Ours is speculative too. In our case, we a building knowledge graphs for individuals where the schema depends on the person and their needs. We also hope that t

Ours is speculative too. In our case, we a building knowledge graphs for individuals where the schema depends on the person and their needs. We also hope that the relationships between data is where the value lies, rather than in the nodes which consist of ingested data. Arguably, the structure can easily be captured with a NoSQL DB or even a relational database. The precarious life of Graph databases is likely due to this - once a schema is established, why would you choose a proprietary datastore?