You build your first 1000 warheads and develop a delivery method. Soon the other party develops countermeasures. So you still have 1000 warheads, but the delivery is no longer guaranteed.
You develop a better delivery method, and it's easier to build another 1000 warheads than repurpose the first 1000. You end up with 2000.
It won't be long you will have to build more. As the warheads are not consumed, you end up having more and more. But only 1000 of the latest generation, that really counts.
By far the most expensive component of a nuclear warhead is the enriched uranium, which is much easier to recycle into a new warhead than make from scratch.
Between 1940 and 1996, the U.S. government spent at least $9.49 trillion in present-day terms[5] on nuclear weapons, including platforms development (aircraft, rockets and facilities), command and control, maintenance, waste management and administrative costs. Saving ~8 trillion was very much a viable option. https://en.wikipedia.org/wiki/Nuclear_weapons_of_the_United_...
By comparison China built about 600 warheads since 1964 which was a significant deterrent for the US. The US on the other hand built ~70,000.
PS: The Minuteman-III was in service from 1970 to now, it’s scheduled to be retire in 2030. While bombers where replaced ICBM’s where hardly a quickly outdated weapon.
As in the bombers example, the previous generation is hardly quickly and completely outdated. They're still nukes. And that's exactly the reason to keep them in service.
Then, when a better tech becomes an option, it doesn't replace the previous weapon, it complements it.
So the options are: 1) retire old weapons, recycle them to build weapons which are expected to be better, keep the military potential about the same, reduce costs but still spend a lot and 2) keep what's in service and build a better weapon. Option 1 is hard to justify politicaly. Scraping something perfectly functional that costs billions sounds like a career ending decision. It also affects too many people. Option 2 is almost a guaranteed way to improve military potential, and a way to have a stellar career. Assuming that funding is available.
So yes, the decisions are not always the most economically efficient. That's a wrong objective function. Especially if we're talking about defense.
If the new weapon systems are about as good as the old ones then what’s the point of building them? Presumably there is a significant threshold for spending 100’s of billions of dollars on a significantly improved nuclear stockpile.
Justifying it to the American taxpayer is meaningless when it’s hidden from them. But, justifying it to Congress was just a question of pork spending which is really what the excessive nuclear stockpile was. It’s no accident that all 3 major branches of the military had Nukes, though currently that’s down to Air Force and Navy.
You develop a better delivery method, and it's easier to build another 1000 warheads than repurpose the first 1000. You end up with 2000.
It won't be long you will have to build more. As the warheads are not consumed, you end up having more and more. But only 1000 of the latest generation, that really counts.