ABC News has projected Trump will carry the state of North Carolina and its 16 Electoral College votes. Although Harris was not likely to carry that state anyway, the projections still hurts her; as you can see using our handy-dandy election simulator, when we restrict the set of possible Electoral College outcomes to ones that include a Trump win in North Carolina and Florida, Harris loses ground probabilistically because those outcomes are likelier to occur in simulations with more Republican victories than not.
Internally here at 538, we have also been running a version of this model that updates itself with a likely win in Georgia, based on how few votes are left there, as well as a tight race in Pennsylvania, Michigan and Wisconsin, based on the lagging Democratic turnout there. That model assigns a 6-in-100 chance of Harris winning the majority of Electoral College votes after everything is said and done.
To be clear: That 6 percent chance is not nothing. It's possible that our models are extrapolating incorrectly about outstanding urban votes in Wisconsin, for example, where about 24% of the vote is left to be counted, or inferring incorrectly the result in Nevada, where no votes have been counted yet. But those states don't have enough votes to put Harris over the top if she loses Pennsylvania, anyway. There, 84% of the vote has been counted, and Harris trails Trump by 3 percentage points.
