IPv4 Address Exhaustion: How the Internet Ran Out of Room

IPv4 address exhaustion is the story of how a system built for four billion addresses slowly ran dry — and how the internet kept growing anyway. Here is what happened, why, and what came next.

When the original internet addressing system was designed in the early 1980s, its creators chose a 32-bit number for every address. That gives roughly 4.3 billion unique combinations. At the time, that felt effectively infinite — the network was a research project connecting a few hundred machines, and the idea that ordinary households, phones, watches and doorbells would one day each want an address was pure science fiction. IPv4 address exhaustion is the name we give to the slow-motion collision between that fixed supply and the internet's explosive growth.

If you want the ground-level picture of what an address is before we talk about running out of them, our guide to what an IP address is is a gentle place to start. This article picks up the bigger question: how do you run out of four billion of anything?

Why IPv4 address exhaustion was inevitable

Four billion is a big number, but it shrinks fast under real-world pressure. Three forces made exhaustion a matter of "when," not "if."

The internet went mainstream. Through the 1990s and 2000s the network stopped being an academic curiosity and became something almost everyone used. Each new home, business, server and mobile device wanted connectivity, and the count of things wanting to connect grew far faster than anyone had planned for.

Early allocations were generous. In the internet's first years, addresses were handed out in enormous, tidy blocks. Some organisations received a block of over 16 million addresses — a so-called "Class A" — when they needed only a fraction of that. Those early grants locked away a large slice of the total space that could never be fully reclaimed.

Not every address is usable. A meaningful chunk of the 4.3 billion is reserved for special jobs: private ranges, the loopback address, multicast, and other technical uses. Our tour of reserved and special IP addresses covers those in detail. Subtract them, and the pool available for the public internet is smaller than the headline figure suggests.

Who hands out addresses, and how they ran dry

Addresses are not distributed at random. There is a tidy hierarchy. At the top sits IANA (the Internet Assigned Numbers Authority), which holds the master pool. Below it are the five Regional Internet Registries (RIRs), each responsible for a part of the world:

IANA gives large blocks to the RIRs; the RIRs give smaller blocks to internet service providers and large organisations; and those, in turn, assign individual addresses to customers. You can read more about the bodies at the top of this chain in who actually runs the internet.

Exhaustion happened in stages. First, IANA distributed the very last of its central free pool to the RIRs in early 2011 — the top of the tap ran dry. After that, each RIR kept serving its region from its own remaining stock. Because the regions grew at different speeds, they hit their own limits at different times, with the fastest-growing regions running short first and others following over the following years. The moment each registry reached its final reserves, it switched to strict rationing: tiny allocations, waiting lists, and rules to keep a small emergency supply for brand-new networks.

The workarounds that kept the internet running

Here is the surprising part: the internet did not break. It adapted, using a handful of clever techniques that stretched the existing supply.

NAT: sharing one address among many

The single biggest lifeline was Network Address Translation. Instead of giving every device its own public address, NAT lets an entire home or office share one public IPv4 address, with a router juggling the traffic behind the scenes. This is why the phone, laptop and smart TV in your home can all be online through a single public address like 192.0.2.1. If you have ever wondered how that trick works, our explainer on what NAT is walks through it, and the related idea of public versus private IP addresses explains the two kinds of address involved.

CIDR: ending the wasteful class system

Earlier, addresses were carved into rigid "classes" of fixed sizes, which wasted huge amounts of space. Classless Inter-Domain Routing (CIDR) replaced that with flexible block sizes, so a network could receive exactly as many addresses as it needed rather than being forced up to the next giant tier. Our guide to subnet masks and CIDR unpacks the notation you will see, like /24.

Reclaiming and trading addresses

As pressure grew, unused blocks from the generous early days were returned or sold. A genuine market in IPv4 addresses emerged, where organisations with spare addresses transfer them, under registry rules, to those who need them. Addresses became a real asset with a real price — a striking outcome for something once given away freely.

IPv6: the actual fix

All of the above buys time. The permanent answer is IPv6, a redesigned addressing system with 128-bit addresses. Where IPv4 offers about 4.3 billion addresses, IPv6 offers a number so vast it is hard to picture — around 340 undecillion. An IPv6 address looks like 2001:db8::1 rather than the familiar dotted-decimal 192.0.2.1.

The catch is that IPv4 and IPv6 are not directly compatible: a device speaking only IPv6 cannot talk straight to one speaking only IPv4. So the world runs both at once during a long transition, a period that has stretched on far longer than optimists once hoped. We explore the two side by side in IPv4 vs IPv6, and the reasons the switch has crawled along in why IPv6 adoption has been so slow.

What exhaustion means for you

For most everyday users, IPv4 exhaustion is invisible. Your provider almost certainly places you behind NAT, so you share a public address with others and never notice. You may already have IPv6 connectivity without realising it. The effects show up more for businesses: obtaining a fresh block of public IPv4 addresses can now cost real money, which nudges organisations toward IPv6 and toward more efficient use of the addresses they hold.

The larger lesson is a hopeful one. The internet faced a hard, finite limit baked into its own foundations decades ago — and rather than collapsing, it engineered its way through with a mix of clever workarounds and a properly designed successor. Curious what address you are using right now? You can always check over at IP Animals.