Confusing speed tests with IPS & various VPN protocols

This topic has been covered before a few times but I was hoping to get some help wrapping my head around what I'm seeing.

I tested 5 VPN configurations both with and without IPS turned on.
As  previously documented IPS kills basic throughput (without a VPN).

I'm puzzled, though, with the results using various VPNs.
In most instances, IPS improves throughput on downloads and not with uploads Open VPN (TCP) excepted.
This seem counter intuitive.

Also Wireguard performance is horrible through the UTM - also counter-intuitive.  Wireguard is supposed to be much faster.
The fact that Wireguard is so much slower on domestic (US) connections is simply bizarre.
Is someone taking that extra time to snoop on the connection?  [Conspiracy theory !]

Hardware is my own - generally overkill, though I don't recall what specifically it is.
CPU Utilization during tests

Can anyone think of a rationale for these results?

Thanks IA.

Added a fun conspiracy theory
[edited by: SalishSwede at 8:45 AM (GMT -7) on 18 Jun 2024]