One group is building a 100,000+ part device, with much HIGHER accuracy than astro, using a miserable wavelength (UV) for $200M a machine.

Not only that, they are doing UV light emission generation (dual laser) at 50,000 per second just to generate the light.

Actuation speeds are actually ridiculous in these machines as they image. Precision is ridiculous (and far higher than astro), lifetimes are pretty good so far. Reflectivity? EUV is one of the hardest wavelengths to manage reflectivity at. Does it even do anything on a standard polished surface? I would expect it to be absorbed pretty much 100%.

In terms of size, if you look at the individual elements the TMT is not one mirror, it is many smaller mirrors.

I think your comment sort of explains why astro is trapped. There is not a willingness to consider options.