How to Get a Public IPv4 Address Without an ASN or BGP in 2026

An ASN and a BGP session are only needed to announce address space yourself. Most people who want a public IPv4 address just want it to work, and there are two practical ways to get one without touching a registry at all.

Either ask whoever already hosts the server, or route one through a tunnel when nobody is hosting it for you yet. This guide covers both, with the actual setup steps and the mistakes that make a routed address look configured while it still fails.

Your server already has a host

A VPS, dedicated server, or cloud instance sits on a network the host already announces. Adding another public IPv4 there is normally a support ticket or a panel click, never a registry transaction.

None of this touches BGP. The host already announces its space, so adding an address just extends what it routes to the account.

Rent a public IPv4 without an ASN

A home connection, a machine with no host willing to hand out more addresses, or a server that needs an address independent of its current provider all share the same gap: nobody is announcing a route on its behalf. A routed-tunnel provider closes that gap. It keeps the announcement on its own network and delivers the address to the server over a tunnel.

Checking whether this applies takes a minute: compare the address the router shows with what a public IP checker reports from outside. A mismatch, or a router address inside 100.64.0.0/10, means CGNAT is in the way. The full check is in Why CGNAT stops you hosting a server at home.

Choosing a tunnel protocol

All three protocols can carry a routed address. They differ in what they need from the connection on the customer's end.

Most home and CGNAT setups end up on WireGuard for that third row. The full protocol comparison, MTU and key management included, is in GRE vs. WireGuard.

Setting it up in five steps

1. Confirm the problem is real. Compare the router's WAN address with a public IP checker before configuring anything.
2. Pick a protocol. WireGuard for most CGNAT connections, GRE when the server already has a stable public IP to tunnel from.
3. Order the tunnel and the address. A routed-tunnel provider issues both together, with no registry membership or ASN involved.
4. Configure the tunnel and the return route. The server needs a policy route so traffic sourced from the new address leaves through the tunnel, not the default connection. The command-by-command version is in How to get a public IP for a home server.
5. Test from outside the network. A phone on mobile data, not a laptop on the same Wi-Fi, is the only test that actually proves the address is reachable.

Mistakes that make it look broken

• No return route. The tunnel accepts inbound traffic but replies leave through the regular connection instead, so nothing ever comes back.
• No keepalive on the CGNAT side. WireGuard's NAT mapping expires when the tunnel sits idle, and the next inbound connection has nowhere to land.
• Testing from inside the same network. A local test can succeed over the LAN while the public path is still broken.
• A firewall rule scoped to the wrong interface. The service port has to be open on the tunnel interface specifically, not just on the one the server used before.
• Assuming the address moves itself. A routed address stays where the tunnel points it. Changing servers or providers means repointing the tunnel, not editing DNS.

Getting the address itself

Several companies route addresses this way, compared in full in Best IPv4 providers in 2026. Taipan Transit is one of them, listing routed IPv4 at €0.70 per address per month with no minimum count, on tunnels from €2 a month, cheaper than the cloud and dedicated-hosting rates compared in What an IPv4 address costs in 2026. Every announcement carries a signed ROA, and a Letter of Authorization is available on request.