Google's IPv6 traffic passes 50%, while APNIC Labs measures 42%

Google's IPv6 statistics have passed the 50 percent mark for the first time: around half of the users reaching Google's services now do so over IPv6. The figure comes from the company's continuous monitoring of IPv6 connectivity among its users, and was highlighted on the APNIC blog, which calls it a significant milestone.

Two measurements, two numbers

APNIC's own measurement programme, run by APNIC Labs, records 42 percent worldwide IPv6 capability — a gap wide enough to need explaining. APNIC Labs distributes its tests through online advertising placed via Google Ads, which appears in browsers, games and apps wherever Google advertisements run. It seeks the broadest possible exposure in every economy, holds no personally identifiable information, and publishes only collations at ISP, economy and region level rather than raw measurements.

The difference comes from weighting. Because advertising demand — and with it sample volume — varies by day and by market, raw sample counts cannot simply be aggregated into a global figure. APNIC Labs aggregates the measured capability of each economy and weights it by that economy's estimated internet user population, using external sources such as World Bank statistics, so large internet populations such as India, China and Indonesia contribute proportionally more than daily sample volumes alone would suggest.

Adoption looks different in every economy

Google does not publish per-region IPv6 statistics and limits its per-economy data to totals, so the shape of adoption is hard to read from its numbers alone. APNIC Labs data shows that individual economies — India, Viet Nam and Saudi Arabia among them — follow curves that differ markedly from the global average. Within a single economy, APNIC's results generally align with Google's totals and with data from Cloudflare, Akamai and Cisco, which suggests the underlying measurements are comparable, while APNIC's global figure tends to sit below Google's.

Living in a two-protocol world

The long road to 50 percent is sometimes read as evidence of systemic failure. APNIC argues the opposite: deploying IPv6 has required substantial technical effort and capital investment, and operators who built commercially viable IPv4 networks earlier naturally sought to maximise the return on that investment, while newer entrants — especially in mobile, such as Reliance Jio's network in India — have found IPv6 the rational choice because it can reduce total cost of ownership.

Today's internet mixes direct IPv4, IPv4 mediated by address translation at home or carrier-grade NAT, and IPv6 — and managing that translation is not materially simpler than the alternatives, so claims that IPv4 is working fine overlook the operational complexity already in place. Interoperability has instead emerged at higher layers, where TCP, UDP and QUIC operate independently of the IP version beneath them. Where native dual-stack support is missing, the cause is usually pragmatic — operational complexity, or legal and regulatory requirements around data access and geolocation — rather than resistance. IPv6, the blog concludes, is now used every day, on fixed and mobile networks in developed and developing economies alike.