With stuff like Facebook opengraph (e.g. og:price) and other meta tags meant to help search engines and social networks get this sort of data to display inline, do you think it's inevitable that complex scraping will no longer be needed in a practical sense since everyone will be inadvertently optimizing their markup in a way that you could write a really simple parser to grab the data?

Most companies will use resources not in their datacenter and not identifiable. Executives know it's sketchy, but they do it anyway.