Well, apparently it counted as a separate application - so that required redundant (virtual servers) plus back-end databases. Add DR environment and live replication of VMs and databases, database backup retention planning, off-site storage costs. Ultra high performance SAN storage for all of that. Add pre-prod and test environments as well and it soon adds up.... apparently.

Turns out the internal infrastructure recharging model didn't scale down and as it was a "process" it couldn't be changed.