It's time to start shifting a few safe R states into the likely R column, if you think any of them are capable of cracking. For my part, I'll suggest Alaska. I was tempted to add Missouri and Montana, but this could be something of a fluke.
Don't get your hopes up on Missouri.
Add Montana to the list, though! And Utah too.
Elliott Morris has suggested that there may be a partisan non-response effect occurring (i.e. Republicans are discouraged and less likely to answer polls), in which case this would likely be a temporary bump. But it could also be the start of a real shift.
Well, then the question becomes: will they be discouraged from voting in November? It really does seem like, in the polarized era, the swing voters that matter are not those who swing between parties, but those who swing between voting and not voting.
