What carries over
The parts of Tend that decide what to change are platform agnostic: demand forecasting from workload history, the operations-research models that turn a forecast into a capacity decision, and the business constraints and approval rules you set. Those do not assume BigQuery. For how they work today, see the technology overview.
What does not
Databricks prices compute in Databricks Units, billed per second across clusters, SQL warehouses, and serverless, with the underlying cloud instance charged separately. BigQuery bills reserved slot capacity. The forecasting layer transfers; cluster sizing, auto-termination policy, and rollback have to be built against Databricks' own semantics. That work is not started.
What support would cover
SQL warehouse sizing against forecast demand, auto-termination and auto-scaling policy, cluster right-sizing including instance-family and spot selection, and DBU attribution per job and per team. Nothing here is committed. Treat it as direction, not as a plan with a date.
If you run Databricks
Tell the team. Design-partner demand decides which platform comes after BigQuery, and Microsoft Fabric and Databricks are both unstarted. For current status, see the roadmap.