The internet ran out of addresses in 2011. Officially. The last block of IPv4 addresses was allocated to regional registries, and from that point on, getting a fresh IPv4 address became a matter of buying one from someone who already had it.
This was a known problem for decades. The fix was designed, standardized, and ready to deploy. And then the internet spent fifteen years very slowly getting around to deploying it.
That fix is IPv6. Here's everything you actually need to know.
The Original Problem: 4.3 Billion Isn't Enough
IPv4, the address format the internet was built on, uses 32 bits to represent an address. That gives you 2 to the power of 32 possible unique addresses, which works out to about 4.29 billion.
Four billion sounds enormous. Until you consider that there are about 5 billion smartphone users alone, plus laptops, desktops, servers, smart TVs, routers, game consoles, IoT devices, security cameras, fridges that can order milk (yes, these exist), and the entire infrastructure of the internet. Four billion addresses is nowhere near enough, and everyone knew this from the early 1990s.
The internet standards people started working on IPv6 in 1994. The specification was finalized in 1998. The year we're talking about is now. You do the math on how fast the rollout has been.
IPv4 Addresses: What They Look Like
IPv4 addresses look like this: 192.168.1.1
Four numbers, each between 0 and 255, separated by dots. Each number is 8 bits (an octet), and four octets gives you 32 bits total. The human-readable dotted decimal notation is just a friendlier way to display what is fundamentally a 32-bit number.
Some ranges are reserved. 192.168.x.x is private network space (your home router uses this internally). 127.x.x.x is loopback (the address your computer uses to talk to itself). Various other ranges are reserved for multicast, documentation examples, and other technical purposes. The actual usable public address space is somewhat smaller than 4.29 billion as a result.
IPv6 Addresses: What They Look Like
IPv6 addresses look like this: 2001:0db8:85a3:0000:0000:8a2e:0370:7334
Eight groups of four hexadecimal digits, separated by colons. Each group is 16 bits, so eight groups gives you 128 bits total.
How many possible addresses is 2 to the power of 128? It's 340,282,366,920,938,463,463,374,607,431,768,211,456.
That's 340 undecillion. Or roughly 340 followed by 36 zeroes.
To put that in perspective: you could assign every atom on the surface of the Earth an IPv6 address and still have addresses left over. Actually you could do that for many millions of Earths. The number is so large that address exhaustion is genuinely not a concern for any timescale that matters to civilization.
IPv6 addresses can be abbreviated. Consecutive groups of zeroes can be replaced with a double colon (but only once per address). Leading zeroes in each group can be dropped. So 2001:0db8:0000:0000:0000:0000:0000:0001 becomes 2001:db8::1. Much more manageable.
Why the Transition Has Been So Slow and Painful
If IPv6 was ready in 1998 and solves the address problem completely, why isn't everyone using it by now?
Because the internet is enormous, owned by thousands of different organizations, and IPv4 and IPv6 are not directly compatible.
A device that only speaks IPv4 cannot talk to a device that only speaks IPv6 without some kind of translation happening in the middle. That means to fully transition, every router, switch, operating system, application, and service needs to be updated. This is not a flip-the-switch situation. This is a replace-and-upgrade everything situation.
Large cloud providers and ISPs have mostly made the transition. Google, Facebook, and Amazon all run IPv6 for their own infrastructure. But enterprise networks, smaller ISPs, embedded systems, and legacy equipment have been slow. Really slow.
There's also just a lack of urgency in the short term, because of the band-aid solution that's kept IPv4 functional past its expiration date.
NAT: The Band-Aid That Became Permanent
Network Address Translation is the reason IPv4 has survived. It works like this.
Your ISP gives you one public IP address. Your home router sits between your devices and the internet. Internally, all your devices get private IP addresses (the 192.168.x.x range) that are non-routable on the public internet. When any of your devices wants to talk to something on the internet, the router translates the private IP to your single public IP and keeps a table of which internal device made which request, so it can route the responses back correctly.
From the internet's perspective, your entire household looks like one IP address. From your household's perspective, you have four devices, each with its own address.
NAT is technically clever and it works, mostly. It lets a single public IP serve dozens of devices. It helped defer the IPv4 exhaustion crisis by decades.
But it introduces complexity. Peer-to-peer connections become harder (see: WebRTC leak issues). Certain applications break or need workarounds. Network troubleshooting gets more complicated. And it doesn't scale infinitely.
