Your IP changes without warning, and your carrier charges extra to keep it still. Nine diagrams show why that happens, what a static IP unlocks and when you don't need to pay for one.
an IP is the address the internet sees · dynamic is lent and may change, static is reserved just for you · the price climbs because IPv4 ran out and a reservation can't be shared · only pay if the other side demands an address that doesn't change
An IP is the address that identifies a connection on the internet, like a house number on a street. When you open a site, it needs to know where to send the reply. That address is the public IP, and it is what this post is about. The devices inside your home have another one, an internal address that only works in there.
→ Static or dynamic describes only the green part of the drawing: the address your carrier gives you.
A dynamic IP is lent. It works like a hotel room: while you're there it's yours, but next stay it may be a different one. A static IP is like your house number: it doesn't change, and nobody else uses it. For browsing the difference is invisible. It shows up when someone needs to find you.
Nobody decides to swap it. The router asks the carrier for an address every time it connects, and the carrier hands over whatever is free in its stock. When something interrupts the connection, the IP goes back to stock and another customer may get it. On reconnecting you receive whatever is available at that moment.
→ A modem restart isn't the only trigger. Power cuts, signal loss and carrier maintenance do the same.
The carrier doesn't have one address per customer. It has a stock, and hands it out in three ways: splitting one IP among many (CGNAT), lending one IP at a time (dynamic) or setting one aside just for you (static). Someone behind CGNAT browses normally, but nobody outside can reach their network.
Many customers share the same public IP. Cheapest for the carrier. For you, nobody outside can reach your home.
An IP just for you, only while the connection lasts. Then it goes back to stock and may be another one. Default for many home and small-business lines.
An IP set aside just for you, online or offline. It never goes back to stock or to another customer.
You can find out in one minute which of the three you are in:
Each step consumes more of the scarcest item on the bill: the IPv4 address. Split among many, it's cheap. Lent to one customer at a time, it costs more. Held for one customer all the time, it's the most expensive, and it usually comes in a business plan with different support and contract. The carrier isn't inventing a charge: it is billing for the reservation of a resource that ran out.
IPv4 has about 4.3 billion addresses in total, and the regional registries have run out of free stock. An address became a pricey resource.
It can't be lent to anyone else, not at night, not with your modem off.
It usually comes in a business plan: faster support, and sometimes a connection-quality guarantee.
If all you do is browse, nothing. A static IP matters when contact starts from the outside: you want to watch the shop camera on your phone, or a system only accepts connections from a registered address. With a dynamic IP that works until the day the IP changes. With CGNAT, it doesn't work at all.
Works every time. The address you saved on your phone never expires.
Works until the IP changes. Then it breaks, and you only find out when you need it. The workaround is DDNS: a name that follows the IP.
There is no front door: the IP belongs to everyone. Only a tunnel/VPN or moving to a public IP fixes it.
The uses where it is worth the money:
Cameras, recorder, server or shop system reachable from outside, always at the same address.
Banks, APIs and firewalls that only accept the IP you registered. If the IP changes, access drops.
Head office and branches linked by an address that doesn't change, no reconfiguring after each swap.
A site, ERP or system hosted at your company, with a stable address for customers and staff.
A self-run mail server needs a stable IP to build reputation and have reverse DNS.
PBX/VoIP and software licensed per IP work better with a predictable address.
Many people buy a static IP as a precaution and pay every month for something they never use. Before paying, answer in order. If the answer is "I don't need it", you save. If it is "there's a cheaper path", you save too. A static IP enters when the other side demands an address that doesn't change.
→ In automation the most common case is the first one: a system that only allows access from a registered IP. It is the kind of detail that shows up when we wire CRM, spreadsheets and APIs together, as in the n8n post.
The order that avoids the classic mistake: buying a static IP before knowing whether anyone demands an address that doesn't change.
List what must be reached from outside (camera, server, system) and what requires registering your IP (bank, API, firewall).
Run the 1-minute test from diagram 08: is your IP public, or are you behind CGNAT?
Walk through the four questions of diagram 09. In most cases the answer shows up right there.
If the answer isn't a static IP, try DDNS or a tunnel before paying. They are usually free or very cheap.
If it is a static IP, ask the carrier in writing: is the address public and dedicated, the monthly price, the delivery time and whether it comes with its own contract and support.
After signing, note the IP, register it in the systems that require it and repeat the diagram 08 test to confirm the address doesn't change.
A dynamic IP is lent by the carrier and changes with a restart, an outage or the end of the lease. A static one is reserved just for you, and costs more because IPv4 is scarce and a reservation can't be shared. Pay when a system requires your registered address or someone outside must always find you in the same place. Otherwise, DDNS or a tunnel usually do the job.
A static IP is a public address reserved just for you, and it doesn't change. A dynamic IP is an address lent by your carrier, and it can change when the router restarts, the connection drops or the lease expires. For browsing, video and social media the difference is invisible. It shows up when someone outside needs to find you or when a system requires your registered address.
Because an IPv4 address is a limited resource and the regional registries have run out of free stock. A static IP is held for a single customer all the time, even with the connection off, and usually comes in a business plan with different support. A dynamic IP goes back to stock and is reused by other customers.
Only if you want to watch them from outside over the internet and don't use alternatives. With a dynamic IP you can use DDNS, which is a name that follows the IP, or a tunnel or VPN. If the carrier uses CGNAT, only a tunnel, a VPN or a public IP will work.
CGNAT is when the carrier makes several customers share the same public IP. To test it, compare the WAN IP on the router page with the IP a lookup site shows. If they differ, or if the WAN starts between 100.64 and 100.127, you are behind CGNAT and nobody outside can reach your network directly.