I'd like to see some evidence that publishing a journal requires each article to be costed at 2995 dollars (a suspicious looking figure to me).
I'm an academic. I get asked to peer review articles for free. We do it as part of our workload. I have colleagues who edit journals. They do this for free. I author articles: I do this within the costs of my project, the journal gets my article for free. Authors work for free, reviewers work for free, editors work for free. It's just the production and publicity team that get paid (the publishing house). We don't even expect them to roll the presses and produce paper versions these days, we are happy with web links to PDFs.
So we need to think hard about what the costs are in putting an online journal live onto the internet.
Why do academics continue to publish in closed journals? because generally they are still the high impact ones (with a very few exceptions). So I, and many other contract researchers like me, tend to publish in closed journals because these look better on the cv. Philosophical high ground is all well and good but when you've got a child to feed and a house to pay for you have to be pragmatic and keep in a job.
I can imagine this might change over the next 20 years or so as more and more folk start open access journals and they are gradually given greater impact ratings.
Personally I think we're going to see a few universities taking the lead with open access journals and this might break into the monopoly held by a small number of publishers right now. If you're doing it not-for-profit you can do it cheaper than a commercial publishing house that has to show profit to its shareholders.