The post is about a milestone that sounds cleaner than it is. Google now sees more than half of user access arriving over IPv6, but APNIC’s own measurements put worldwide IPv6 capability closer to 42%. The gap comes from how they measure. Google sees traffic reaching Google properties, which overweights places and services where IPv6 is already strong. APNIC measures broader client capability across the public internet. That led people to a sharper conclusion than the headline. IPv6 has clearly won enough share to matter, especially on mobile networks and in newer builds, but it is still not a completed transition.
The strongest pattern people pointed to is that adoption is highly uneven by network type and geography. Mobile carriers and newer ISPs often run IPv6-first because buying or leasing
IPv4 space at scale is painful. That helps explain very high numbers in places like France and India. Older fixed-line incumbents were the main villains of the story. Virgin Media in the UK came up repeatedly as a case study in how an
ISP can promise IPv6 for more than a decade and still not ship it. Several commenters also tied Google’s weekday dips and weekend peaks to the same split. Consumer and mobile traffic tends to be more IPv6-heavy than office and corporate networks, so the line rises when people are off enterprise networks.
The practical bottleneck is not whether IPv6 works in principle. It is that IPv4 still works well enough for the actors who must do the migration work.
NAT,
CGNAT, IPv4 leasing, and translation layers have kept the old world limping along. That blunts urgency for incumbents, even though it pushes real costs into networks, cloud bills, and ugly workarounds. A lot of people framed the result as a supplement, not a replacement. You can be a heavy user of the internet today and still hit important IPv4-only gaps, with GitHub cited as the most embarrassing example. Others noted that large enterprises often have little incentive to touch internal IPv4 setups because their public-facing needs can be handled at a few egress points.
Where comments got most concrete was on what the long transition has broken or preserved. Supporters argued that IPv6 avoids CGNAT overhead, reduces collateral damage from shared IPv4 bans, and restores cleaner end-to-end connectivity for things like gaming,
WebRTC, and self-hosted services. Skeptics replied that consumer routers now default to stateful firewalls anyway, so IPv6 does not magically bring back open peer-to-peer networking, and in practice bad
CPE firmware, poor peering, and flaky IPv6 endpoints still create operational pain. Several operators described real incidents where IPv6 paths were slower or outright broken while IPv4 was fine, especially for package mirrors,
CDN paths, and tunnel-based setups. That pushed the conversation to a sober landing point. IPv6 is no longer hypothetical and no longer optional to understand, but it is also nowhere near a clean cutover. The internet is settling into a messy equilibrium where IPv6 carries a huge and growing share of traffic while IPv4 remains mandatory for compatibility and for too many important services.