It's hard for an ISP to justify switching. It costs money now, so your competition who do not switch can use the money saved to outcompete you. Switching costs are also likely to fall the longer you wait. Customers don't demand it.
If I was CEO of an ISP, I would have ordered a plan, but not put it into action just yet.
I work at a small ISP and I consider this a misconception.
Any network equipment your ISP bought or gave you in the last 10-15 years supports IPv6. Equipment only has a certain lifespan anyways, so there really has been no "cost" to upgrade. Maybe a bit of software needed updating.
In my experience, the problem is customers don't care that much about IPv6.
When your customer is an IT professional who has been configuring networks the same old way for years, they expect to be sent the IPv4 subnet information. It would be a negative customer experience if we gave them an IPv6 subnet by default. Since they don't see the need and aren't used to it, it's bound to cause frustration for them.
We do support IPv6 for any customer that wants it.
I think IPv6 will happen when web sites can no longer get IPv4 addresses. Then people will start saying, "my favorite site is v6 only!", so the IT people hear about it and start to care.
> I think IPv6 will happen when web sites can no longer get IPv4 addresses
With vhosts and the proliferation of CDNs, will this ever happen? If my site is behind CloudFront, I don't need any IPv4 addresses of my own
The only thing that IPv6 really solves for end-users is peer-to-peer. If video games, VoIP, etc have lower lag on IPv6 (due to not having to go though a mediating server) customers might demand it.
If I was CEO of an ISP, I would have ordered a plan, but not put it into action just yet.