Why are these publishers needed? What service do they provide these days? It seems their role is similar to publishers in other media (TV, movie, music, etc) that can and have been replaced due to the distribution ease of the internet. Aside from a distribution platform, what do these publishers provide?
However, in academia, the "prestige" of the journals you have published to is hugely important. Good journals, in theory, are more selective about what they publish, so the assumption is that if you have publications in a good one you must have done more groundbreaking research. To an academic, this is the kind of thing that affects whether you get grants, whether you get paid more, whether you can get a job elsewhere.
Some journals are pretty much talked about in hushed whispers, like "that PhD student got an ___ paper in their first year, they must be some kind of genius". In reality, research tends to be more about who you know, what kind of work you're doing, and whether you're in the right place at the right time.
I wonder why nobody created a single peer-reviewed platform for all scientists to use? Elsevier is just piggybacking on other people's work and funds.
Obligatory: https://xkcd.com/927/
What happens when the investors and shareholders of 'even bigger hegemon' start wanting their profits? Elsevier II.
It should be noted that this selectivity is not just on quality, but also e.g. on how "groundbreaking" the results are. This results in papers in top journals actually being retracted more often than in others, since they're more likely to be false.
1. They coordinate the submission of papers, and their dispatch for peer review.
2. They handle layout, typesetting, copyediting, etc.
3. For journals with print editions (of which there are many), they handle distribution.
All of those things could be replaced by the internet, but they take time, and they cost either time or money. And I much prefer reading journals with professional typesetting over those that rely on LaTeX.
There's also just the weight of history - many "good" journals aren't open access, and an academic's career trajectory is largely determined by publishing in those good journals. Going exclusively open access (in many fields) is either:
1. Expensive (paying for open access publication in a commercial journal)
2. An affectation by senior people who already have established careers
3. A huge risk by junior researchers making a principled stand at the cost of potential damage to themselves.
Article Processing Fees for their journals range from $1500 to $3000 per article: http://journals.plos.org/plosone/s/publication-fees
Does it really though? I can say it takes 10,000$ but does it really take that much money?
If you create an electronic document of a few MByte size, 1500$ should cover hosting it for the next 1000 years! The largest cost (besides writing the paper) is reviewing it and that cost is already externalized by publishers.
And indeed, there are selective open-access journals, whose content is free to access once it's been accepted for publication; and there are non-selective, crappy journals which are not open-access.