# Who owns the networks behind the websites of German municipalities?

> Review of all 10,747 official municipalities, 10,438 analysed

- Publisher: Nexthosting
- Authorship: Erik Friedl (ORCID: https://orcid.org/0009-0008-2784-4314), Nexthosting
- DOI: https://doi.org/10.5281/zenodo.24430484
- Type: Own measurement
- Category: Public sector
- Version: 1.0.0, As of Nov 21, 2026
- Period: Nov 21, 2026
- Sample: 10,438 of 10,747 Municipalities evaluated
- License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/deed.en)
- Original page: https://nexthosting.net/en/studies/gemeinden-website-hosting
- PDF: https://nexthosting.net/en/studies/gemeinden-website-hosting/studie.pdf
- Translation of the German original. The original is authoritative. https://nexthosting.net/studien/gemeinden-website-hosting

## Summary

For at least 1,041 of 10,438 analysed municipalities, the website is served by hosting, cloud or CDN services controlled from outside the EU.

## Research question

Which networks serve the official websites of German municipalities, and for what share of municipalities is the service demonstrably controlled by companies or investors outside the EU?

## Key results

- **Confirmed municipalities** (10,440 of 10,747): For 10,440 of the 10,747 official municipalities, the website was confirmed, by its content or through an official source, as the website of the municipality or its administration.
- **Network registered in Germany** (93.9 %): For 93.9 % of the analysed municipalities, the website is in a network with a German registry entry. The entry names the network holder, but neither its parent group or investor nor the server location.
- **Service under control from outside the EU** (at least 1,041 of 10,438): For at least 1,041 of 10,438 analysed municipalities (around 10.0 %), the website is served by a hosting, cloud or CDN service that is demonstrably controlled by a company or investor outside the EU.
- **Including transit and city networks** (at least 1,454 of 10,438): If transit, city network and data centre operators are added, the website of at least 1,454 of 10,438 analysed municipalities (around 13.9 %) is in a network with documented control from outside the EU. The website may be operated there by a different service provider.
- **Service under control from the USA** (at least 819 of 10,438): For at least 819 of 10,438 analysed municipalities (around 7.8 %), the website is served by a hosting, cloud or CDN service controlled by a company or investor from the USA.
- **Financial investors** (at least 237 of 10,438): For at least 237 of 10,438 analysed municipalities (around 2.3 %), control of the hosting, cloud or CDN service lies with a financial investor outside the EU.
- **Nameservers** (at least 1,926 of 10,438): For at least 1,926 of 10,438 municipalities with a confirmed website and a determinable nameserver network (around 18.5 %), at least one nameserver of the website domain is in a network with documented control from outside the EU.
- **IPv6 address in DNS** (41.4 %): For the website of 41.4 % of the confirmed municipalities, an IPv6 address is entered in DNS.

The study examines, for all 10,747 official municipalities in Germany, which networks serve their website. 10,438 municipalities with a confirmed website and a determinable network were analysed.

93.9 % of these municipalities use a website in a network with a German registry entry. The entry names the network holder, but neither its parent group or investor nor the server location. For 703 municipalities, the website is in the network of a German-registered hoster that is demonstrably controlled from outside the EU.

If services with a foreign registry entry are added, the figure is at least 1,041 of 10,438 analysed municipalities (around 10.0 %), of which at least 819 are under control from the USA. Including transit, city network and data centre operators, the figure rises to at least 1,454 municipalities (around 13.9 %). Only what is backed by ownership sources that were read and cross-checked by a second AI agent was counted. The actual shares may therefore be higher.

The study makes no statement about server locations, data outflows or legal violations. For email, researchers from the University of Fribourg, the Lucerne University of Applied Sciences and Arts, TH Ingolstadt and TH Nürnberg have already examined the municipalities in Germany, Austria and Switzerland. This study adds the websites and their nameservers.

