"IBM declined to re-implement the architecture in a more modern technology."
They must have had something else they were trying to sell them.
in the secure world, a further compounding problem is that the government demands exclusive access to many of the technologies you develop - making it difficult or impossible to leverage the work in the market. since that line isn't well defined, and the risks are so high, both you and the government err on the side of caution.
there is a constant struggle to keep your best engineers from being locked in a scif for the rest of their lives.
many people on the government side see themselves as the overall architects of the system with absolute fiat authority. this can lead to huge uncosted scope changes, difficult or impossible deliverables, the whole project lurching into the swamp because of some tertiary issues, etc.
despite the high price tag and hefty support contract, a lot of these deals are strategically poor and don't end up making you alot of money. in the limit your entire company functionally gets absorbed into the government.