowa and then consolidates his position with a win in New Hampshire?
The possibility brings to mind the famous line from Admiral Isoroku Yamamoto: Sanders can run wild for two weeks in February, but then what?
The consensus is that the Clinton machine will grind Sanders down over. The last poll in Nevada had Clinton at +23; in South Carolina, the RealClear average has her at +40. Even if Sanders picks up a great deal of momentum, it seems unlikely—highly unlikely—that he will be able to push past these firewalls.
But momentum can be an unpredictable force in politics and at that point her entire theory of victory will be predicated on Sanders' weakness with black and Hispanic voters. That's probably a good bet. But maybe not one you'd want to put the mortgage on.