## Method

- **Population**: All politically independent municipalities according to the municipality directory of the Federal Statistical Office, as of 30.09.2026. Unincorporated areas are not counted.
- **Website per municipality**: The addresses come from Wikidata and were matched using the official municipality key. 33 missing addresses were researched individually. If a municipality has no website of its own, the website of its office or its collective municipality is used.
- **Content check**: The home pages were retrieved once, redirects were followed and robots.txt was respected. A site is confirmed automatically if it contains the name of the municipality and typical terms of an administration. Unclear cases were additionally checked using the home page and the legal notice. All 794 addresses still open after that were checked individually, and the 311 that remained unclear were checked in a second round using municipality lists of districts and state portals. Wrong addresses were replaced by the documented official address and measured again. Content of sites blocked by robots.txt was not used as evidence in the analysis. For 67 confirmed municipalities, the stored address was blocked: 64 were confirmed through an official third-party source, 3 through a home page that was not blocked. Parked and default pages never count as a confirmed website.
- **Network measurement**: For the target address, CNAME, A, AAAA and NS records were queried via DNS. The first address was assigned, using public routing and registry data, to the network (Autonomous System) that announces its most specific address block.
- **Control of the networks**: For 306 networks, the holder and registered seat were recorded, and for 104 of them an ownership statement with at least a medium level of evidence is available. Of 47 networks with possible control from outside the EU, the 40 with at least a medium level of evidence were checked individually using mandatory disclosures, annual reports, legal notices and registry data, and subjected to an attempt at refutation. Networks of the same owner group were assessed uniformly. One large transit network was removed because the registered holder of the address block used there has no longer belonged to the group since the sale of the European business. A network with only one municipality and its own country of control is not counted, to protect that municipality. 31 networks are counted in the results. 10 further networks were only added through the individual check or the renewed network assignment, were not checked for their owners and are therefore not counted. The ownership information of 13 networks with reported sales transactions or older sources was checked again on 22.11.2026 and remained unchanged.
- **Control check**: A fixed random sample of 150 automatically confirmed municipality sites was checked without knowledge of the automatic result: 146 were official sites, 0 were wrongly assigned and 4 could not be decided. At 95 % confidence, an error rate of up to around 2.4 % remains possible for the automatically confirmed sites, and up to around 6.7 % if the undecidable cases are counted as errors. The individually checked municipalities are not covered by this sample.
- **Use of AI**: Research, individual checks, source checks and the control check were carried out by AI agents. A second agent cross-checked the ownership assignment and the upgrades made in the second round. The result figures are calculated from the data by script.

### Measurement setup

- Measurement date: 21.11.2026
- Municipalities: 10,747
- Confirmed municipalities: 10,440
- Input addresses: 8,202
- Distinct target addresses with network: 7,846
- Analysed municipalities (web): 10,438
- Analysed municipalities (nameserver): 10,438
- Networks with recorded holder: 306
- Networks with documented control from outside the EU: 31
- DNS resolvers: 1.1.1.1 and 8.8.8.8

## Sample

Full survey of all politically independent municipalities according to the official municipality directory, as of 30.09.2026. Included in the analysis are municipalities whose website was confirmed, by its content or through an official source, as the website of the municipality or its administration, and whose network could be determined.

## Limits of what the study can show

- **Minimum values**: A network only counts as controlled from outside the EU if the ownership source was read and the assignment survived an attempt at refutation. Three networks of one owner group did not pass this check because the country of the manager was disputed (United Kingdom or Guernsey), and they are counted through the uniform group assessment. For 202 of the 306 networks, with 55.6 % of the analysed municipalities, the owner has not been established. They are not counted. The shares are therefore lower bounds for the analysed municipalities. Relative to all 10,747 municipalities, the 1,041 municipalities with such a service correspond to around 9.7 %. The study deliberately does not report a share of German or European owners.
- **No server location**: What was measured is which network an address belongs to, not where the server is located. A network under foreign control can operate servers in Germany. No statements about data outflows or legal violations can be derived from the data.
- **Network and website operator**: Of the 1,454 municipalities with documented control from outside the EU, 413 are in transit, city network or data centre networks. They are therefore shown separately from the hosting, cloud and CDN services. For transit networks, the address block may belong to a different company, including a German one. What is counted is the network that announces the address in routing. For upstream CDNs, only this network is visible, not the origin server.
- **Excluded municipalities**: 307 municipalities are not included in the shares: for 150, the known address does not lead to an official website, and for 157, the assignment of the address was not unambiguous. For a further 2 confirmed municipalities, no network could be determined.
- **Shared websites**: Many small municipalities share the website of their administration. The 10,438 analysed municipalities lead to 7,846 distinct target addresses. Each municipality counts once, so a shared website counts several times. Calculated per target address, the website is served by such a service for at least 967 of 7,846 addresses (around 12.3 %).
- **Differences between states**: Differences between states and size classes are often due to individual shared IT service providers and are not an assessment of the states or municipalities.
- **Snapshot**: The measurement comes from one day and one measurement point. Ownership and hosting can change at any time. The date of the respective ownership source is stated in the data, and for some networks the most recent explicit source dates back several years.
- **Definition of control**: Control here means: the network belongs to the company, to a documented subsidiary of the group, or to an investor with documented controlling influence. For financial investors, the country of the manager applies, not the seat of individual funds. The group outside the EU comprises countries with different standing towards the EU under data protection law. The study does not assess this.

