An incorrect IP location does not usually mean that a website saw your GPS and got it wrong. IP geolocation is an estimate from the start.

(For the underlying concept, read What Is an IP Address, and What Can Websites Learn From It?.)

GeoIP databases estimate an IP's country, region, or city from address ranges, ASNs, ISPs, operator-published Geofeeds, routing and latency measurements, historical observations, and correction data. The result describes the approximate location of a network egress or address range, not necessarily the user's physical location.

It is very common for someone in one city to see a nearby city, state capital, or major metro area. A different country more often points to roaming, a VPN, a proxy server, corporate egress, transferred address space, or delayed database updates.

The short answer

• IP geolocation is not GPS and cannot identify a street address precisely.
• Country-level results are generally more reliable than city-level results, but both can be wrong.
• Centralized ISP egress, mobile networks, and Carrier-Grade NAT (CGNAT) commonly shift city results.
• A roaming eSIM can use a local cellular network while reaching the internet in another country.
• VPNs, corporate gateways, and proxies expose the egress node's location.
• Address ranges can be sold, leased, or moved before databases update.
• Lookup sites often disagree because their sources and refresh cycles differ.
• Read the country, ASN, ISP, network type, time zone, and confidence together instead of focusing on a city name.

Check your IP location, ASN, and network environment

About the Caylet model (1.6.0): Caylet distinguishes a confirmed database finding, absence from an available source, and no data. The last two do not prove safety. The composite IP risk score uses valid numeric values from Scamalytics and AbuseIPDB only. IP2Location and Feodo Tracker appear separately, while IPinfo Lite provides ASN context only.

How is IP geolocation produced?

An IP packet does not include a statement such as “this user is currently on this street.” A GeoIP service must map an IP or prefix to a location through indirect evidence.

1. IP address registration data

Regional Internet Registries maintain records about address ranges and the organizations that hold them, including:

• the address range
• organization name
• contact information
• registered region
• ASN

These records manage network resources; they are not a user-location service. A company registered in New York can operate addresses elsewhere, especially when it is a multinational carrier or cloud provider.

2. ASN and BGP routing data

An Autonomous System Number (ASN) identifies a network system that announces routes on the internet. ASN and Border Gateway Protocol (BGP) data can show a GeoIP provider:

• which network announces a prefix
• likely upstream carriers
• whether the operator changed
• where the network appears

BGP's purpose is to route traffic, not to publish exact geography.

3. Operator-published Geofeeds

RFC 8805 defines a standard format through which an ISP, business, or other operator can publish country, region, and city mappings for its IP ranges.

A Geofeed is useful operator evidence, but it is still not live device GPS. It describes how the operator believes the range should be located, and the standard is intended for coarse geolocation rather than precise residential identification.

4. Active network measurements

Some providers send requests from multiple measurement points and analyze:

• round-trip latency
• routing hops
• network topology
• timing differences between measurement sites

Latency can eliminate physically implausible locations and estimate proximity to a network node because data cannot travel faster than light. Routes are not straight lines, however. Packets can detour through remote backbones or encounter congestion, so latency is not the same as geographic distance.

5. Corrections from ISPs, customers, and users

GeoIP providers may accept ISP range data, business corrections, user reports, and location observations gathered through services. These can improve accuracy but require validation because submissions may be outdated, too broad, or simply wrong.

How accurate should an IP location be?

IP geolocation may be expressed at several levels:

1. continent
2. country or territory
3. state, province, or administrative region
4. city
5. postal code
6. latitude and longitude

The more specific the level, the greater the usual chance of error.

Country-level geolocation

Country data is often the most useful level for regional content, language or currency selection, content licensing, compliance screening, and initial risk review.

It can still be wrong when traffic uses international roaming, a VPN or proxy, multinational corporate egress, recently transferred address space, a satellite or global network, or an outdated database.

Region- and city-level geolocation

City results are less reliable. MaxMind explains that geolocation precision can range from roughly 5 kilometers to hundreds of kilometers and should not be used to identify a particular person, household, or street address.

