SQL is the language of relational databases, and databases, relational or otherwise, are all about data. As you have seen, data come in every imaginable shape and size, and need to be structured to some degree in order to be queried with SQL. After all, it is Structured Query Language. With ever-faster processing speed and ever-cheaper storage, the need for preprocessing the data (conforming it to some data model in order to be queried) diminishes, and in some cases goes away. So does the need for SQL.
The latest buzz in the community is the NoSQL database. As with every latest and greatest idea, you might have heard about it before. It is a dumping ground for the data. The reinvention of the idea given up for dead in the late 1970s was made possible by the unique combination of the cheap storage, fast CPU(s), and ubiquitous high-speed infrastructure to glue it all together.
There are new usages for SQL, as data are being increasingly distributed across the globe, and data consumers adapt to ever less preprocessed data with new computing paradigms.