Elsevier and other large publishers publish many journals on behalf of societies, so the distinction between the interests of professional organisations and publishers doesn't really work.
I think there is a clear distinction: Elsivier cannot go to an open access model, because it would cease to exist from lack of revenue, while the ACM and IEEE can continue to exist with an open access model because of membership dues.
The ACM already makes a lot more on conferences than dues, I think.