It's a pity that we don't know for sure (officially) which approach Liberty Global used in other jurisdictions. It's reasonable to assume that whatever they did elsewhere they'll do in the UK too.
Lacking that information, I think I'll cast my guess with @ravenstar68, because the *FAT* long tail of IPv4 that would result if IPv6 were given automatically only to new customers makes no sense to me. The fat long tail would have the following disadvantages:
• By needing to keep all of their existing IPv4 resources around, it wouldn't let Virgin move rapidly to IPv6-only internal infrastructure. That would be very bad for them since VM would lose the large savings in cost, manpower, equipment and reduced headache / security risk which provide most of the corporate incentive for ISPs to move fast to IPv6-only internally.
• It wouldn't give Virgin any information about how many IPv4 blocks can be freed up for selling off while the market price of IPv4 addresses is high. The loss of that easy one-off windfall does not seem very strategic, and for a profit-motivated company, unlikely.
• It would waste the high degree of compatibility and operational transparency of the two protocols, which for most common applications are completely interchangeable and work fine side by side. Most users wouldn't notice the change at all, giving Virgin a quick win. Losing that low-hanging fruit would delay their IPv6 transition plans and burden them with high costs for a long period, a very high price to pay.
Transitioning to IPv6 with a long IPv6 tail is a very bad package deal, so I expect Virgin / Liberty Global to switch everyone over at once, and then roll back to native IPv4 (without IPv6) for the relatively few who want it. Trickling out IPv6 is such a bad package deal that it's never even mentioned as an IPv6 deployment M.O. --- the normal use of "long tail" is to refer to the long tail of legacy IPv4, slowly fading away into the sunset as IPv4-only equipment dies or is modernized.
PS. "IPv4-as-a-Service" (in the sense of an extra revenue stream for ISPs) is only ever mooted as a long-term goal, an option that becomes viable only after IPv4 usage has become niche. It's unlikely to affect the decision process for present-day ISPs when considering how best to transition to an IPv6 world, because a niche IPv4 is beyond the visible horizon.
Morgaine.
"If it only does IPv4, it is broken." -- George Michaelson, APNIC.
