Databricks supports MDM in the way that any off the shelf database also can--you just have to write all the code to handle the data standardization, survivorship and entity resolution rules. You can absolutely do MDM in Databricks, the medallion architecture corresponds nicely to how traditional MDM systems categorize data, and the flexibility of Delta Live Tables pipelines makes it easy to write all the code you need. Streaming tables can be used for near real time MDM, while the scalability of the Spark/Photon compute means you can also handle gigantic batches of data.
You can import Python libraries to assist with your coding and use those libraries in your Delta Live Tables pipelines. But you are "build first", and I've worked at places with that mentality and it's fine, you just have to maintain a lot of code (but, you can get very customized MDM at potentially lower cost).
An advantage that Databricks has is the Databricks Marketplace. You can subscribe to services from Dun and Bradstreet or Experian which will do the standardization and entity resolution for you, then return the golden records. Another great feature Databricks has is the data expectations, which are used to measure data quality before any MDM work is done, and can quarantine bad data to keep it from ruining your golden records. Both the Marketplace services and expectations are used in your DLT pipelines.
I've built entirely custom MDM systems on SQL Server. They work, and they easily fit in to your enterprise's service inventory, but they really require a dedicated team to maintain the system. MDM systems like Reltio are also lesser impact on your enterprise, but are SaaS so all you do is maintain the rules. I describe systems like Informatica as "an enterprise lifestyle"--your business conforms to how they run.
In the end, all MDM systems cost money and take effort. The decision largely comes down to whether you want to pay your own developers and system engineers, or someone else's. All MDM systems have the same need for data quality, data governance and business rules governance, so those are the same across the board. It is possible for different companies to take different paths, and all be right.
