Yes, a replacement will cost money. But probably less than a year contract with Palantir.
Mistakes were made (I would argue intentionally because of local wallet minima with stakeholders), there is a sunk cost to using Palantir. The country can't even develop it further, because it's Palantir's. That's not an argument to keep using it.
The country needs to develop the ability, to develop critical national infrastructure (this includes software, obviously is not limited to it). I would also argue it already has the ability, we need to prevent it from withering away.
Palantir will do the same as local companies (building on top of open source) except it needs to make a large tech profit and with its monopolistic ability, It will capture the value for itself.
Procurement has a misaligned efficiency incentive. Procurement and governments want a single provider so it's less direct cost on them on managing projects. Sadly, they've actually unnaturally forced a monopoly, resulting in serious costs and inefficiencies in the long term. We need a way to encourage multiple providers for the same thing, and allow new companies to join in even when they're late. Just like B2B and B2C.
I wonder if anyone has some thoughts in this area