I think so.

I wouldn't be surprised if one day we could compress to just the semantics of the film - characters, movement, lighting, props, textures, etc. and then "re-render" on the other side of the pipe.

If not for film and video, I think this will be used for "world models" / interactive video models that are playable "video games" / Holodecks.

This will probably come much later down the road as there are so many other problems to work on before "video models are also compression".