Perhaps the best way to figure out if a power generation method is worth the trouble and expense is to see if people are already doing it. While there are various good bets being missed, none of them are that big and easy.
The most attractive methods that amount to the same thing as generating power are power saving methods. If you made a hobby out of tinkering with saving power, like with insulation and cutback thermostats and more efficient cars, you’d save quite a lot. Especially if you have access to a factory or some other large user of power.
Now, about power generation methods, as existing power supplies get more expensive of course the break even point for new methods gets closer and closer without the methods themselves getting any improvements at all. When you keep improving them, so much the sooner and better.
An interesting test case is photovoltaic solar cells. They can’t create a net increase in power availability for at least about 30 years. The reason is that they also take a lot of power to manufacture. Assuming various exponential growth rates in the solar cell industry, the industry itself consumes more power than its already-delivered products are generating, for at least 30 years, and that’s at an optimum growth rate. A faster growth rate has the industry consuming much more power, and a slower rate has their delivered products contributing much less. In the distant future, the industry’s products will eventually generate more power than the industry itself consumes (unless something unforseen happens to stop the industry or to make it unexpectedly more efficient), and it may be worth going down that road, but numerically it isn’t a way to increase available energy in the next several decades.
Cite? I read this in, I think, the Economist a few months ago. Might have been Scientific American though. But I think the more important point is that, since power generation equipment takes power to make, you have to analyze both sides of the relationship to see whether a new technology that works will contribute anything at all.