Amazon lost money every quarter as well for years. There is a huge market in government and defense IT globally that Palantir could potentially tap into. Historically, it's been IBM's gig, but that could change very quickly.
Features, requirements, scope, datasets. It's not like services are cheap and easy to make.
The other part of .Gov contracting is integrating so tightly that you become a defacto standard that is not easily replaced. Then lobbying to ensure you keep that position.
Government and defence IT is a service business - where do you invest the money?
Note: I'm not saying they don't invest the money, just curious where it goes.