When you use all of that data, on a conservative end, you end up with 600 million pages that have been copied and not paid for. These are copies that are not licensed, nor have transactional licences been obtained for them that are not available under open access licences.
On May 22nd, 2018. See this statement in context.