The author really needs to define "data" and "stored" for this article to make sense. For example, I work for a large manufacturing company and the combination of our ERP/MRP DB and our manufacturing systems DBs totals about 20-25TB, with about 10-15% growth per year on average. That's small potatoes and
barely what I'd consider big data in the context of analytics. However, totally separate from this structured data, well, there's a whole of other structured data created by and stored in other app databases, but this is predominantly stuff that supports business automation, not core business systems. I'd guestimate this is in the 5-10TB range, but since most of that data is siloed and unrelated across apps, an analyst would never be looking at more than a few hundred gigs at a time for any given purpose. Then we get to unstructured data, which is what I thought this article was intending to discuss. With file servers at every one of our sites, with 20000+ employees using Google Apps/Drive, with who knows how many Dropbox accounts (not to mention internal FTP), and internal CMSes, we probably have about one PB of unstructured data, not including the stuff in the cloud. We generate email at about 2TB per month, and until recently backed it all up on tapes we ship to Iron Mountain. This is separate from the default business process of imaging a terminated employee's PC and keeping that permanently just in case (stays on a server for a while and is then backed to tape).
So, I'm not sure how to interpret this article, given what I know about how my own company works and where our data lives.