It's cost-shifting, and not understanding the thing they seek to regulate
The network can only relate connections to a subscriber. In IPv4 the subscriber is allocated a particular IPv4 network. This was a /16 (~65,000 addresses) at the start of the internet and is a /32 (one address) these days. That doesn't change with IPv6 -- the subscriber is allocated a particular IPv6 network. That is currently a /32 for the largest of sites (<4 billion subnets) and a /60 for the smallest of sites (16 subnets, enough to separate a home's laptops, phone, TV and IoT so the home router can have access policies between these).
The agencies complaint is essentially one of an error in their "systems analysis". They assumed that a subscriber having one address was the way it always was and the way it always would be. Both assumptions are incorrect. Now the agencies are trying to change the world to match their deployed software rather than fixing the errors in their systems which arose from their faulty understanding of the technology. In short, that their interception and monitoring software needs to monitor subnets (of which single addresses are a special case) not only be able to monitor single addresses.
What's really odd is that the agencies are shooting themselves in the foot. IPv4 only tells the agencies the subscriber home router, as everything behind that is NATed. IPv6 tells the agencies the subscriber home router and allows a device behind that router to be identified. That is, after raiding the house the agencies can quickly determine which device accessed the resource. That's mostly true even of privacy addressing, as the device usually logs the privacy address used.
The power to make the world match the way you want it to be only appears in two sorts of literature: fantasy and histories of totalitarian states. Neither is a suitable model for agencies' behaviour.
(Anon, as personal view, not employers)