It ends up being a question of time and cost. Even if you make no changes to the design, the construction and verification of a new copy will still take years (

It ends up being a question of time and cost. Even if you make no changes to the design, the construction and verification of a new copy will still take years (verification of space payloads this expensive takes FOREVER!) and likely still cost hundreds of millions of dollars. I wasn't involved in the finance side of things but I'd guess that at least 40% of the effort would need to be duplicated for a project that size. But most likely, technology will have progressed in the meantime and you'll want more accurate sensors or more powerful on payload data processing or more powerful/efficient radios so you can get more of the data to the ground. Spy satellites only have a useful life of around 20 years due to obsolescence of sensors. So you're going to end up wanting to upgrade the new one so you don't just lose 1/4 of it's useful life. So you'll probably end up designing a lot of parts again.