Eco Domain (.eco) avatar

The carbon footprint of domain registries

In this report we look at the carbon footprint of different domain registries and explore intensity measures for comparing them.

Cyclists biking in a bike lane

Every top-level domain (TLD) is operated and managed by a domain registry. Each domain registry maintains the database of all domain registrations for the TLDs it operates and is responsible for handling requests to look up the DNS records for those domain names. Doing this in a performant and resilient manner requires running technical infrastructure in data centres around the world. The manufacturing, provisioning and operation of this digital equipment takes energy and currently that energy is generated by burning fossil fuels. So running a domain name registry, just like most other human activity, contributes to climate change (see our article on the Climate Impact of a Domain Name for more details).

As the domain registry for the .eco TLD, we take our role as members of the global environmental community seriously. We want to ensure that the environmental impact of our activities are minimized. This is why we have been carbon neutral since 2022 and have annually published a greenhouse gas report giving transparency into our carbon footprint and how we are working to reduce it.

We are not the only members of the domain industry involved in this endeavour. A number of other domain registries have also made a strong commitment to climate accountability and transparency by publishing an annual climate impact report of their operations. Some, like Afnic (.fr), Nominet (.uk) and DNS Belgium (.be) have been doing this for longer - since 2013 for Afnic! We’re excited that .eco is part of a growing group of organizations in the domain industry that have made commitments to climate action.

The Data

We have pulled together the latest carbon footprint report data from the registries that we know of that are taking climate action. This is both to recognize their work and to provide a simple basis for comparison. The intention in doing so is not to shame or criticize any of the registries listed here. It is more the opposite: these organizations should be commended for their work and transparency. By having the data collated in one place, it provides some overall insight into the footprint of operating domain registries.

Also, please note that this analysis just deals with the overall carbon footprint of these organizations (including Scope 1, 2 and 3) measured in metric tonnes of carbon dioxide equivalents (t CO2e). It is important to underscore that even though these organizations are in a common industry and provide similar services, they differ in terms of the specific services they manage, how and where they operate and what they have chosen to include in their carbon footprint.

Registry Overall footprint
(t CO2e)
Reporting year Intensity per employee
(t CO2e)
Intensity per domain
(g CO2e)
.be (DNS Belgium) 124.7 2023 3.4 72
.ca (CIRA) 1006.7 2024 7.7 295
.eco (Big Room) 3.8 2024 1.9 400
.fr (Afnic) 690 2023 3.7 147
.uk (Nominet) 1075 2024 3.8 103

As domain registries differ considerably in terms of their size and operations, we have included some measures of carbon emissions intensity.

Carbon emissions intensity divides the total footprint by another measure, such as number of employees, total revenue or number of domain names under management, to normalize the total value and provide a better basis of comparison.

There are pros and cons associated with different intensity metrics and we will explore them further below.

Carbon intensity per employee

Dividing the total footprint by the number of people employed by the organization is useful because it offers a broad basis of comparison both within and across other companies beyond the domain name industry. Employee headcount is readily available, especially for larger or public organizations. This allows for a comparison between a small company like ours and much larger registries like Nominet or CIRA.

The per employee intensities are quite consistent across all of the organizations listed (within an order of magnitude). For an organization like CIRA, operating in a large country with a cold climate like Canada, it’s perhaps unsurprising that it would have a higher per employee carbon intensity due to more energy required for heating and transportation.

This intensity metric can also be compared with the typical per capita carbon intensity of a citizen of the country where these registries are located.

Registry Intensity per employee
(t CO2e)
Intensity per capita (2023)
(t CO2e)
.be (DNS Belgium) 3.4 7.1
.ca (CIRA) 7.7 14.0
.eco (Big Room) 1.9 14.0
.fr (Afnic) 3.7 4.1
.uk (Nominet) 3.8 4.4

While this intensity measure has the advantage of being broadly comparable, it focuses on employee headcount which is not something that should be cut to drive carbon efficiency. Employment is a good thing and we don’t want organizations reducing headcount to improve their carbon emissions intensity. It is also not correlated with a unit of production, which is a better and more targeted measure of efficiency.

Carbon intensity per domain name

Domain names under management (DUMs) is a common measurement used in the domain name industry. It represents the total number of domain names that are in a registered state at a given point in time under a top-level domain. Most registries publish their DUM counts, so this metric is widely publicly available.

Using this measure as the basis for calculating carbon intensity has the benefit of giving a domain name holder an idea of the climate impact of their domain name. For example, someone with a .fr domain could consider that the carbon cost of their domain name in 2023 was approximately 147 gCO2e. They could potentially include this count in their own carbon footprint.

Registry Domains under management Intensity per domain
(g CO2e)
.be (DNS Belgium) 1,731,806 72
.ca (CIRA) 3,414,858 295
.eco (Big Room) 9,500 400
.fr (Afnic) 4,693,878 147
.uk (Nominet) 10,400,000 103

There are definitely economies of scale at work in this measure. There is some baseline overhead associated with operating a registry. Having a large DUM count means that the carbon cost can be allocated across a larger number of domains, lowering the per domain intensity. In other words, TLDs with a large number of registered domains will generally have a lower carbon emissions intensity per domain. Though that’s not exclusively true as .be fares very well according to this metric.

Towards a better carbon intensity metric

One thing that we don’t like about using DUM count as the basis for calculating carbon intensity is that a domain name is largely inert. After registration, it’s just a row in the registry database, consuming very little energy by itself. Many domain names are registered and never used, and hence are responsible for next to no emissions.

What consumes most of the energy associated with the technical infrastructure of operating a domain registry is serving DNS requests. Domain registries handle billions of authoritative DNS requests per year in data centres running all over the world. Performing this function is an essential part of operating a TLD and so it should be accounted for when considering carbon intensity.

We have an announcement that we will be making soon regarding carbon reporting for members of the .eco community, so watch this space.

Wrapping up

Disclosing the carbon footprint of an organization's operations is still relatively new. It is not mandated, although legislation requiring publishing this information is coming soon to some jurisdictions. Organizations choosing to publish this information are leaders who are making a real commitment to sustainability and transparency, and they should be commended.

Now that we have a few registries in the domain name industry providing this data, we can start to get a better sense of the climate impact of domain names and think about how to best share this information with domain holders.

As additional registries disclose their own carbon footprint, we will add them to this report expanding the analysis, so continue to check for updates. Let us know if you have any comments or suggestions for additional ways to analyze this data.