Here's my thought: text and images were easy enough to produce that they made sense for mass peer to peer social platforms. Video is hard enough to produce, and host, that a platform focused on asymmetric publisher-consumer relationship was feasible. Podcasts fall in the middle; they're not so easy to make that just anyone can do it, but they're also not out of the reach of amateurs and hobbyists to create and host. I think companies like SoundCloud just couldn't make a go of taking it over.