Ah, this old topic.
Look, at the end of the day, unless someone can actually get a copy of Canon's engineering schedule, testing costs, design documents, etc, guessing at how much video costs to add is just that, pure guesswork.
However, even if sales volumes go up, video is NOT free. It ads complexity that needs more testing (not as cheap or quick as people might think), it requires more hardware and software to operate, and chances are there were design meetings where they had to weigh various decisions based off how they would impact still vs video performance.
Outside DSLRs, plenty of dedicated still and video cameras still exist. They tend to fill niches now, places where volume is already low and thus the reduced engineering costs and less weighted design decisions really matter. If Canon put the same resources into developing a dedicated still camera instead of a multimedia camera, chances are it would be functionally superior in some non trivial ways, just like Canon's dedicated video cameras are generally considered better then their multimedia ones.
Part of what we have here, as a marking issue, is that there is a larger market for video then stills in general, so dedicated video cameras make sense because plenty of videographers would probably complain at the design compromises but not enough still photographers do.. and still photographers who are really particular move to one of the niche cameras.
Realistically, multimedia cameras, esp at the consumer end, are probably here to stay. But just like monochrome ones, a few specialized models will probably stick around. Or who knows, maybe some open source kit plus 3D printers will just let people design whatever they hell they want.. I am already seeing that start to creep in to the video market....