Privacy & VPN

Residential IP Address vs Datacenter vs Mobile: What Sites See

What a residential IP address is, how it differs from datacenter and mobile IPs, how sites classify them from registry, ASN and list data, and how to check.

In this guide

A residential IP address is one a consumer ISP assigns to a home connection; a datacenter IP belongs to a hosting or cloud provider; a mobile IP is a carrier address shared by many phones behind carrier-grade NAT. Websites tell them apart from registry records, the announcing network's ASN, published hosting ranges and reputation lists, not from your device.

Test it now

Cloudflare edge
IP websites see216.73.216.115
LocationColumbus, United States
Internet providerAnthropic, PBC

Read at the Cloudflare edge from this request. The homepage adds your provider, network type and the VPN, DNS and WebRTC checks.

Show details

The IP types explained: residential, datacenter, mobile and ISP or static residential

Connection type is a label for the network an address belongs to. Nothing in the address itself says "home" or "server"; the label comes from who holds the block and how that network is known to use it.

Residential. An address an ISP hands to a subscriber line: cable, DSL, fiber or fixed wireless. Databases record it with values such as MaxMind's Cable/DSL connection type and residential user type. They are usually leased dynamically and shared by every device behind the home router.

Datacenter (hosting). An address held by a hosting or cloud provider and used by servers. Large providers publish their ranges, which makes this the easiest type to identify; MaxMind's is_hosting_provider field is "true if the IP address belongs to a hosting provider."

Mobile. An address on a cellular carrier's network, MaxMind's Cellular type. Carriers commonly place handsets behind carrier-grade NAT, so one public address is shared by many subscribers and the handset's own address sits in the 100.64.0.0/10 range that RFC 6598 reserves for that purpose.

ISP proxy or static residential. Marketing terms, not registry categories. They describe an address registered to an ISP, and therefore residential in registry data, but routed to a server rather than a subscriber line. The detection section explains why that matters.

TypeWho holds the blockWhat a lookup showsShared byReputation starting point
ResidentialA consumer ISP (an LIR in registry terms)ISP as registrant and announcing ASN; generated PTR name or none; often on the Spamhaus PBLOne household per address, changing over timeNormal user for browsing; refused for direct email
DatacenterA hosting or cloud providerHosting company as registrant; address inside a published cloud range; PTR often names the facilityOne server, or many tenantsExtra scrutiny; +20 in the CheckMyIP fraud-score formula
MobileA cellular carrierCarrier ASN; PTR usually missing; location shows a carrier hubMany subscribers at once behind CGNATShared: any block or challenge lands on everyone behind the gateway
BusinessA company's own networkThe company or its ISP as registrant; Corporate connection typeOne organizationUsually neutral

What is a residential IP address, exactly? Allocation and routing, not the device

A residential IP is defined by the chain of custody of the address. RFC 7020 describes the registry hierarchy as "rooted in the Internet Assigned Numbers Authority (IANA) address allocation function," which "serves a set of 'Regional Internet Registries' (RIRs)" (AFRINIC, APNIC, ARIN, LACNIC and RIPE NCC); the RIRs in turn serve Local Internet Registries, "typically ISPs, end users, or child LIRs."

Each registry records what happened to a block. ARIN's policy manual (version 2025.1, March 3, 2026) defines an allocation as "a block of IP addresses issued from ARIN directly to customers" that "may be further reassigned or reallocated," and a reassignment as addresses "sub-delegated to an organization by an upstream provider for the exclusive use of the recipient organization." The RIPE Database records its own version of this hierarchy as a status on each inetnum object: ALLOCATED PA, ASSIGNED PA, SUB-ALLOCATED PA and so on.

A residential address is one inside an ISP's allocation that is used for subscriber access, and two records say so: the registry names the ISP as the holder, and the global routing table shows the ISP's autonomous system announcing the prefix. The device at the end of the line is irrelevant. A server on a home fiber connection has a residential IP; a laptop connected through a cloud VPN has a datacenter one.

The two records can disagree. The IP lookup page notes that the registered organization and the announcing operator differ when a company rents addresses from a provider, and RIPE's SUB-ALLOCATED PA status exists to record that hand-off. That gap is where the "ISP proxy" category lives, and it is why classifiers look at more than one signal.

How websites tell them apart

No website sees the type directly. It assembles a verdict from several sources, roughly in this order.