## Data and evidence

- [daten.csv](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/daten.csv): Measured values per municipality (anonymised, CSV) (SHA-256 `7abbfbb53672e0f36be37f5503234f13231385306ea9e7e9e31f8b137ab43374`)
- [laender.csv](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/laender.csv): Results by state (CSV) (SHA-256 `54ab7e34f99be290fbdf3d2ff1d808c82bd388d223b6f74b4732036b1875dbf1`)
- [groessen.csv](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/groessen.csv): Results by population (CSV) (SHA-256 `6d7e6f8de1cf30d78c4e6995dfcebe0f507d558ada16a8c13693d2e6c49f9617`)
- [netze.csv](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/netze.csv): Networks with ownership evidence (CSV) (SHA-256 `7cb2aac93a23300882306c30ab105b1b617a71afac0879a3a50016f009da8361`)
- [stichprobe.csv](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/stichprobe.csv): Result of the control check (CSV) (SHA-256 `13f400f0a2f709c1bc6feb1d4f6cd94b053d2dc3b7d31f0b371b09b26fd898dc`)
- [01-wikidata.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/01-wikidata.mjs): Script 01-wikidata.mjs (SHA-256 `141679f4f2a695acea47c3a0a08d644442605e23bbbb0369db688bdba4e8b2bd`)
- [02-liste.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/02-liste.mjs): Script 02-liste.mjs (SHA-256 `374be3a36cea6b205a02826483f21237ec174671e10575c44df3a16809281b1f`)
- [03-messen.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/03-messen.mjs): Script 03-messen.mjs (SHA-256 `cc15d2de15ce9bebd0afaf83e07b3551af984eb9da61d35ff6c61da68f9baca3`)
- [04-destatis.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/04-destatis.mjs): Script 04-destatis.mjs (SHA-256 `8dafc11e154553ab36319164d20fb12e3309e82f22b80a1a3b5058c07a3ba54b`)
- [05-luecken.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/05-luecken.mjs): Script 05-luecken.mjs (SHA-256 `345eb2f11ef1a66f225e6e269988fe0508ac87885fcf07fd587679bbdb81b3fc`)
- [06-startseiten.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/06-startseiten.mjs): Script 06-startseiten.mjs (SHA-256 `f827bc35a0364a42bc30ebfca4e7fcaf75e2bd174228146a5ee9c0b904056d92`)
- [07-universum.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/07-universum.mjs): Script 07-universum.mjs (SHA-256 `d6ef47d83ab8c2dfd32b7d5435f51225faa7ed3efe16ba8dc0fd4bedb4b2c78e`)
- [08-nachpruefung.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/08-nachpruefung.mjs): Script 08-nachpruefung.mjs (SHA-256 `6f4af08be28e2f24f1e0bbee09a89c146de76e59342af927b86216c8f2ef8405`)
- [09-zusammenfuehren.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/09-zusammenfuehren.mjs): Script 09-zusammenfuehren.mjs (SHA-256 `55f7d133c6dad42cdb39e6331d8d239a219581494d88699f2e3f3cf44164dc94`)
- [10-rdap.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/10-rdap.mjs): Script 10-rdap.mjs (SHA-256 `733e61b164a97ce2c02ea25b23ea259c312fe05068dfdeeb92d658bbf4798bca`)
- [11-netzliste.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/11-netzliste.mjs): Script 11-netzliste.mjs (SHA-256 `b5bd39ec73b6b31d395f72126e5a7d27464799cf2d7467c58134ae84f2a7cb0f`)
- [12-nichteu-liste.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/12-nichteu-liste.mjs): Script 12-nichteu-liste.mjs (SHA-256 `6837ce147c98eef183f387fe408a0c4b446ff9348fc6e2dcd7d5bf99ded2b3df`)
- [13-stichprobe.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/13-stichprobe.mjs): Script 13-stichprobe.mjs (SHA-256 `acfaf328419338c14fc286fcb12c872ae1149eaf4bf68fb2d9a73f70aeb3e373`)
- [14-studie.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/14-studie.mjs): Script 14-studie.mjs (SHA-256 `64b065d96a292c3d2838e420d6c086e37c3d16c8e7e16b8be4e0a8def78701d8`)
- [15-einzelpruefung-liste.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/15-einzelpruefung-liste.mjs): Script 15-einzelpruefung-liste.mjs (SHA-256 `8f8ea7f9ed860e1fdf5c8417544fcd96268bb28baab56f2aff2ea303410b487e`)
- [16-einzelpruefung-runde2.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/16-einzelpruefung-runde2.mjs): Script 16-einzelpruefung-runde2.mjs (SHA-256 `3114be18f7ca002c729b092ee6be6615edc440cc9774cb7ec024efd7b76e3bc3`)
- [17-einzelpruefung-abschluss.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/17-einzelpruefung-abschluss.mjs): Script 17-einzelpruefung-abschluss.mjs (SHA-256 `0527838c7ecd210c3e55e8751dab7e05e8b8d83cafde9415f644018150034b3a`)
- [18-netze-nachmessen.mjs](https://nexthosting.net/studien/gemeinden-website-hosting/dateien/18-netze-nachmessen.mjs): Script 18-netze-nachmessen.mjs (SHA-256 `56031e05ec918aa8ec5c0d1528980da85060a35d6b3e61d2a88873ed7e7a1f94`)

