What Is an IPv4 Proxy and How Does IPv4 Work

IPv4 defines 4,294,967,296 possible addresses, and 3.687 billion of them had been allocated by the end of 2025. The pool has stopped growing. Geoff Huston’s annual address report on the ARIN blog recorded a contraction of roughly 237,000 addresses across 2025, about 0.01 percent of the total. Every hosting account, mobile subscriber and proxy server now draws on that fixed supply, which is why a single IPv4 address carries a market price, an inherited reputation history and, in the RIPE region, a waiting list.

What IPv4 is

IPv4 is one version of the Internet Protocol, the standard 32-bit IP. The full name in English is Internet Protocol Version 4. This protocol is designed for the transmission of traffic on the Internet. It represents a value of four decimal numbers separated by dots. What it might look like: 55.113.22.44. Those 32 bits allow 4,294,967,296 unique combinations, and the free supply ran out long ago. IANA passed its last blocks to the regional registries in 2011 and now holds three /24 prefixes, 768 addresses in total, unchanged for thirteen years.

How IPv4 works

Information transfer via IPv4 protocol takes place using TCP/IP technology. Every device on the network is assigned an address, and each packet carries a source and a destination so routers can forward it hop by hop. The one address per user assumption behind that design no longer holds. Cloudflare’s research team describes how a single IPv4 address can now represent hundreds or thousands of people because of carrier-grade NAT, VPNs and proxy middleboxes, so blocklists and rate limits built on the idea of one address, one accountable user end up hitting unrelated groups.

How much IPv4 space is left in 2026

Registry stock is a rounding error against demand. Across all five regional registries, roughly 3.9 million addresses sit in the available pool and 11.2 million in reserve, which together equal about 0.41 percent of the 3.687 billion already allocated. Around 33 million addresses changed hands through registry transfers during 2025 alone. The combined free and reserved pools of every registry on the planet would cover less than half of a single year of trading, which is the practical reason new address space now comes from the secondary market instead of from an application to a registry.

Source: Geoff Huston, IP Addresses Through 2025, ARIN, January 2026.

What an IPv4 address costs

Prices for IPv4 fell through 2025 and kept falling into 2026. Huston’s series built from IPv4.Global auction records puts the mean at 22 US dollars per address in the forty days to 10 January 2026, with a low of 9 dollars for a /14 block, after a decline of roughly 50 percent across 2025. A half-year review published on CircleID counts 596 transactions covering 5,016,064 addresses in the first half of 2026, at an average of 20.04 dollars and a median of 20.00 dollars, with ARIN blocks averaging 18.95 and RIPE blocks 20.95.

Renting an address from a cloud provider costs more than owning one within months. AWS has charged 0.005 dollars per public IPv4 address per hour since 1 February 2024, which RIPE Labs calculated as 43.80 dollars a year for an always-on address. At the 2026 market average of 20.04 dollars, a purchased address covers that annual rate in roughly five and a half months. The comparison has a floor: a /24, meaning 256 addresses, is the smallest block that routes globally, so the realistic entry point is about 5,100 dollars plus registry membership and compliance work.

Why IPv6 has not replaced IPv4

IPv6 crossed the halfway mark in 2026 without reducing the need for IPv4. Google’s measurement recorded 50.10 percent of its users arriving over native IPv6 on 28 March 2026, as reported by Internet Society Pulse. APNIC Labs, which weights samples by estimated internet population per economy, measured 42 percent worldwide in April 2026. Huston’s routing review for 2025 puts it plainly: 57 percent of the user base was still IPv4 only at the end of the year, so a service reachable over IPv6 alone loses most of the internet.

What makes an IPv4 proxy usable

A proxy server acts as an intermediate link between the user and the requested web resource. It replaces the address the destination site sees with the address of the proxy. Compatibility is the practical argument for choosing IPv4 over IPv6 here. Because most users and many corporate networks remain IPv4 only, sites are built to accept IPv4 first, and some platforms accept nothing else. The routing system reinforces that: /24, /23 and /22 prefixes made up 84 percent of the IPv4 routing table at the end of 2025, and a /24 is the smallest advertisement that propagates across the entire default-free zone.

