For military operations it might actually be a possibility to just not relaunch though. Explode the engines/anything sensitive and just leave it there. They're surprisingly cheap to build (compared to a lot of the expenses incurred during combat operations), at a (public and old) target of $5 million/ship.
Though the Falcon 9 is $55 million, now that's comparable to a modern fighter jet.
Bear in mind that it is only the Starship (upper stage), not the Super Heavy booster, that would be landing at the destination and hence potentially subject to abandonment. Also note that Starship, unlike Falcon 9 (or at least to a much greater extent), is designed to be produced efficiently in quantity.
I think you're probably looking at a few more million in launch costs, but those aren't recovered by recovering the ship anyways.
https://en.wikipedia.org/wiki/Propellant_depot
It would have to deal with the extra weight of all the fuel on landing, but it would enable operations in completely remote areas.
More realistically, if you have access to methane, you can just refuel the rocket. Theoretically if you had access to water for hydrogen and atmospheric carbon dioxide, you could manufacture artificial methane - needing just electricity.
Musk has mentioned that before: https://www.hindustantimes.com/world-news/elon-musk-plans-to...
Accordingly, the Air Force science and technology investments will include "novel loadmaster designs to quickly load/unload a rocket, rapid launch capabilities from unusual sites, characterization of potential landing surfaces and approaches to rapidly improve those surfaces, adversary detectability, new novel trajectories, and an S&T investigation of the potential ability to air drop a payload after reentry," the document states.
Thanks for that.