Safety isn't that much of a concern, it's a simplification but all of the places we're likely to put an orbiting station for the forseeable future are in the same ballbark of risk in terms of ease of access and return, obstacles, etc.

You can't separate physics and dollar cost since dollar cost is directly related to payload, quite simply delta v costs money.

Launches play a big part... Like, the ISS is low enough that it suffers a lot from atmospheric drag and its orbit needs frequent boosts (and that'll get worse as the atmosphere warms). But obviously higher up = harder to get to. So that's just a plain old compromise, but it's ultimately one that can be handled within a really wide range- raising it occasionally is just a question of fuel, so that's meant that the low-ish orbit has worked well and that's probably still true.

The other being the orbital path of course, since you have to have the launches intersect with the orbit. And that's simple phyics really but not simple human-stuff. Where will we be launching from in 2040, and how will we be launching? Will we have equatorial stations, or more mobile floating launchers? Will be still be using chemical rockets for everything with no other options in sight, or will we be kettling stuff up, or have a big railgun up an equatorial mountain, or be launching payloads from the moon, or getting close to any of those? It all gets insanely complicated, right down to "which US politicians want to keep launching from their state" or "which countries will be friendly and stable enough to invest this stuff in"

An ISS replacement in the short to medium term, I bet 20 scottish pence would end up at a similar orbital height, but with a different track to suit current US launch sites and less or no thought to Russian cosmodromes. In the longer term I'd expect payload delivery to get easier and therefore a higher altitude to become more desirable, especially with a warming atmosphere, but for now it's almost certainly still better to be lower and to get mass there easier