538 is revving up their state level presidential polling, and in their first election update the results provide an interesting window into the party realignment effect that the Trump candidacy may have. There's been plenty of talk about how Trump's unique populist appeal might redraw the electoral map, and the state polling shows what some of that might look like.
Trump does currently poll higher than Romney's 2012 election performance in a number of blue states (11.3% better in Connecticut, 7.7% better in Maine, 6.7% better in New York) though he doesn't come very close to looking like winning in any of those states. Clinton currently polls only 6% above Trump in Connecticut, where Obama beat Romney by 17.3%, but given how entrenched the Democratic lead is in Connecticut, even that large swing isn't enough to bring it into Trump's column.
In swing states, Clinton is currently polling slightly better against Trump than Obama performed against Romney, but the effect is pretty small. The only potential flips indicated so far for the polling would be North Carolina going to Clinton when it went to Romney in 2012, while Nevada currently has Trump out-polling Clinton even though Obama won it by 6.7% in 2012.
The biggest swing is in red states, where Trump is currently polling 9.1% worse than Romney's 2012 performance, though even this isn't enough to flip any states other than a current freak result of Kansas going from a 21.6% Romney victory to a 2.6% Clinton lead in the polls over Trump. (I'd be kind of shocked if Kansas goes for Clinton, but as we keep saying: it's an unusual year.)
Overall what we're seeing right now is a lot of movement in the levels of support within specific states, showing how Trump is peeling off some working class White support in blue states while losing various factions of the conservative coalition in red states, but at the moment the forecasted result would actually be pretty close to the 2012 electoral map. If this is a party realignment in progress, it's not yet a massive shake up of the kind which nearly inverted the electoral map between 1960 and 2000.