Because I do a lot of network troubleshooting and having no IPv6 would be a significant style cramper. I expect to be able to do end-to-end IPv6.
That’s not entirely the point though. The point of the VPN is that to the outside world, to the internet, your traffic originates with the VPN provider and your IP address as seen by the outside world is that assigned to and by the VPN provider (not the IP address assigned to and by your ISP). Conceptually that works just as well with IPv6 as it does with IPv4.
A VPN provider that supports IPv6 would have to be careful not just to assign you the same IPv6 address every time you connect to the VPN service. A VPN provider that supports IPv6 might also want to take care to enforce IPv6 address reuse across different customers over time.
IPv6 is more fun because not only can every device have its own IP address, every device can have a zillion IP addresses, so most (privacy-concerned) IPv6 users get a new random IPv6 address every X hours (when not using a VPN so that they themselves are determining the IP address). So you can have your own IP address but it is not the same IP address all the time even within a fairly limited amount of time. (For clarity, the top 64 bits of the IPv6 address typically won’t change but the bottom 64 bits will randomize every X hours.)
In the original specification, the top 64 bits might have been assigned by your ISP and the bottom 64 bits might have been derived trivially from your MAC address, and this address is globally unique, globally visible and unchanging, so there was a significant deterioration in privacy as compared with IPv4! You can of course still do that but I think for quite some years Linux out of the box defaults to the more private address-randomizing option.