The contents might already be on SciHub and other sites, but I can already see the bandwidth spike they're going to receive from doing this... from those sites wanting to complete their collection, among others.