If a result says “Charlotte,” a reasonable interpretation is:

The provider believes this IP's network location is broadly associated with the Charlotte area.

It does not prove that the user is inside the city limits or that the returned coordinates identify their building.

Accuracy radius and confidence

More complete products may report an accuracy radius and city or country confidence. A city with a 10-kilometer radius means something very different from the same label with a 200-kilometer radius. Displaying a city without confidence or radius can make the result look more precise than it is.

Why does an IP show another city?

1. Centralized ISP egress

Home broadband traffic does not have to enter the public internet in the customer's city. An ISP may aggregate several cities at a regional core. External routing evidence then points toward that core.

For example, a customer lives in a small city, the ISP's main internet breakout is in a nearby large city, and the entire prefix is mapped to the large city. This does not imply VPN use.

2. Small towns map to nearby metro areas

Without enough evidence for a small town, a provider may return a nearby major city, regional center, state or provincial metro area, or another representative location with higher confidence. MaxMind also notes that smaller towns may map to larger neighboring metro areas.

3. Dynamic address assignment

Many home connections use dynamic IPs. The same prefix may serve customers in different cities over time. A database based on an earlier assignment pattern can continue to show the old city.

4. Network restructuring

An ISP may merge networks, relocate a core gateway, split ranges, add data centers, or acquire another ISP. Routing changes immediately, while external databases need time to reevaluate the range.

Why are mobile IP locations especially imprecise?

4G and 5G networks generally do not assign every phone a public IPv4 address precisely tied to the city of its cellular connection.

Carrier-Grade NAT

CGNAT lets many mobile users share a set of public IPv4 addresses. A website sees the carrier's centralized internet breakout rather than the mobile network serving the handset.

A phone in Los Angeles can therefore reach the internet through a gateway in another city, which is the city shown by the database. This is ordinary network architecture, not failed GPS.

A moving phone does not necessarily change IP immediately

A phone can move between cities while the mobile network preserves its path through the same core node. Conversely, a stationary phone can receive a new IP after reconnecting. A mobile IP's city is unsuitable for real-time phone tracking.

Why can a roaming eSIM have an IP in another country?

Separate three concepts:

1. the country where the phone is physically located
2. the country of the cellular network it is using
3. the country where data reaches the public internet

They need not match.

Home Routing

With Home Routing, a phone connects to a local partner mobile network, but data returns to the eSIM's home or partner core before reaching the internet elsewhere. A phone in Japan may use a Japanese cellular network, route through a Singapore core, and appear to websites as a Singapore IP.

Local Breakout

With Local Breakout, data may reach the internet in the visited country or a nearby region, producing an IP closer to the user's location. The route depends on the eSIM, plan, and destination country.

An eSIM is not itself a VPN.
Roaming architecture can nevertheless place the IP in a different country from the phone.

How do VPNs, proxies, and corporate networks change location?

VPN

A site normally sees a VPN server's egress IP. A U.S. egress will generally geolocate to the United States even if the user is elsewhere.

The exact city can still be wrong because an advertised VPN location differs from the server, the service uses Anycast or a virtual location, registration differs from the data center, or the database is stale.

Proxy server

A proxy similarly exposes its egress IP. A residential proxy can use an address presented by a home or mobile network, but a residential-looking network type does not mean the user lives at that location.

Corporate and school networks

Multinational organizations may route offices through headquarters, a cloud security gateway, a SASE or Zero Trust platform, or a regional data center. An employee in Hong Kong can legitimately show a Singapore or U.S. IP because of corporate policy.

Why do transferred, leased, and moved IP ranges create errors?

IPv4 space is scarce and ranges can be transferred or leased between organizations. A U.S. company may sell a range to an operator in Europe or Asia. BGP routing changes first, while some GeoIP databases retain the old location. Different sites can then disagree for days or weeks.

Even after an operator submits a correction, the provider must receive and verify it, update its database, distribute the release, and wait for customer sites to deploy it. Corrections do not propagate instantly across the internet.

Cloudflare's dedicated-egress documentation warns that geolocation updates across databases may take at least several weeks, illustrating the gap between submitting an update and every site displaying it.

