> The interesting point here is that Sci-Hub isn't really a threat to the publishers.

I'm not sure if you can say that. Consortia of universities in both France and Germany have recently decided not to renew their contracts with one of the big publishers, and although it's hard to pin down to what extent motivations contribute, one could very well make the case that they have been able to do so because they have less pressure from academics to retain subscriptions - thanks to Sci-Hub.