>Doesn't fusion demand constant refueling (like most power plants)?
I don't think the half-life of deuterium/tritium is particularly long, so even if you could store it, it might not last decades before use. Though on the other hand, it's conceivable that you could design the reactor so that it produces deuterium/tritium on site from hydrogen cracked from seawater.
But realistically the reason we want fusion over fission in the first place is the proliferation concern. All of the other arguments against fission are NIMBY nonsense that could be applied equally to a whole swathe of chemical processing facilities that nobody seems to care anything about because they're not OMG nukular. And there is hardly a major proliferation risk from a reactor surrounded by military personnel on a vessel designed to securely carry actual nuclear bombs, so it makes the specifically military application kind of questionable.
We really need fusion for two reasons. 1) There is a better chance that we can actually build a large number of reactors without roving gangs of idiots destroying their cost effectiveness with gratuitous and intentional red tape, and 2) we can build them in countries we don't trust without making it easier for them to blow up the world, which will help to fight global warming vs. those countries continuing to burn oil and coal.