Yeah. They say IPv6 is too complicated, so they will rather have all of the workarounds you mention. And they say IPv6 is lacking features from IPv4, including all of the workarounds you mention. And they want to be able to communicate with IPv4-only hosts without using dual-stack or any transition solution.
Once IPv6 have added all of those workarounds while becoming simpler and simultaneously achieve the ability for IPv4-only and IPv6-only to communicate without using a transitioning solution - they will eventually update.
At least that's what their explanation sounds like to me. Maybe they are just looking for excuses to procrastinate for another 30 years.
@kasperd @travis no, you really do need hardware support for everything when routing platforms are in the 2-50Tbps range these days.
Taking Arista as an example, only the latest generations supported IPv6 VTEPs for vxlan, and even their x86 software based DPDK platforms have only basic IPv6 forwarding support, but lack support for ACLs and VRFs. On the AV side, support for MLD for multicast is comically lacking across the board
I was referring to the IP layer and not a tunnel layer which isn't part of IP.
Yes some use cases of tunnels require those to be implemented in hardware. If you are using VXLAN I recommend you investigate whether your hardware supports Geneve.
Geneve is an IETF standards track protocol designed to replace VXLAN. Some sources say that Geneve is already widely supported in hardware, but I am not in a position to verify those claims.
Travis
Named Bird
Nick Buraglio
kasperd
Kenneth Finnegan
Jima 
Mara