Checksums: https://nexthosting.net/studien/gemeinden-website-hosting/pruefsummen.txt

The numbered scripts document the measurement and analysis. The ownership and individual checks were carried out by AI agents, and their results are summarised in netze.csv and stichprobe.csv. The individual verdicts for each municipality are not published, so that no municipality is exposed. For the 6 networks that are grouped together, the evidence is not named in order to protect the municipalities. A complete re-run from the published files is therefore not possible. The core results can be recalculated from daten.csv, and the ownership information can be checked using the evidence addresses. Tables by state and by size, as well as figures per target address, are based on the unpublished assignments.

### Fields of the data files

- `inhalt`: Result of the website check: bestaetigt, nicht_offiziell or nicht_entscheidbar
- `web_auswertbar`: Website confirmed and network determinable
- `netz_register`: Registry country of the network: DE, EU, nicht_EU or unbekannt
- `web_kontrolle`: dienst: hoster, cloud or CDN service with documented control from outside the EU; netz: transit, city network or data centre operator with documented control; kein_nachweis: not counted
- `web_region`: Region of the documented control: USA, UK/Kanalinseln, Sonstige or zusammengefasst
- `web_rolle`: Type of control: konzern, finanzinvestor or zusammengefasst
- `web_funktion`: Function of the network: hoster, cloud, cdn, netz_rechenzentrum, dns_domain or zusammengefasst for fewer than 5 municipalities
- `dns_auswertbar`: Website confirmed and at least one nameserver network determinable
- `dns_nicht_eu`: At least one nameserver in a network with documented control from outside the EU
- `dns_usa`: At least one nameserver in a network with documented control from the USA
- `ipv6`: The target address has an AAAA record in DNS
- `gruppe`: State or population class
- `gemeinden`: Number of official municipalities
- `bestaetigt`: Municipalities with a confirmed website
- `web_basis`: Analysed municipalities for website shares
- `dns_basis`: Analysed municipalities for the nameserver share
- `anteil_dienst`: Minimum share with a hosting, cloud or CDN service under control from outside the EU, in percent
- `anteil_dns`: Minimum share with a nameserver under control from outside the EU, in percent, relative to dns_basis
- `asn`: Number of the Autonomous System or a summary row for networks with fewer than 5 municipalities
- `region`: USA, UK/Channel Islands, Other or Mixed
- `kontroll_land`: Country of control as ISO code or Several. GB/GG: manager in the United Kingdom, fund structure in Guernsey, in part together with a Canadian investor
- `rolle`: Group, financial investor or Mixed
- `funktion`: Function of the network: hoster, cloud, CDN, transit, city network or data centre, DNS and domains or Mixed
- `gemeinden_web`: Confirmed municipalities whose website is in this network, values from 1 to 4 shown as under 5
- `gemeinden_dns`: Confirmed municipalities with a nameserver in this network, values from 1 to 4 shown as under 5
- `quellenstand`: Date of the most recent dated document on the ownership relationship, with a note for undated sources
- `belege`: Addresses of the ownership evidence that was read, separated by spaces
- `pruefurteil`: Verdict of the control check: offiziell_eigen, offiziell_verwaltung, nicht_offiziell or nicht_entscheidbar