  1. Registry records (RDAP and WHOIS). The holder's name, the block's type (ARIN shows values such as DIRECT ALLOCATION; RIPE shows the statuses above) and the registration date.
  2. The announcing ASN. The autonomous system that routes the address has a public owner, country and set of upstreams. PeeringDB's Network Type field adds self-described categories such as Cable/DSL/ISP, Content and Enterprise, with the caution that it "is self-declared and might be rarely updated."
  3. Published hosting, VPN and Tor ranges. Cloud providers publish machine-readable lists of their address space, the Tor Project publishes its exit list, VPN ranges are collected in public lists, open proxies are reported through DNS blocklists, and Apple publishes the egress ranges of iCloud Private Relay. Matching against these lists is the simplest test there is.
  4. Reverse DNS (PTR records). RFC 8501 observes that "many ISPs generate PTR records for all IP addresses used for customers" and that an ISP "has no knowledge of its residential users' hostnames," so it serves "a wildcard response or a dynamically generated response." A generated name is a residential tell; a name that encodes a facility is a hosting tell; no name is normal for mobile and many home lines.
  5. Policy lists. The Spamhaus Policy Blocklist is "a dataset containing end-user IP address ranges" that "should not be attempting to directly deliver unauthenticated SMTP email." ISPs submit their own ranges, and a listing "is NOT a result of any actions undertaken by the end users," so a PBL entry is close to an ISP's own declaration that a range is consumer space.
  6. Commercial usage-type databases. MaxMind's connection type database returns Cable/DSL, Cellular, Corporate or Satellite; its web services add a user_type (business, cellular, hosting, residential, school and others) and anonymizer flags including is_anonymous_vpn, is_public_proxy, is_tor_exit_node and is_residential_proxy. IPinfo describes active measurement: "We connect to services directly, analyze handshakes, scan protocols, and verify exit node behavior."
  7. Behavior. Address data is an input, not the verdict. Cloudflare's bot engines run "a number of heuristic checks to identify automated traffic," identify "headless browsers and other malicious fingerprints," and train machine learning on "large volumes of anonymized data from across Cloudflare's global network." Google's CAPTCHA appears when "a network you use, including VPN networks, seems to be sending automated traffic."

Check what your own IP looks like

Five CheckMyIP tools run these checks against any address, including yours, and each shows which source produced each value and when it was fetched.

  1. Open the IP lookup tool. It shows the registered organization, the announcing ASN, the network range with its registry and allocation date, the PTR name, and a connection-type label "derived from the ASN and the range lists," with a confidence level. The VPN, proxy, Tor and Private Relay flags each name their list and fetch date.
  2. Open the ASN lookup for the ASN shown: owner, registration country, issuing registry, announced prefixes, and peers and upstreams from RIPE RIS data. Some small ISPs have no ASN of their own and announce through their upstream, which makes their customers look like the upstream network.
  3. Open the WHOIS tool and enter the address. It queries the responsible registry over RDAP and shows the handle, type, range, registrant and abuse contact, with the warning that sub-assigned blocks appear as separate records and that the country is the registration country, not a location.
  4. Run the proxy check. It compares the address with the Tor exit list, VPN and datacenter range lists, published cloud ranges, the Private Relay egress list and proxy-flagging DNS blocklists, then adds browser-side checks for your own connection (time zone and languages against the address's location, the WebRTC address, the DNS resolver's network, proxy headers). The verdict is one or more of residential, mobile, business, data center, VPN, proxy, Tor or relay, with reasons.
  5. Run the IP fraud score. It sums fixed weights for the signals it finds (Tor exit +40, botnet command server +50, open-proxy evidence +35, known VPN range +30, hosting or cloud range +20, Private Relay +10, +15 per blocklist listing up to +45), capped at 100, and states that the number "is not a probability of fraud. It is a transparent sum of published list memberships."

The my ISP page is the quick version of steps 1 and 2 for your own connection. If the lookup labels your home line as datacenter or flags a VPN you are not using, the likely cause is a list that covers your ISP's range or a rented block whose registrant and announcing network disagree; the ASN and WHOIS pages show which.

Why it matters: trust scores, CAPTCHAs, streaming blocks and email deliverability

Risk scoring. A datacenter address starts with points against it on any list-based score, as the +20 hosting weight above shows, and a VPN range starts with more. The fraud-score page states the limit of the method: the score "is a property of the address and its history, not of the person using it," and "Sites that block on score alone block legitimate users."

CAPTCHAs. Google's "Unusual traffic from your computer network" message is tied to shared networks: "You could get this message if others that use the same Virtual Private Network (VPN) you use are sending automated searches," and some VPNs "make it impossible for Google to tell the difference between abusive and non-abusive traffic." An address shared by one household rarely trips this; an exit shared by thousands does.

Streaming. Netflix's help page for "You seem to be using a VPN or proxy" explains that such services "can make your device or network look like it's connecting from somewhere other than your physical location," and its troubleshooting has you compare your IP address and country with your home details.

Email. Here the hierarchy flips. A residential address is the wrong kind for sending mail directly: the Spamhaus PBL exists so receivers can refuse unauthenticated SMTP from end-user ranges, and RFC 8501 notes that "most email providers will not accept incoming connections on port 25 unless forward and reverse DNS entries match."

