Good stuff, thank you. I was going with the idea that some form of higher technology would allow Osmium to be refined in greater quantities at a lower cost, but using it for specialty rounds makes more sense.
Does the fic. universe have any functional VonNeuman type tech? Or at least highly automated mining? If so, then yes. You can just turn the machines loose and they'll do whatever you want for "free", save the overall entropy of the Universe at large and wherever it is they get their energy from.
Otherwise the the statistical distribution of rare metals and heavy metals works against you in an economic sense no matter what.
Another though is that high density metals are like ball ammo, they might just sail on through a target, carrying a lot of wasted energy with them. Less exotic materials that aren't as hard/dense will vaporize more completely, and create more secondary damage.
Another consideration is the ship's drive efficiency and technology in the fic. universe. If it's GOOD, it's possibly just easier to get your KEW's up to speed and let them go, then make a course change, than it is to launch them with some sort of "gun".
This falls back on the Niven's "Kzinti Lesson" namely: "Any drive technology that's sufficiently efficient, powerful, and advanced, is itself a weapon that is equally efficient powerful and advanced.
You might be looking at a situation where offensive KEW's are gotten up to speed by the ship and released since you are choosing the time and place of firing, but defensive KEW's are shot from "guns" because you may need to lay down CIWS fire anywhere, independent of your maneuvers. (Presumably if you could maneuver/thrust to fire back and intercept incoming KEW's, then you could just get out of the way/evade too.)
You could also touch on all sorts of strategy like "bright" KEW's vs. "dark" KEW's, where a spread of high albedo, radar reflective, or warm in the IR spectrum KEW's gets an enemy ship to move to evade, only to run into a second formation of "dark" projectiles that are chilled to reduce their IR signature, are not radar reflective, or have a low visible abedo.
Then I imagine there's all sorts of complicated mathematics based on how many KEW's you've got. The targeting precision, the distance you're firing at, the closing velocity, and the enemy's Delta-V to evade at velocity X when they've got Y% probability of detecting the incoming KEW's, which mandates firing a spread of size Z to cover the area they can't escape from in time.