The reason is: the second one doesn't cost much less than the first. They are hand-built. It's structurally very complex, so building takes a long time. Piec

The reason is: the second one doesn't cost much less than the first. They are hand-built. It's structurally very complex, so building takes a long time. Pieces are sent cross-country multiple times to assemble more and more components ("integrate") and tested in stages. The assembly and vacuum test of one module uses the whole chamber, so you can't do two at once. I'm using modest language ("very complex") - but the actual complexity is more like "uniquely mind-boggling". (I happen to know the MIRI cryocooler system engineer - just another Harvard PhD whose thesis required liquid helium.) For more on the MIRI cryocooler, which is the critical path item in JWST, and has been for some time, see: https://en.wikipedia.org/wiki/MIRI_(Mid-Infrared_Instrument) https://en.wikipedia.org/wiki/MIRI_(Mid-Infrared_Instrument). * The cost of these flagship missions, and the all-or-nothing problem that you point out, has caused conflict in the Astrophysics community. It might be better to have a pipeline of lower-cost but more frequent missions. A recent National Academy of Science report on exactly this question (https://www.nap.edu/catalog/24857/powering-science-nasas-large-strategic-science-missions https://www.nap.edu/catalog/24857/powering-science-nasas-lar...) concluded that the large missions did offer a unique benefit in pushing boundaries of what's possible. That report pinned the real problem on getting correct cost estimates, not the notion of large missions in itself.