Residential vs datacenter IP: speed, stability, reputation, cost, the honest trade-offs

Speed. The type says nothing about the line itself; a datacenter address sits on a server in a facility built for throughput, a residential address on whatever the home connection is.

Stability. Residential addresses are leased, and when a lease moves the next subscriber inherits whatever reputation the address earned, which the IP ban guide covers. A hosting address is assigned by the provider from its own allocation; there is no lease to a changing household, so a range's history stays attached to the provider's customers.

Reputation. A residential address is the default state a website expects from a visitor, so it attracts the least scrutiny for browsing and the most refusal for sending mail. A datacenter address is expected from servers and crawlers, so services aimed at people treat it as a question mark. The fraud-score page puts it this way: "A VPN user on a shared exit inherits the exit's history; a cloud server inherits its range's reputation." That cuts both ways.

Cost and availability. A hosting provider can give a server a new address from its own allocation in seconds. A residential address exists only on a subscriber line and cannot be obtained from any registry; the only ways to use one are to be the subscriber or to route traffic through someone else's connection, the subject of the last two sections. The proxy server guide describes how each proxy type carries traffic.

Mobile IPs: shared by thousands behind CGNAT, and why blocking them is costly

A phone rarely holds a public address of its own. RFC 6598 reserved 100.64.0.0/10 to "number the interfaces that connect CGN devices to Customer Premises Equipment," and packets with those addresses "MUST NOT be forwarded across Service Provider boundaries." The public address a website sees belongs to the carrier's NAT gateway, which multiplexes many subscribers; the CGNAT guide explains how to tell whether a fixed line is behind one too.

Three consequences follow. Geolocation is coarse: the IP lookup page warns that "for mobile networks and VPN servers the city is often the provider's hub, not the user's," so a city mismatch means little on cellular. Reverse DNS is usually absent, which the lookup page calls normal for mobile. And any block placed on a mobile address lands on everyone behind that gateway, the shared-IP problem the IP ban guide documents, so a site that cares about false positives has to challenge mobile traffic rather than block it. "Mobile proxy vs residential proxy" refers to this difference: a mobile address is shared by a crowd by design, a residential address by one household.

Residential proxy detection: how it works and where it fails

A residential proxy is traffic that exits to the internet from a residential address belonging to someone other than the sender. Every registry signal says "home," so detection relies on three other things.

Measured datasets. Providers connect to proxy services and record which exit addresses they hand out. IPinfo calls residential proxy detection "a separate dataset with its own methodology" and acknowledges that residential proxies "bypass standard privacy detection entirely" without it; MaxMind's is_residential_proxy flag is "true if the IP address is on a suspected anonymizing network and belongs to a residential ISP." Such data is only as current as the last measurement.

Consistency checks. The browser reveals its time zone, languages, WebRTC candidates and DNS resolver. The proxy check treats a mismatch as evidence, with the caveat printed on the page that a time-zone mismatch "is not proof of anything" on its own.

Behavior. Request rate, page sequence, headless-browser fingerprints and account history are what the Cloudflare engines and Google's automated-traffic detection act on. An address that looks residential and behaves like a script is scored as a script.

Where it fails explains why an innocent home connection can be challenged. An address registered to an ISP but served from a hosting facility (the "ISP proxy" case) passes every registry test and is caught only by measurement or behavior. A VPN client on a home router makes the home line itself the exit, so there is nothing in the address to detect. And a compromised device is a genuine residential endpoint: the Justice Department's 911 S5 indictment describes malware that compromised "millions of residential Windows computers worldwide," associated with "more than 19 million unique IP addresses," and the Resident Evil measurement study found "many proxies run on likely compromised hosts including IoT devices," among them web cameras, DVRs and printers. A household whose address lands in such a pool inherits its reputation.

Sources

Questions

How do I look up my home IP address?

Open the CheckMyIP homepage or IP lookup tool from any device on your home network. It shows the public IPv4 and IPv6 addresses your connection presents, the ISP and ASN they belong to, and the connection-type label (residential, mobile, business or datacenter) with the sources behind it.

Is a residential proxy illegal?

Using a proxy is not illegal in itself in most places, but the legality depends on your jurisdiction, how the addresses were obtained, and what you do through them. The US Department of Justice prosecuted the 911 S5 service because its addresses came from malware-infected computers and customers used them for fraud.

What are the types of IP addresses?

Addresses are grouped several ways at once. By version, IPv4 or IPv6. By reach, public or private. By assignment, static or dynamic. By connection type, the subject of this guide, residential, business, datacenter or mobile, which describes the network that holds the address rather than the device using it.