Part of the value of Nostr is being able to choose who stores your data, or even store it yourself, creating censorship resistance through optionality and redundancy.

This doesn’t scale very well to videos, at least not as currently constructed and with current technology. I’m no computer scientist or coder, but I know that my NOSTR relay wouldn’t be able to handle much traffic to read from it, nor would I want many people to write to it. At best, maybe you’d have two or three major relay companies competing and all storing your content, but media with lots of data really benefits from economies of scale.