My expectation is that the next dominant platform will offer something unique which creates enough value that I'm willing to overlook the unfinished parts that deduct value.

When Neo4j and other graph databases went big in the past decade, they value they created: "finally! I don't have to faff around with RDBMS tables and Cobb purists to store and query my non-relational object-graph!" - despite the fact that Neo4j then (and still?) didn't support running serving databases concurrently or schema enforcement or even transactional atomicity (I think they fixed that recently?)

So in this context, what business-value does PolarDB add or create that makes it worthwhile to deal with its expected short-life?