The exact numbers depend on the form of fusion in question, but fusion does have some several places where it has quite substantial x^n growth possibilities, where "n" is definitely greater than one and can be greater than two at times, sometimes even substantially so. This means that there is some real, concrete hope for improvement in a way that, say, solar could never improve more than 4-5x where it is now because the absolutely best it could ever hope for is 100% efficiency. At the core, this is because as you get the plasma hotter and more confined, the rate of fusion goes up very quickly, much much beyond linear increases.