According to the article, they're 70% of publishers' revenue (straightforward extrapolation from 30% margin). I can't think of any sense in which that's minuscule.
Research costs may be correct. Do you know where I can find data on average funding for published studies? Or how many government funded studies were published at what budget?
Sorry, I was wrong about publication cost vs the cost of published articles. They're comparable.
Let's say that I'm writing a paper. Maybe the underlying research cost $100K. So paying $3K for publication isn't much. And let's say that I need to review 100 papers for background, discussion, etc. At $30 each, that's also $3K. Even less, for subscription-based access.
So maybe it's a wash for grant-funded academics who publish in journals that charge $3K to publish and $30 per copy. But it certainly isn't a wash for academics with marginal funding, not-for-profits, etc, who publish at much lower cost.
And furthermore, they're minuscule compared to costs of journal access.
Edit: I'm wrong on the second point. They're actually comparable. See comment below.