Journals do minimal typesetting, and practically no editorial work. Paper reviews are done by academics, free. Editors are mostly academics, not depending a paycheck from publishers. There is cost, like marketing cost, but none that justifies $35 per article to the general public.
That's either a bad lie, or you don't understand how this works. As a small example: Science, Nature, and various other journals produce high quality videos (like this: https://www.youtube.com/channel/UCv0aU2eKry3kdSTnFa8QAWA ). And they have many in-house designers and editors who do important work like identifying from a sea of work what is important and what is not, etc.
It's neither, I think it's just that different fields operate differently. In CS most papers are published in conferences. Authors do their own typesetting, editing, figure making, etc. Peer review is organized on a volunteer basis by the program chair, who is (as far as I know) not usually compensated for their efforts.
This means we don't get many fancy videos unless we or our university's PR people make them, but it still seems to work fine for getting science done.
