The earliest modules are inseparable, maybe literally. Zarya is attached to Unity on the US side, which has the truss. Zarya has been so intertwined and might be cold welded to Unity. Zvezda, on the other side of Zarya, still handles a lot of station control. You could replace Zarya with basically a new self-sufficient space station, and, at that point, why take on the baggage of the rest of it.
Oh yeah, I forgot that was a thing. I agree, if literally the entire station is cold-welded together, that might make it too difficult. But if it's "just" Zarya and Unity that have the cold-welding problem, why not at least disconnect and reuse some of the modules on the other side of Unity, or at least some solar panels and stuff?
If you think about it, even just attempting to salvage part of the station could be an interesting and useful experiment in and of itself, regardless of whether they expect for it to be successful.
at that point, why take on the baggage of the rest of it.
Because being attached to the legacy ISS, at least in the short term, would preserve a lot more space for activities than launching a single module standalone and letting the rest burn up? I mean, sure, if we already had a half-built replacement station in orbit right now, I'd say let the ISS go. But we don't. Right now it's very questionable whether we'd have anything flying by the time it's scheduled for de-orbit, and I find both that and the notion of replacing the relatively-gigantic ISS with something Skylab-sized (at least for a few years) to be an unacceptable downgrade.
Bio contamination is my guess. Even the ISS is getting quite a few gross spots that are becoming mini ecosystems. Of course the experiments are on purpose, but there is a lot of astronaut filth and other basic growth. Much like how an old house cannot be kept up to date by making additions and rennovations for ever, the ISS cannot go on for ever. At least not without something akin to a serious rennovation, but at that point...
Much like how an old house cannot be kept up to date by making additions and rennovations for ever
FYI, houses do not work that way. They can, in fact, last forever as long as you keep up with the maintenance. The main reason they don't is that people either abandon them or want to build something else in their place (i.e. they become functionally obsolete).
Now, if you want to argue that the ISS is functionally obsolete, I could see your point... but then again, it's hard to argue that a thing is obsolete when nothing exists to replace it. Even if the engineers aren't thrilled about trying to interface new modules with the decades-old stuff, I don't think that's a good enough excuse to throw away the entire existing thing and start over from scratch.
Other companies are working on inflatables and the Bigelow owner was a loon, so don't mourn that loss.
NASA's Commercial LEO Destinations program is working on private space stations. Orbital Reef will include an inflatable made by Sierra Space. There's also Starlab, which is a single launch big can, and Axiom, which is like ISS 2.0 and will bud off the ISS.
And there are other companies working in stations. Gravitics and Vast are making cans. Max Space has concepts for inflatables.
At the rate things are going, just launch a kitted out Starship and have that serve as a temporary space station. That should tide us over for however long it takes to design, build, and launch a newer station.