Honestly, as I know the industry, I wasn't really thinking of the review process as something Elsevier provides that is worth much.
Aside from sending a few emails, the review process is largely managed by academia, although I suppose academia may be happier thinking they are being managed by Elsevier than their peers (who tell Elsevier who to enlist for the work anyway...).
But the non-academic work is what you have to do yourself if you break the lock in.
It is not trivial and as it scales it gets increasingly complex.
There are two moats that protect Elsevier's business.
The first moat is that academics don't trust each other and perpetuate the "Publish or Perish" mindset.
In some disciplines (e.g., Liberal Arts, Math), there isn't enough money flowing through academic circles for this to be a real problem. But for the Life Sciences and some Physical Sciences in particular, there is a lot of government and industry money sloshing around, and academics seem perfectly willing to undercut each other and themselves as a group to get a bigger slice of it.
Elsevier takes advantage of that.
Maybe academics will get mad enough at them over it to unite long enough to break their ability to take advantage of them. Seems like it would be in their best interest to do so.
The second moat is that building a web platform to efficiently distribute hundreds of thousands of pages of content a year while still maintaining access to all the previously published pages is a seriously non-trivial task.
Now just providing that web platform can be done for much less money than is currently paid to publishers.
But, as I'm sure all the web developers and sys admins on hacker news would be happy to point out, hosting terabytes of data and generating consistently formatted HTML/CSS from thousands of individually created Word docs and graphics with nifty social media integrations and customized search/browse options is not a thing that can be done for free at the scale Elsevier is operating at.