The science done by a bunch of smaller IR telescope is different from what you can do with JWST. (Basically you are reducing resolution but get better time coverage of transients.)
These missions don't even need to be physically small.
Then launch the more expensive one.
[0] https://en.wikipedia.org/wiki/Space_Interferometry_Mission#I...
If you're able to deploy orbital observatories with a StarLink sort of cadence (keep throwing up cheap satellites as quickly as possible), would you think we'd reach more mature observation tech faster? Maybe observing from Earth is dead long term if you drive innovation up and costs down with the observing systems and the supporting data transport networks.
TLDR How do we make decades of progress in years instead?