IPv6 Proxy

An IPv6 proxy routes your request through an IPv6 exit address rather than the IPv4 addresses that almost all proxy supply uses. Vendors sell them cheaply because of simple arithmetic. IPv4 contains about 4.3 billion addresses in total. A single IPv6 /64, the standard allocation handed to one home subscriber, contains 2^64 addresses, roughly four billion times the whole IPv4 internet. Scarcity is what makes IPv4 proxies expensive, and IPv6 has no scarcity, so the per-address cost collapses.

The constraint that kills most use cases sits on the other end of the connection. The target has to publish an AAAA record and serve over IPv6 or there is nowhere for the request to go. W3Techs put IPv6 support at 31.1% of all websites in its July 2026 survey (w3techs.com, read 2026-07-27), which leaves roughly two thirds of the web unreachable this way. The failure mode is worse than a plain error: many HTTP clients fall back to IPv4 when the AAAA lookup comes back empty, so the request leaves from your real datacenter egress with no proxy in front of it. You get a 200 and clean HTML from an address you never meant to expose, and nothing in the logs flags it.

On the sites that do serve IPv6, defenses generally reason about prefixes rather than individual addresses. A /64 is the smallest block a single subscriber is expected to hold, so scoring the whole /64 as one identity, or a whole /48 for a business allocation, is the obvious design. That erases the address-count advantage completely: a million addresses inside one /64 behave like one address under a prefix-level rate limit. Vendors rarely publish their prefix diversity and targets rarely publish their scoring policy, so treat this as the safe default assumption rather than a measured fact.

The practical step is to test before you buy. Run an AAAA lookup across your actual target list and count how many answer. If it is under half, IPv6 is a supplement to a specific set of jobs, not a supply strategy. Then force IPv6-only resolution in your client so a missing AAAA fails loudly instead of leaking. Cheap IPv6 volume is genuinely useful against targets that serve it and do not defend hard. It is not a substitute for IPv4 residential supply on defended sites, and no amount of address count changes that.