Earlier quoted context omitted.
I had the same thought, but do your dozen folks use video calls on your instance? Does it work well enough? I'd definitely expect to be able to host a text-only instance for $5, but maybe they also need to provision for more demanding features? I only use Element / Matrix as an IRC replacement for now so I couldn't tell.
AFAIK, Matrix does not support Video chat yet? But the VoIP/audio chat is using WebRTC, so Matrix is only the signalling server and the audio is p2p between the participants. As such the strain on the hosting server should be minimal
Element (and some other clients, but I don't think it is part of the standard) also supports multi-user voice and video using Jitsi which works very well but you loose e2e encryption (although IIUC Jitsi has some experiments for this) and need to run it separately.