Efficiency in existing operations
Increased efficiency from current operations is the highest immediate payoff. Many industrial plants vent a lot of heat into the air or local waters. Lots of that heat, though no longer in a form usable by industrial processes, can be used for heating purposes. It's called co-generation by some, and is one of those uses for waste (heat in this case) which can payoff in a big way. In addition, the example of the sawmill which uses sawdust and scraps to heat the plant and generate some electrical power is another example. Altogether, this approach, if employed widely, will result in very large savings. Overall, not high capital investment, either.
Another increased efficiency measure would be a way to use superconduction in electrical power distribution. Not yet possible, but much research is being devoted to it. This would have the effect of adding many very large power plants to national power grids after installation.
Increased efficiency in insulation, reduced power requirements (eg, efficient appliances, more efficient light bulbs, etc) would again give the effect of more large power plants for large countries with populations using lower efficiency products.
Nuclear power (as in current fission plants or proposed fusion plants, if we can ever figure out how to make it work) is tempting. There are some problems however. We've not solved the radioactive waste problem, we've not solved the proliferation of weapons problem, and in an increasingly terroristic world filled with fanatics for whom the price anyone else pays to support their goals is a matter of indifference. Huge capital investment, much shorter lifetime than had been expected just after WWII, very large regulatory supervision which the current business climate regards as without any value to anyone, ... Accidents are increasingly likely (eg, Chernobyl, Three Mile Island, assorted accidents in Japan, ...) and can have...