Astronomical telescope mirror manufacturing is a labor intensive, hands-on, non-automated process. And the culture of aerospace is highly risk averse: this comes from the very customers, like the good people at NASA Goddard.
Lunar telescope manufacturing would require some exciting scientific, engineer, and processing improvements that would also pay off for terrestrial manufacturing.
First, assuming they're not planning to house and employ a standard aerospace company, with 1000 engineers, technicians, and managers on the moon, this would be fully automated. Mirror making is anything but automated. The development of highly automated methods for processing and testing mirrors would be quite a move forward. It would also have direct benefits for conventional manufacturing.
Second, making a mirror on the moon would seem to require a tolerance of risk currently not accepted. Every time a mirror is moved, a crew of people must oversee the affair, sign the (physical) paperwork, and manually inspect the mirror afterwards. For lunar construction, this would have to become an assembly line that ran without that direct oversight, paperwork, or crews. Enabling more efficient methods would certainly benefit normal processes as well.
Moreover, the task of creating such a facility would keep many, many aerospace workers employed for years :)