This is the fundamental problem with fusion. It probably can’t ever be economically viable for power generation.
Suppose we get a working research reactor that produces net positive power in the reaction itself. What’s going to be in this baby? It’s going to have, at a minimum, thousands of miles of superconducting cables, building sized banks of lasers, and so on. One horrible problem here is that superconducting magnets need to be kept cool (probably using liquid helium), and yet somehow you have those magnets next to a neutron source that is heating everything back up, creating steam.
Similarly, lasers are not very efficient, so even if you have positive gain in the fusion reaction, you’re driving a laser that only converts a few percent of input energy to light with energy from a heat engine that is also limited in efficiency.
So you’d need huge fusion gains for it to even work at all. Ok, so let’s posit they find a way to get these gains.
How is it going to ever be economical to build a power station that requires immense quantities of high grade components, immense quantities of rare elements like rare earths, and the whole thing has to be aligned and tested by crews of PhD physicists.
Contrast that to something a little simpler. We could just make a factory that creates solar panels completely autonomously, reducing labor costs to near zero. We’d build a second factory that does the same thing for lithium iron batteries. Solar panels and lithium iron batteries are far simpler devices than anything that goes in a fusion reactor, and they don’t require more than minimal training or education to use.
In the far future, you’d just make solar cells directly and put them in space so you don’t need batteries at all. (you’d make the solar cells in a robot factory on the moon and launch them with a mass driver for minimal cost). This effectively is just building the energy collection part of a fusion reactor, and letting the sun do all the rest of the steps. It’s always going to be cheaper than building a fusion power reactor, no matter how advanced future technology becomes.
With all that said, there is a form of fusion called “aneutronic fusion”. It requires certain extremely difficult fusion reactions. However, you get beta particles (electrons) directly as the output. You slow the moving electrons down between 2 charged meshes and get power out directly.
Also, there’s at least 1 use case for fusion : rocket engines. You don’t try to generate electricity, you just let the high velocity particles from the fusion reaction fly away from your spacecraft. This gives you better thrust/weight than generating energy and then powering electric engines (such as ion engines or VASIMIR). It’s less fuel efficient, though. So the only use is manned missions inside our solar system (the semi-bad fuel efficiency means you’d never get enough velocity for interstellar missions)
See here : http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20050160960.pdf