Skip to content
InternetData

What do free datacenter IP lists cover?

We measured X4BNet, ipcat and a cloud range list against our Hosting IP and VPN IP databases: the hosting space each covers, and the VPN exits each catches.

Published 3 min read

On this page
  1. The lists side by side
  2. X4BNet's datacenter list is good
  3. A cloud range list isn't a datacenter list
  4. A frozen list falls behind
  5. Where X4BNet's misses are
  6. Choosing one
  7. How we measured
  8. Getting the databases

"Block datacenter IPs" is one of the first rules teams write against bots and VPN traffic, and the list behind it often comes from GitHub. Three free ones are in wide use: X4BNet/lists_vpn, with 903 stars and a rebuild every day; client9/ipcat, with 529, archived and unchanged since February 2023; and jhassine/server-ip-addresses, with 221, also rebuilt daily.

In June, a Hacker News reader asked of X4BNet's list: "Does this have most of the VPN IP addresses?" The one reply was "Not sure but I'd assume so."

We measured all three against our Hosting IP and VPN IP databases as built on 2026-10-09, IPv4 only: how much of the hosting space we classify each list covers, and how many of the 1,198,247 VPN exits we attribute at high confidence it catches.

The lists side by side

ListIPv4 addressesAlso in Hosting IPHosting IP's high it coversVPN exits it catches
X4BNet datacenter395.0M97.2%87.4%75.9%
X4BNet VPN3.2M98.2%0.7%33.9%
jhassine177.6M99.5%41.9%0.8%
ipcat96.0M92.1%17.6%12.5%
Hosting IP, high confidence420.0M100%100%72.0%
Hosting IP, all of it776.1M100%100%90.6%

Every range in Hosting IP carries a confidence: how sure we are that it's hosting, with medium less sure than high.

X4BNet's datacenter list is good

97.2% of what X4BNet's datacenter list holds is hosting in our data too, and it covers 87.4% of what Hosting IP holds at high. It catches more VPN exits than Hosting IP's high ranges do on their own, 75.9% against 72.0%. The 11.0M of its addresses outside Hosting IP aren't necessarily wrong: from our data alone, we can't tell its mistakes from our own misses.

Its VPN list is the narrow option. Nearly all of its 3.2M addresses are in the datacenter list too, and they catch 33.9% of VPN exits: about 8 addresses blocked for each exit caught, against 434 for the datacenter list.

A cloud range list isn't a datacenter list

jhassine's list holds the published ranges of AWS, Azure, Google Cloud and Cloudflare, as its README says. 99.5% of it is hosting in our data, and it covers 41.9% of what Hosting IP holds at high, so it's both accurate and large. It catches 0.8% of VPN exits. VPN companies rent their servers from other hosting companies, which is why blocking the big clouds stops almost no VPN traffic while it still blocks the virtual desktops companies run there.

A frozen list falls behind

ipcat's last change was in February 2023, and its open issues include "Lots of inaccuracies", "still maintained?" and "No vultr, sucuri, hostgator?". 92.1% of its 96.0M addresses are still hosting in our data, so what it lists has mostly stayed right. What it lacks is the problem: it covers 17.6% of what Hosting IP holds at high and catches 12.5% of VPN exits.

Where X4BNet's misses are

X4BNet's datacenter list misses 288,618 of the 1,198,247 exits:

  • 54,538 sit in Hosting IP's high ranges that it doesn't list yet. X4BNet builds the list from network numbers (ASNs) and adds one when someone opens an issue with evidence, as 23 "ASN Addition" requests have since August 1.
  • 141,196 sit in Hosting IP's medium ranges.
  • 92,884 sit outside every hosting range, ours and X4BNet's both. Only a VPN list catches these.

Choosing one

For bot filtering, where the question is whether a request comes from a server rather than a person, X4BNet's datacenter list is a reasonable free start. It's rebuilt daily, so update your copy daily too. A cloud range list answers a narrower question, and one that stopped updating in 2023 misses most of today's hosting space.

For VPN decisions, check for the VPN itself. A VPN list catches exits wherever they sit, outside hosting space included, where no datacenter list reaches.

How we measured

We downloaded each list on 2026-10-09 and merged its entries into non-overlapping ranges, IPv4 only: X4BNet's 44,422 datacenter and 11,037 VPN CIDRs, jhassine's 53,079 CIDRs and ipcat's 3,429 ranges. We compared them with Hosting IP and VPN IP as built the same day.

The VPN side is every IPv4 address VPN IP publishes with a named provider at high confidence, counting only those that send traffic out rather than only take connections in: 1,198,247 exits. Hosting IP is a curated catalog of hosting and cloud providers plus our own classification and scanning of datacenter networks.

Getting the databases

Hosting IP and VPN IP are rebuilt daily as gzipped CSV and MMDB downloads. Their docs list every column: Hosting IP, VPN IP. To license them, see pricing or talk to us.

Written by

  • Uman Shahzad

    Founder & CEO

    IP intelligence data expert for over 8 years, former lead engineer at IPinfo, Uman has pioneered many systems and practices used by IP data companies to operate at scale, provide high-quality data and offer the best developer experiences.

Get new posts by email

Or follow along in a feed reader.

RSS feed

Put this data to work.

Everything these posts report on comes from the databases we license, and you can test them first.

00 · Evaluate

Test the real build first.

An evaluation license puts a full current build in your hands, with no sampling, so you can measure it against your own traffic before anyone talks about terms.

01 · License

Then license what you use.

Standard, for detection and decisions inside your own product and operations.
Redistribution, when the data ships onward inside what you sell, priced by scope rather than by seat.
Annual terms, a named contact, and monthly, quarterly or yearly payment.