## Sources

1. Statistisches Bundesamt (2026): Gemeindeverzeichnis, Alle politisch selbständigen Gemeinden, Stand 30.09.2026. https://www.destatis.de/DE/Themen/Laender-Regionen/Regionales/Gemeindeverzeichnis/Administrativ/Archiv/GVAuszugQ/AuszugGV3QAktuell.html (Retrieved on Nov 21, 2026)
2. Wikimedia Foundation (2026): Wikidata Query Service. https://query.wikidata.org/ (Retrieved on Nov 21, 2026)
3. Team Cymru (2026): IP to ASN Mapping. https://www.team-cymru.com/ip-asn-mapping (Retrieved on Nov 21, 2026)
4. RIPE NCC (2026): RIPE Database, Registration Data Access Protocol. https://apps.db.ripe.net/db-web-ui/query (Retrieved on Nov 21, 2026)
5. American Registry for Internet Numbers (2026): ARIN Registration Data Access Protocol. https://search.arin.net/rdap/ (Retrieved on Nov 21, 2026)
6. RIPE NCC (2026): RIPEstat. https://stat.ripe.net/ (Retrieved on Nov 21, 2026)
7. U.S. Securities and Exchange Commission (2026): EDGAR, Unternehmensmeldungen. https://www.sec.gov/edgar/search/ (Retrieved on Nov 21, 2026)
8. UK Companies House (2026): Companies House, Unternehmensregister. https://find-and-update.company-information.service.gov.uk/ (Retrieved on Nov 21, 2026)
9. Bischof, Huser, Ochsner, Petrlic, Portmann; Springer Nature (2026): Email Provider Dependencies and Email Security in Municipalities Across Germany, Austria, and Switzerland. In: Computer Security. ESORICS 2026 International Workshops (SecAssure 2026), Lecture Notes in Computer Science (angenommenes Manuskript). https://secassure2026.mxmap-project.org/ (Retrieved on Nov 21, 2026)
10. Technische Hochschule Nürnberg Georg Simon Ohm, idw (2026): Sicherheits-Check: Forschungsteam untersucht kommunale E-Mail-Server im DACH-Raum (Pressemitteilung vom 26.10.2026). https://idw-online.de/de/news879675 (Retrieved on Nov 21, 2026)

## Own interest

Nexthosting sells hosting in Germany itself, so it competes with companies in this analysis and uses data centres and networks of other companies. These include services of network operators that are listed here as controlled from outside the EU, among them a CDN. The method, analysis scripts, data and the ownership evidence for the listed networks are therefore disclosed.

## Changes and corrections

- 1.0.0 (Nov 22, 2026, Version): First version.

## How to cite this study

Friedl, E. & Nexthosting. (2026). Who owns the networks behind the websites of German municipalities? (Version 1.0.0) [Report]. Nexthosting. https://doi.org/10.5281/zenodo.24430484

```bibtex
@techreport{nexthosting2026gemeindenwebsitehosting,
  author     = {Erik Friedl and {Nexthosting}},
  title      = {Who owns the networks behind the websites of German municipalities?},
  institution= {Nexthosting},
  type       = {Study},
  year       = {2026},
  version    = {1.0.0},
  url        = {https://nexthosting.net/en/studies/gemeinden-website-hosting},
  urldate    = {2026-11-22},
  language   = {english},
  doi        = {10.5281/zenodo.24430484}
}
```
