“Support” is an interesting word.
Nothing stopping you from buying multiple hubs. The issues are how you might separate out the logic and devices.
As long as you are ok with them operating independently (or figure out the interactions between) then its just fine.
I’d say, try to find our max… We certainly are. Busy systems might slow down for sure on complex execution or lots of devices sending commands, question is if its tolerable, or time for a second hub.
We are excited to get real world user feedback on how the hub performs. Having such an open system, it is always possible that one bad device can cause problems with the mesh, or bad code, etc. Thankfully that is limited to your system and can’t effect others performance in the “cloud”.
I suspect some of the community devs that will want to bang on a hub and test the limits might want to buy one for their WAF factor to run the house, and the other to test with.