Most IP addresses aren’t fixed—they change frequently due to factors like ISP policies, hosting providers buying or leasing more IPs, and more. These changes affect how IP data providers like IPinfo obtain and update their IP address information, and how customers ensure that they’re always getting the most up-to-date data.
So just how often do IP addresses change? Last year alone, IPinfo serviced more than 75 billion API requests for over 500,000 users. Our in-house data experts track trends like this one to help our customers optimize IP data. Here are our most recent findings.
Static IP addresses remain fixed, while dynamic IPs change regularly—anywhere from every few hours to every few months—based on a variety of factors, from ISP reassignment to geolocation changes. We’ve found that 1.3% of IPv4 addresses change cities every day, 16% change in a month, and 44% change in a year.
Average change rates for IPv4 addresses in IPinfo datasets in 2024:
Average change rates for IPv6 addresses in IPinfo datasets in 2024:
For downloadable database users, these changes influence how frequently companies need to download and refresh IP insights. That’s why IPinfo’s database is updated every 24 hours, giving users access to industry-leading IP data whether using our API or downloads. If IPs are changing location every day, even on a small level, then the data must reflect that.
Learn more about our IP address location accuracy.
Enhance user experience and strengthen your security with rich geolocation data.
Average change rates for IP address data in 2024:
While these changes are fairly small, the fact that some IPs change location every 24 hours is why IPinfo updates these insights with the same level of frequency.
Average change rates for IP address data in 2024:
Nearly 8% of IPv4 address city locations change weekly. Explore: where are most IPs located?
Average change rates for IP address data in 2024:
Within a month, 16% of IPv4 address city locations change.
Average change rates for IP address data in 2024:
At the year mark, almost half of IPv4 addresses have changed cities at least once.
IP address data is constantly evolving. That’s why IPinfo is so committed to IP data accuracy. Beyond the numbers, here are some real-life examples of why IPs change over time.
ISPs (internet service providers) need IP addresses for their customers. These IP addresses are seldom used to host content or services. Most of the time residential users will consume content (e.g. stream music or videos), but rarely host it. For this reason these networks are also called eyeball networks, because they’re used to “look at things.”
Ideally ISPs would assign one IP per customer for the whole lifetime of a customer. In this scenario an IP address would never move. However, there is a limited amount of IPv4 addresses (~3.7B addresses usable on the internet) and more people and devices than that on earth. This method is thus not feasible for most ISPs. Instead, ISPs will resort to two workarounds: dynamic IP addresses and shared IP addresses.
A dynamic IP is an address that is assigned to a customer for a limited amount of time. Typically a few days to a few weeks. If the address is not renewed within this time interval, then it is returned to the ISP pool of unused IPs to be assigned to other customers. This avoids wasting IP addresses if a customer router is powered off. This also means that the IP address might change location if it is assigned to another customer in a different location.
A shared IP is an address that is assigned to multiple customers at the same time. This is achieved with a technique called Carrier-grade NAT (also called CGNAT) where customers are assigned a private IP address (i.e. non-globally unique address) within the ISP network, and all the traffic exits through a central router to which the shared public IP is assigned. In this context IP geolocation might be imprecise if the IP is shared amongst customers over a large geographical area. Shared IPs can also be dynamic and be assigned to different groups of customers over time.
Those two techniques are also used by mobile carriers, which must deal with a large number of mobile devices, and the fact that those devices are on standby most of the time. Thus assigning static IPs to those devices would be wasteful.
IPinfo data provides several fields that are useful for ISP addresses:
Hosting providers are the opposite of eyeball networks: they are where the content is hosted. Because servers are always powered and must always be accessible, they are often assigned static IP addresses. This means that the geolocation of a hosting IP will never change, unless the IP is transferred to another datacenter, or another company.
Because server IPs (generally) can’t be shared IPs, hosting providers must own enough IPs for all of their servers (physical and virtual). This can become a challenge for large providers like Google Cloud or Amazon Web Services, which host multiple millions of servers. Because all IPv4 addresses have been assigned by regional internet registries (RIRs) to companies, the only way of obtaining IPv4 addresses today is to buy or lease them from other companies.
In 2024 there were several large transfers of IPv4 addresses: 3.2M from Ford to Google, 2.2M from Windstream to Amazon and 1.5M from Nokia to Microsoft, among others.
Regarding IPv6, we observe almost no transfers, which are not necessary since IPv6 addresses are abundantly available from RIRs for free, or a modest fee.
Another consideration with hosting providers is that their servers might change owners quite frequently. For example, one customer might set up a server for testing, or to host a website for a short amount of time, and then return the server to the provider. So while the geolocation of a server IP is stable, the usage that is made of the IP might change over time. At one point in time the IP might be used to host webpages, but at another time it might be used to host a VPN service.
Most businesses will use a standard ISP to access the internet. Sometimes with a dynamic and shared IP, sometimes with the option to get a static IP for a fee. However larger companies might want to own their own IP addresses, to be able to change ISP without changing IPs, or even to use multiple ISPs for redundancy.
In this scenario it is important to distinguish between the owner of an IP address, and the origin autonomous system (AS) of an IP address. The owner is the person or company that is registered as such in the RIRs. The origin autonomous system is the organization that is announcing the IP address to other autonomous systems on the internet. Sometimes these can be the same organization, but not always: one can own IP addresses but not necessarily the infrastructure to make these IPs reachable from the internet.
As such, it is possible to observe changes in the origin AS of an IP, but not in its ownership, although such changes are relatively rare.
Some users rely on VPNs and proxy servers to protect their privacy, which results in more frequent IP address changes. These tools help mask their real location, making it harder for websites to track them and reducing exposure to cyber threats. However, frequent IP changes from anonymizers can also pose challenges for businesses that rely on IP data for security, fraud prevention, and user experience. The chart below shows how frequently IP addresses that we identify as IP anonymizers change IPs.
The IP address change rates for IPv6 are generally lower than for IPv4. There are two main explanations for this: not all networks use IPv6 and thus it’s not as widespread as IPv4, and since there is an extremely large number of IPv6 addresses available it is easy, and recommended, to assign one IP address per device instead of sharing them as done with IPv4. IPv4 addresses have been exhausted, so they must be recycled.
As a result, IPv6 addresses can be considered more stable than IPv4 addresses from a geolocation and ownership point of view.
Some of our customers, knowing how often IP addresses change, use that information to decide how frequently they need to download our database. If a use case, for instance, requires the highest level of accuracy, these users may want to download our data every day – our data is refreshed every 24 hours.
However, if small deviations on the city level don’t affect your use case, you could download our data with weekly or monthly frequency. In general, we don’t recommend downloading the datasets only once a year because, as we’ve seen, almost half of IPv4 addresses change city location every year.
IPinfo provides accurate, real-time IP data that identifies VPNs, proxies, and other anonymizers, along with contextualized information about IP ownership and geolocation. This data empowers businesses to make more informed decisions while balancing security, fraud prevention, and user experience.
As always, our data experts are available to help you determine the best approach for your specific use case. Additionally, our team helps build out custom datasets or enterprise solutions such as these:
Our data experts enjoy helping customers use their IP address data to its full capacity.
Have a question or unique use case? Connect with a data expert today for customized solutions!
Locate users, customize experiences, eliminate site risks, and much more.
Max currently works as a data engineer on IP geolocation at IPinfo. Previously, he was a postdoctoral researcher at Sorbonne Université where he worked on multipath traceroute measurements.