Why do IP lookup sites disagree?

The same IP may produce different cities, regions, countries, ISP names, network types, and VPN or proxy findings. This does not necessarily mean one service is fabricating data; they use different inputs and methods.

Compare whether the country, ASN, ISP or organization, network type, update time, and accuracy or confidence agree. A city disagreement alone does not establish that one provider is absolutely wrong.

Why do WHOIS, ASN, and IP geolocation differ?

They answer different questions.

A range may be registered to a New York company, announced by a nationwide ISP, used by a North Carolina subsidiary, and routed through a Charlotte core. All four facts can coexist.

ARIN also explains that its registry database is not a GeoIP service. Third parties may use registry records, but ARIN does not control their location inferences.

Does Anycast make geolocation harder?

Yes. Anycast allows the same IP to be announced from multiple geographic locations. Network routing directs a user toward an appropriate node.

One Anycast address may simultaneously serve New York, Los Angeles, London, Tokyo, and Singapore, making a single-city label an oversimplification. Anycast is common for CDNs, DNS services, DDoS protection, global cloud networks, and enterprise security gateways.

When looking up a website server, DNS resolver, or global security-service IP, do not assume the database's city is the only physical location.

Why can IPv4 and IPv6 show different locations?

A dual-stack network can use different providers or routes for IPv4 and IPv6. IPv4 might pass through a VPN while IPv6 goes directly to an ISP; IPv4 might use mobile CGNAT while IPv6 is native; or the two databases may update at different times.

This deserves more attention than a city-only mismatch. It may reflect two egress paths, a VPN that does not fully handle IPv6, or inconsistent provider data.

How does IP location differ from GPS or browser geolocation?

IP geolocation

• requires no user permission
• usually estimates only a broad location
• primarily reflects network egress
• is readily affected by VPNs, roaming, and corporate networks

GPS and device geolocation

• may combine GPS, Wi-Fi, cellular-network data, and sensors
• usually requires operating-system or browser permission
• can be far more precise than an IP
• primarily reflects device location, not network egress

Account or billing address

• is supplied by a user or financial institution
• describes account data, not necessarily a live location
• cannot by itself prove whether IP geolocation is correct

An IP in Singapore, GPS in Japan, and account address in the United States are not inherently contradictory. Travel, roaming, corporate routing, and account history determine whether the combination makes sense.

How should you read Caylet's IP location result?

Step 1: Start with the country, not the city

City differences of tens or hundreds of miles are common on mobile and centralized networks. If the country agrees, the ASN is plausible, and the time zone is consistent, a city difference alone is rarely cause for alarm.

Step 2: Review the ASN and organization

These fields help distinguish a home ISP, mobile carrier, corporate network, cloud or data center, VPN or proxy provider, and other specialized networks.

Step 3: Compare the IP and device time zones

Agreement is useful context, but disagreement can result from a device not updating after travel, a company computer fixed to headquarters time, roaming through another country, or remote desktop use. Time zone is one consistency signal, not proof of identity.

Step 4: Review WebRTC, DNS, IPv4, and IPv6

If the public IP shows the United States while WebRTC observes another country's external IP, DNS uses a different region, or IPv6 shows another country, the connection may have multiple egress paths or an incomplete proxy configuration.

Step 5: Distinguish source states

• true: the source explicitly detected the condition
• false: the source explicitly did not detect it
• unknown or unavailable: insufficient data

Unavailable never means definitely absent.

Use Caylet to check IP location and environment consistency

What should you do when an IP location is wrong?

For users

1. Check whether a VPN, proxy, Private Relay, or corporate security tool is active.
2. Test Wi-Fi and mobile data separately.
3. See whether IPv4 and IPv6 show different countries.
4. Compare the country, ASN, and ISP across two or three sources.
5. Test again later to rule out caching or a temporary route.
6. Contact the ISP if major services consistently show the wrong country.
7. Submit a correction to the specific GeoIP provider.

Restarting a router may assign another dynamic IP, but it does not guarantee a database correction.

