That's an divisive statement. I'd blame people who were unwilling to invest the time in properly modelling their data on the rise of NoSQL.

Transactional consistency and data normalisation - pffft.

SQL is still doing very well running things behind the scenes.

The problem is not data modelling. The problem is ensuring eventual consistency and synchronization of data in multiple datacenters around the world.

Talking about running things behind the scenes, mainframes with pre-SQL NoSQL DBs (ADABAS, IBM IMS) and even with no DBMS in a modern sense at all (running on TPF, working directly with Direct Access Storage Device records) are still doing very well.