The James Webb Space Telescope, nicknamed the telescope that ate astronomy, has cost an enormous amount of money due to the need to get everything aligned absolutely perfect before launching because there will be no chance for repair.

In comparison, with adaptive optics ground based telescopes have a lot more leeway to be fixed and corrected.

(That said, if astronomical interferometry can achieve the same result as a 100Metre telescope, then let's do it that way! How is less important than making sure we have something in our pipeline within the next decade that can be capable of directly measuring the spectra of exoplanets.)

Extremely large telescopes use segmented mirrors to avoid this problem.

here is a telescope idea that is theoretically super scalable

http://www.pellegrino.caltech.edu/aarest1/