In the overwhelming bulk of cases firms have a column store that they generate cubes from. OLAP is run against the cubes. Some put warehousing in between, though that changes little. Cubes are fundamentally a form of caching because historically it was prohibitive to do large-scale aggregations in real-time. With massive memory servers, and more importantly flash storage that improves aggregate performance at the enterprise scale by many magnitudes -- a million times faster analysis and aggregation is entirely possible -- that historic caching step becomes a hindrance and maintenance/timeliness issue. So it's discarded.
That's all 100% true. It has happened in many orgs.
Not all, of course. But as a general trend.
The proof of this? Go to any serious columnar database provider and search for the words 'OLAP cube'. You will find that they are careful to say 'OLAP workload', but not 'OLAP cube' — because in the strict definition of the term, an OLAP cube or data cube is an entirely different architecture.
Relevant sources are included at the bottom of the piece.
endorphone's comment has it right.