It can be cost-driven, but it's just as likely to be complexity-gated (w/ cost frequently being implementation driven and a proxy for complexity). The cost comes from actually unwinding a gnarly legacy system (that may be less than well understood) and making the successor system work. Also, Gall's law has entered the chat:

A complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works and cannot be patched up to make it work. You have to start over with a working simple system.