It's particularly staggering because it's unlikely they've correctly verified the status of all 200k, and are just dumping out a database query. (Whenever one checks these things, it typically turns out that many are actually CC or otherwise Open Access, PD because federal government work, preprints permitted to be shared, etc.)

One wonders if they sent Researchgate a DMCA and thus have perjured themselves countless thousands of times, and why Researchgate is rolling over?