From a gameplay perspective, solar is the best, and I hold out for it as much as possible. From a realism perspective, it really shouldn't cost more. It's possible to just use cheap reflective mirrors to focus the sunlight for heat to drive turbines. (Salt, water, whatever) Solar could easily be argued down in cost, or there could be a steam based solar plant that consumes water and takes a good chunk of metal to build. Rather than nerfing solar, I think something like a volcano powered geothermal plant should receive a big buff. I also wouldn't mind seeing the power output of solar be reduced if the cost stayed the same. Making solar take advanced parts just rubs me the wrong way. I like the idea of having more modifiers affecting things based on the planet you're on. Atmosphere density, rain, distance + star type(?), elevation(?) all affecting solar energy output. You may be better off with some other power plant if your planet is a couple AU from the parent star.
I'd also agree with the general thread premise that the tech tree could use some adjusting. I saw a desire for more control over how research is focused, related to the way you can easily kill a game by failing to get a power plant while opening up too many research paths. Some way to direct focus onto a section of the tree could fix that issue; it would take away from some of the random nature of the current research system, but that might be for the better. There's still a ton of randomness the player has to account for.