Superconducting loops and hydro seem to be the most realistic options at the moment. Hydro is limited by geography, superconducting loops by technology, though there has been some real progress on the latter the last couple of years.

A combination of wind + superconducting load levelers already works quite well in that grid load fluctuations are dealt with efficiently (and in a very compact package) allowing windfarms to feed old and fickle electrical grids.

The knowledge gained there can be applied to longer load shifts but it still is a real challenge.