It is precisely because they are using far away rented servers that they have less commitment and interest in learning witness skills than if the witness node was living with them.

I disagree, I don't think that has anything to do with it. Granted if someone buys hardware they likely care more about Hive than renting, but correlation is not causation.

I keep my own backups of block_log and do my own snapshots which I can transfer to whichever Hive node needs them over my Gigabit Ethernet LAN. Even the 330Gb block_log only takes less than one hour to transfer.

I will agree this is far easier on your own LAN.

While I agree that the ability to replay is also important for hard-forks and emergency patches, replay speed is NOT about internet connection or where a node is located.

Its about single core CPU performance, RAM and I/O speed.

The fact is that the machine I specced above will beat 90% of servers in those factors.

Yes, but that wasn't my point. My point is you cannot rely on snapshots, especially when a witness is needed most (during hard forks and chain down events).

Regardless, I have no problems with running an API node on a home network if you have the resources to do it, a witness node I am not so agreeable about. I merely was saying the Internet is more of the critical factor especially as the chain gets larger and providers make their data caps smaller.