Address history matters more than address novelty in 2026. Huston’s transfer analysis shows that more than 55 percent of the address volume traded during 2025 was originally allocated between 2000 and 2012, and that 40 percent of the volume was twenty years old or older. Genuinely new space is not being issued in meaningful quantities, so almost any address entering a proxy pool today carries thirteen to twenty five years of use by earlier holders. Reputation is inherited with the block rather than created at the moment of purchase, which is why a provider’s replacement policy matters.

How platforms classify IP addresses

Large platforms no longer treat an IP address as a single category. Cloudflare published a classifier that separates single-subscriber addresses, carrier-grade NAT addresses and VPN or proxy addresses, trained on more than 200,000 CGNAT addresses, 180,000 VPN and proxy addresses and roughly 900,000 addresses belonging to individual subscribers. The model reports 0.98 accuracy and a macro-average AUC of 0.9946. Prefix-level features contributed most to that result, because CGNAT occupies large contiguous blocks while VPN and proxy addresses scatter across many prefixes. The company a /24 keeps influences classification as much as the individual address does.

Shared addresses attract more restrictions than their traffic justifies. In the same Cloudflare study, the median share of bot-scored requests was almost identical for CGNAT and non-CGNAT addresses, 4.8 percent against 4.7 percent, and the mean was actually lower for CGNAT, 7 percent against 13.1 percent. Addresses identified as CGNAT were still rate-limited three times more often, because many unrelated people behind one address trip customer-configured firewall rules. A dedicated address avoids that specific failure mode, since the only behaviour attached to it is its own user’s.

Where IPv4 proxies are used

Business use of IPv4 proxies concentrates on tasks that need a stable, separate outbound address per workflow. Agencies handling client accounts, research teams collecting publicly available pricing data, QA engineers opening regional builds of their own product and marketers verifying how campaigns render in a given country all rely on an address that maps to a known location and is not shared with unrelated traffic. Each of these tasks belongs inside the terms of service of the platforms involved, because a proxy changes the network path and nothing about the rules that apply to an account.

• Market and competitor research. Collecting public prices, listings and search results at a request rate the target site permits.
• Ad verification. Checking how a campaign displays to users in a specific country or city.
• Regional QA and testing. Opening a company’s own product from another region before release.
• Account separation for agencies. Keeping client workspaces on distinct addresses where the platform permits multiple managed accounts.
• Access to services from a region where the company operates. Reaching tools and dashboards that restrict connections by network location.

How to choose an IPv4 proxy

Selection comes down to five checks that can all be made before purchase. The variables that decide whether an address works are measurable: who else uses it, which registry and country the block is registered to, whether it appears on public blocklists, and what the provider does when an address stops working. Price per address is the least informative signal on that list, because an unusually cheap address normally means a shared address or a block with a heavy history behind it.

1. Check exclusivity. A dedicated IPv4 address serves one customer. A shared address serves up to three, which means shared reputation and shared rate limits on every service they all touch.
2. Match registry and country to the task. The country recorded in RIR data reflects the holder’s registration, not necessarily what geolocation databases report, so verify the address against the service you plan to use.
3. Test before committing volume. Open the target platform through a single address first and confirm it loads without a verification wall.
4. Ask about the replacement policy. Addresses inherit history, so a provider should swap out any address that arrives already blocked.
5. Confirm protocol support. HTTP, HTTPS and SOCKS cover most software. Anti-detect browsers and automation tools normally document which of the three they require.

The economics have inverted since 2021. Addresses trading above 45 dollars at the end of 2021 averaged 20.04 dollars in the first half of 2026, while the underlying scarcity did not ease at all: registries hold 0.41 percent of the allocated pool, and 57 percent of internet users still reach services over IPv4 only. For anyone buying proxy capacity, that combination means the price of an address matters less than its history, its subnet neighbours and its registered location, because those three properties decide whether a platform treats the connection as an ordinary user.