IPv4 Getting Pricey, IPv6 Still Scary: The Real IP Address Situation in 2025

It's 2025 and the IP address saga continues. IPv4 ran out, prices are jumping around, we've been talking about IPv6 for thirty years, but the actual transition is stuck. Let's figure out what's really happening.

How IPv4 Ended (Spoiler: It Didn't)

IPv4 was invented in 1981. The protocol runs on 32 bits and provides roughly 4.3 billion addresses. In the eighties, that seemed like a huge number. Nobody thought that in forty years everyone would carry three or four devices in their pocket.

The problem was spotted back in the late 1980s—internet growth went exponential. The last free blocks from IANA were handed out in February 2011. That was 14 years ago. IPv4 is still here.

By now, all five regional registries have given out their last addresses, keeping only scraps for special needs. But addresses keep working, the internet didn't collapse. How's that?

IPv4 Prices: Roller Coaster Ride

When a product runs out but demand stays—the market kicks in. That's exactly what happened with IPv4. A whole industry of buying and selling addresses appeared.

Prices in 2025 look something like this:

• $35 to $60 per address—average range, depends on volume
• Small businesses pay $40-55 when buying small blocks
• Large /16 blocks dropped from $50 (mid-2024) to $21 (May 2025)

The price drop on large blocks is interesting. Owners of big address pools figured the market peaked and started selling en masse. Some are breaking up blocks trying to hold prices, but the effect backfired.

Paradox: addresses are simultaneously scarce and available. You just pay for them now, sometimes quite a bit.

Leasing Instead of Buying

High prices spawned a leasing market. Leasing is cheaper than buying, especially if you need addresses temporarily or want flexibility.

The advantages are obvious. Pay monthly instead of dropping a pile of cash upfront. Need more addresses? Lease them. Project closed? Stop paying. Plus decent providers check IPs against blacklists.

Amazon dumped $2.5 billion just buying IPv4, snapping up 100+ million addresses. But Amazon can afford it. For regular businesses, leasing often makes more sense.

NAT: The Crutch Holding Everything Together

While IPv6 is still warming up, technologies extend IPv4's life.

NAT (Network Address Translation) is familiar to everyone—that's when your home router lets all your devices access the internet through one IP. Works, convenient, familiar.

CGNAT (Carrier-Grade NAT) is more serious. Providers split one public address across thousands of subscribers. This solves the shortage problem but creates new ones: harder to configure incoming connections, latency grows, diagnostics become a quest.

Worse: when CGNAT fails, thousands of users fail at once. Each NAT layer adds a failure point.

IPv6: A Thirty-Year-Old Solution Still Being Deployed

IPv6 was developed in 1998 as an IPv4 replacement. The protocol runs on 128 bits and gives you about 340 undecillion addresses—a number with 38 zeros. You could give every atom on Earth an address and still have 99% left over.

Seems like problem solved? Not quite.

Where We Are Now: 2025 Numbers

The adoption figures vary, but here's the picture:

• Globally—43-48% according to Google
• USA crossed 50% in February this year
• Leaders: France (80%), Germany (75%), India (72%)
• Asia overall—around 50%
• Africa—only 4-5%

Summer 2025 saw eight more countries cross 50%: Brazil, Guatemala, Hungary, Japan, Mexico, Puerto Rico, Sri Lanka, Tuvalu. In Tuvalu, by the way, the surge is linked to Starlink—Musk launched his internet there, which runs IPv6-only.

Why So Slow?

The protocol is almost thirty. But the world still hasn't switched. Tons of reasons:

Different worlds. IPv4 and IPv6 are two incompatible protocols. They don't talk to each other directly. You need intermediate tech: dual-stack, tunnels, NAT64. Adds complexity.

Money. The transition costs serious cash. Update or replace equipment, reconfigure networks, train people—all this hits the budget hard. A big company can spend millions and months of work.

Old hardware. Lots of routers, servers, applications simply don't do IPv6. They were built for IPv4, period. Replace them? See the point about money.

No urgency. As long as NAT works and IPv4 can be bought or leased, businesses don't see the rush. IPv4 functions, customers are happy, why change anything?

Don't know how. Admins grew up on IPv4. Retraining for IPv6 takes time and effort. Plus the protocol itself looks scary with its hexadecimal numbers and colons.

Chicken and egg. Only a third of websites support IPv6 in 2025. Providers have no reason to deploy what sites don't use. Sites have no reason to support what providers don't offer. Who should go first?

Who's Ahead, Who's Behind

Interesting thing: mobile operators and home providers are racing ahead, while corporations are dragging.

Mobile networks. Reliance Jio in India—95%+ on IPv6. T-Mobile USA—over 90%. They have it easier: fresh infrastructure, less legacy stuff.

Home vs offices. IPv6 traffic is higher on weekends than weekdays. People at home use modern providers with IPv6. At offices—legacy systems on IPv4.

Corporations. Everything's complicated here. Old applications, multi-tier infrastructure, tons of security policies, need to support both protocols—all this slows the process.

What Should You Do

Startup or Small Business

Start with IPv6 right away. Most VPS providers give it for free. Need IPv4 for compatibility? Lease it, don't buy. Work dual-stack—that's the norm now.

Medium and Large Business

Look at what you've got. What hardware can't handle IPv6? Where are the bottlenecks? What will upgrades cost?

Don't try to move everything at once. Start from the edge, move toward the center. Do your admins understand IPv6? If not—train them. Does your ISP provide IPv6? Cloud provider? If not—pick another one.

Hosting Providers

IPv6 should be default. It's not a feature anymore, it's a basic requirement. Offer the combo: IPv6 plus IPv4 leasing for those who need compatibility. Explain to clients why they need it—most don't understand.

When Will This All End?

Predictions vary. Some say 2045. Others think the 2030s will be the tipping point.

Realistically? IPv4 will live another 10-15 years for sure. Dual-stack is our reality for years to come. Those sitting on IPv4-only will pay more and face growing problems.

Governments are starting to push. The US issued directive OMB M-21-07 for federal agencies. China, India, Europe are taking similar steps. The transition is accelerating, but not dramatically—gradually.

Bottom Line

IPv4 is getting expensive but still available. IPv6 is growing slowly but growing. Full transition—years away.

What this means for you:

Dual-stack will be with us for a long time. IPv4 will remain important, but you need to use it smarter. You can't ignore IPv6—prepare ahead, don't wait until you're forced. Leasing IPv4 often beats buying. Train your team to work with new tech.

No need to panic. But don't bury your head in the sand either. Those who start adapting now will win tomorrow. Those who delay will face problems at the worst possible moment.

The internet is changing. Slowly but surely. IPv6 is coming. The question isn't "if" but "when" and "how ready are you".