For network operators

Consider verifying WHOIS or RDAP organization data, publishing an RFC 8805 Geofeed, linking it through an appropriate registry or routing object, submitting corrections to major GeoIP providers, checking whether leased ranges match actual routing, and supplying verifiable network and location information.

Sites will still adopt the changes at different times because their update cycles differ.

Does a wrong IP location indicate a security problem?

Usually not. A wrong city, coordinates at a city center, an ISP name different from a consumer brand, an adjacent IP and device time zone, or a mobile IP in another city does not prove compromise.

Raise your concern when other evidence appears at the same time:

• an unknown sign-in device
• an impossible sign-in time
• unauthorized password or contact changes
• unknown recipients or transactions
• sessions jumping across countries in a short period
• signs of malware or remote control

Do not interpret a GeoIP city error as proof that an account was stolen.

Common misconceptions

“The GeoIP coordinates are my actual coordinates”

False. They are usually a representative point in an estimated area, not device GPS.

“A nearby city means I am using a VPN”

False. Centralized ISP egress and mobile-network design are more common explanations.

“A WHOIS address is the server or user's location”

False. WHOIS and RDAP describe resources and organizations, not a live device.

“Restarting the router always fixes it”

No. A new IP may come from the same range, with the entire range mapped to one city.

“All IP lookup sites should agree exactly”

No. Their observations, algorithms, and refresh schedules differ.

“A foreign IP proves someone is remotely controlling my device”

No. A roaming eSIM, corporate gateway, or VPN can all create cross-border egress. Use device and account activity as additional evidence.

Frequently asked questions

Why does my IP show another city?

The most common reason is centralized egress used by an ISP, mobile carrier, or company network. A GeoIP database may locate the gateway, registered range, or nearby metro area rather than the device.

Why does my IP show another country?

Common causes include international roaming or eSIM egress, a VPN or proxy, a corporate security gateway, transferred address space, and an outdated GeoIP database. A country mismatch generally deserves more investigation than a city mismatch.

Will restarting my router fix my IP location?

Not necessarily. A reconnect can assign another dynamic IP in the same range. If the issue is the provider's database or the ISP's architecture, a restart does not solve the underlying problem.

Does a wrong IP location mean my account or device was compromised?

Usually not. Routing and database issues are more common. Investigate when unknown devices, unusual sign-ins, changed credentials, or unauthorized transactions also appear.

Why do IP lookup sites show different cities?

They may use different providers, sources, refresh schedules, algorithms, and confidence thresholds. Prioritize the country, ASN, ISP, and accuracy radius.

How can an incorrect IP location be corrected?

Contact the ISP or submit a correction to the GeoIP provider showing the error. A network operator can publish a standardized Geofeed. Corrections require time to propagate and will not reach every site simultaneously.

Summary

IP geolocation is not a precise coordinate read from device GPS. Providers estimate it from prefixes, ASNs, ISPs, Geofeeds, route measurements, and historical data.

Wrong cities commonly result from centralized ISP egress, CGNAT, dynamic assignment, or coarse data. Wrong countries more often involve roaming eSIMs, VPNs, corporate egress, transferred address space, or delayed updates.

Read the country, ASN, ISP or organization, network type, accuracy radius, update time, IPv4 and IPv6, and WebRTC, DNS, and time-zone consistency together. IP location is useful network context, not GPS, proof of residence, or a final identity decision.

Primary sources

1. MaxMind, Geolocation Accuracy
2. MaxMind, IP Geolocation Data
3. MaxMind, Geolocation Coverage
4. MaxMind, Correct GeoIP Data
5. RFC Editor, RFC 8805: A Format for Self-Published IP Geolocation Feeds
6. RFC Editor, RFC 9632: Finding and Using Geofeed Data
7. ARIN, IP Geolocation: The Good, The Bad, & The Frustrating
8. ARIN, Top 5 Questions Asked at an ARIN Help Desk
9. Cloudflare, Radar Glossary: IP Address Geolocation
10. Cloudflare, Dedicated Egress IPs: IP Geolocation

This article provides general education about networking and security. GeoIP data changes with time, source, and network path. It does not identify a device's precise physical location or represent an official decision by any third-party platform.