Failover primary cluster?

you can promote an alternative subcluster to be primary. however this is not automatic.
Its rare that all nodes of a subcluster will be lost at the same exact time.
Withing a (primary) subcluster loosing a single node or a few nodes is fine as long as you have quorum (i.e. n/2+1 nodes within a subcluster are UP).

Creating second primary subcluster is the only way how to move whole primary subcluster to different HW without downtime.
Once you failover all DB sessions to the second primary subcluster, you can drop the first one without any downtime.