It's not just readers who pay. Scientists also have to pay some journals to get their work published (assuming the submission went through) which is totally backwards. Scientists do not get anything monetary from what the readers pay. Now look at other publishing industries. It's the publishers that pay the creators (even if the rates may seem unfair at times, it is still far far better than negative gain for scientific research content creators). Been there multiple times back when I was a PhD student. So I have absolutely nothing against those who use the free sites.
I don't know for sure, but I think that some of their customers are universities, and possibly governments and large enterprises, who subscribe to large swaths of their offerings as bundles.
first, sometimes there are publication fees on the author end.
on the buyer end, university libraries buy huge bundles of journal subscriptions for millions per year to give access to everyone at their institution. this is essentially enterprise sales and the publishers are really good at it using all the typical tactics for winning lucrative contracts.
and the universities don't want to just tell their researchers to use sci-hub, although I think e.g. UC Berkeley essentially did this with a nudge and a wink a few years ago during negotiations.
So I'm not sure why people don't vote with their feet (well, mouse clicks) regarding Elsevier.
Of course, copying and stealing is very bad.