Launch costs are a fairly small part of most space telescopes. Even for something like Hubble which required multiple incredibly expensive shuttle launches, <10%. Sure, if you treat the telescopes themselves as disposable you can bring R&D costs down a bit.

But imagine how hard it would be to build something like the Square Kilometer Array in space (keep in mind it'll already cost billions to build it on Earth). You'd need hundreds of launches to get each of the 50 ft dishes to space. Not to mention the difficulties of on orbit assembly.

And of course before you do that you have to figure out how to do interferometry in space. AFAIK it's only been done reliably at very long wavelengths, except for a few demonstrators that required non-scalable physical connections. I don't think it'd impossible, but it requires some huge advances in both science and engineering (submillimeter positioning, cryocoolers with orders of magnitude improvement at passing vibe checks, etc)