They wouldn't have server costs, but they'd have a lot more costs in terms of support.
People would see a feature advertised, and they'd expect to be able to make it work with, more or less, the same ease as most pieces of AV gear, which don't involve tinkering with NAT, firewalls, port forwarding and so on.
When it didn't work, they'd call Humax and complain. And Humax, lacking the massive database explaining how to set up port forwarding on every cheap router that's out there, would almost inevitably have to pass the buck to ISPs or networking equipment makers.
So, the net results is that while a few technical people would be able to make it work, the vast majority would have problems, and Humax would be able to do nothing much beyond telling them to phone someone else for assistance.
Net result? More phone calls, higher support load, and most users left with what they'd perceive as an unsatisfactory response. That's a very real cost, both in reputation and financial terms.