IPv6 was designed for a world without NAT, where every device has its own globally routable address. Whether that's actually better from a privacy standpoint is a fair question.
Dual Stack: Running Both at Once
Most modern networks don't have to choose between IPv4 and IPv6. They run both, simultaneously, using a configuration called dual stack.
Your device can have both an IPv4 address and an IPv6 address at the same time. When you connect to a website that supports both protocols, your device and the server negotiate which one to use. IPv6 is generally preferred when available because it's faster and has lower overhead.
If the server only supports IPv4, your device falls back to IPv4. If the server only supports IPv6 and your ISP doesn't give you IPv6 yet, the connection might need a translation mechanism or might fail.
This is the transition path. Dual stack lets everyone slowly shift while maintaining backward compatibility. It's unglamorous but it works.
The IPv6 Privacy Implications
Here's something a lot of people don't know. IPv6 can actually be worse for privacy than IPv4 in some configurations.
Under IPv4 with NAT, you share a single public IP with everyone in your household, and that IP rotates when your ISP reassigns it. Linking that IP directly back to you requires going through your ISP.
Under IPv6, by design, every device can have its own globally unique, persistent address. Your laptop could have an IPv6 address that identifies it specifically and doesn't change. Websites could potentially track individual devices across sessions in ways that aren't possible with shared dynamic IPv4.
This is why IPv6 Privacy Extensions exist. This is a feature (RFC 4941) that generates randomized, temporary IPv6 addresses for outgoing connections that change regularly. Modern operating systems like Windows, macOS, and most Linux distributions enable this by default, so in practice you're probably okay. But worth knowing about.
If you're using a VPN, the VPN should route all IPv6 traffic too, not just IPv4. A VPN that only handles IPv4 will let your real IPv6 address leak out on dual-stack connections. This is a specific type of VPN leak to check for.
How to Check If You Have IPv6 Right Now
You can find out immediately whether your current connection has IPv6 support.
Go to the IPv6 Test tool and it will tell you whether you have a working IPv6 connection, what your IPv6 address is if you do, and whether your IPv6 connectivity is functional.
Roughly half of internet users in developed countries have IPv6 at this point, but adoption varies enormously by country and ISP. The US is around 50 percent. Belgium is consistently near the top globally. Some parts of the world are still mostly IPv4.
ISP Support: Who Has IPv6 and Who Doesn't
Major US ISPs have all deployed IPv6 to varying degrees. Comcast has been a leader in the US and has very high IPv6 deployment rates. T-Mobile's network is heavily IPv6, which is why mobile connections often show IPv6 addresses. AT&T and Verizon have deployed it. Charter Spectrum has been slower.
Business ISPs tend to lag residential ISPs on IPv6 deployment because enterprise customers are often more conservative and have legacy equipment that doesn't support it.
If you don't have IPv6 yet, it's probably your ISP's fault, not your device's. Modern devices support it natively. Older routers sometimes need firmware updates to handle IPv6 properly.
What This Means for VPN Users
If you use a VPN, you should check whether it handles IPv6 correctly.
Many VPN providers, especially older ones, only tunnel IPv4 traffic. In a dual-stack environment, your IPv4 traffic goes through the VPN but your IPv6 traffic bypasses it entirely and goes directly through your ISP. This means your IPv6 address (which might be uniquely tied to your device) is visible to every website you visit.
This is a real leak that's distinct from WebRTC leaks and DNS leaks.
Good VPNs in 2024 handle both protocols. Mullvad tunnels IPv6 properly. ProtonVPN handles it. Some VPNs, when they don't support IPv6 tunneling, just disable IPv6 entirely on your device while connected, which prevents leaks but removes your IPv6 connectivity.
Run the IPv6 test with your VPN on. If you see an IPv6 address that belongs to your ISP rather than your VPN provider, that's a leak worth addressing.
The Bottom Line
IPv4 served the internet well for forty years, it's held together with NAT as a patch job, and it's still the dominant protocol. IPv6 is the future and the present, deployed alongside IPv4 in dual-stack configurations, and has enough address space to last until the heat death of the universe.
You probably have both. Check with the IPv6 test tool, see which protocols your connection supports, and if you're using a VPN, make sure it handles both.