The “whole systems” energy sustainability of digitalization: humanizing the community risks and benefits of Nordic datacenter development

Digital platforms and the online services that they provide have become an indispensable and ubiquitous part of modern lifestyles, mediating our jobs, hobbies, patterns of consumption and forms of communication. However, no one is steering this development, or closely looking at the impacts that it may have on remote communities in the Arctic and Nordic region, a hotspot for datacenter development. Moreover, unlike other areas of energy consumption or technology adoption prone to rich, qualitative assessments, such work on datacenters involving local stakeholders and environmental concerns is less common, particularly at a larger scale. In this study, based on novel mixed methods — including corporate data, expert interviews, focus groups, and extensive site vis- its — across three countries, we offer a geographically and technologically bounded assessment looking at the sustainability impacts of datacenters on local communities. We ask: What impacts are occurring as part of datacenter development or planning proposals in Greenland, Iceland, and Norway? What is the actual and anticipated scale of those impacts on local Arctic communities? Finally, what impacts to datacenter development occur at the “ whole systems ” level? We examine not only impacts onsite at existing or proposed datacenters, but an entire range of consequences including the manufacturing of equipment, the laying of data cables, the con- struction of buildings, and issues of the dark web, cryptocurrency mining, hacking, spying, waste and decommissioning. Moreover, we humanize risks and benefits not only across scales, but also categorical types, including local impacts such as boom and bust cycles, the displacement of indigenous groups for land – particularly for power supply - and impacts on employment, especially after datacenters may close.


Introduction
Online services and their related infrastructure (such as digital devices and internet data centers) currently account for roughly 10% of global electricity demand, and approximately 3% of global greenhouse gas emissionsmaking them equivalent to the global emissions from the much more visible airline industry. In the UK and globally, there have been concerns about how electricity supply can sustainably support the growth of data centers and data traffic [1]. They continue to rapidly expand, with growth in internet traffic of about 20% annually. Even conservative models predict that online services and devices will rise to 20% of global electricity by 2030. Andrae [51,52] projects internet infrastructure as constituting a significant bulk of global electricity consumption by 2030, with utlization ranging from lows of 8% to more than 50% depending on method of evaluation and breakthrough technology. The global Covid-19 pandemic has further necessitated the need for more remote working (leading to more need for data capability)balancing growth of datacenters (and attendant issues) with supporting daily life and economies [2].
Datacenters have impacts beyond power consumption. As of 2019, the global footprint of datacenters was estimated at 63.4 million squarefeet, with another 4.3 million square-feet under construction [3]. Across Europe, JLL [4] reports that planning applications for datacenters in 2020 far exceeded those made in 2019, with expectations for increased growth in 2021 over 2020. According to these reports, the expected rollout of 5G, demand for edge computing, quicker downloads, greater efficiency, more devices being connected to the internet, the popularity of online gaming, and a rise in streaming services (among others) were creating exponential growth in both data volumes and processing [3,4].
Yet, no one is steering this development, nor always adequately attending to its impacts on global sustainability [5,6]. Reliable information on internet related energy use, carbon emissions, and the energy profiles of internet data centers is fragmented and often inadequate [7]. Most national regulations and global governance regimes focus on growing information infrastructure, rather than curtailing it or making it more sustainable. All the while, much of industrialized society is in danger of becoming "addicted" to the internet and smart phones [8,9]. Moreover, Hargreaves and Wilson [10] note that many of the energy savings that are made through these devices are instead counteracted by the more, or new, energy-intensive forms of demand that they enablean unintended consequence. When one undertakes a whole systems lifecycle assessment, the networks and datacenters supporting smart phones and computers consume more energy than the devices supported [53]. Hischier et al. [11] estimate that data networks account for as much as 90% of the total energy consumption of tablets and smartphones. As Corcoran and Andrae, note: "there is a strong tendency to push electricity consumption onto the network and data center infrastructure where energy costs are less transparent to consumers" ( [12]: 1).
Even though energy efficiency improvements have been applied to information networks, and proposals and pathways for green data centers exist [13], the IEA have indicated that: "energy use over the long run will continue to be a battle between data demand growth versus the continuation of efficiency improvements" ([14]: 2). As Murugesan [15] warned: "computers and other IT infrastructure consume significant amounts of electricity, placing a heavy burden on our electric grids and contributing to greenhouse gas emissions." But what value does such expansion of data centers, and digitalization, bring to local communities and households? The latter is an increasingly important question, the burgeoning literature of which we seek to add to. Authors such as Von Bargen and Fish [16] have addressed the multi-scale politics of data ownership. Fish [17] offers an evocative "experimental documentary" of Iceland's oceanic internet connection. Starosielski [18] addresses the technological, cultural, political and geographical aspects of undersea networks for another location, the South Pacific.
In this study, based on a wide range of mixed methods-including corporate data, expert interviews, focus groups, and site visits-across three countries, we offer a qualitative, geographically and technologically bounded assessment looking at the impacts of data centers on local communities in the next big energy frontier, the Arctic [19]. We ask: What impacts are occurring or being anticipated as part of datacenter development or planning proposals in Greenland, Iceland, and Norway? What impacts to datacenter development occur at the "whole systems" or multi-scalar level well beyond the Nordic region?
In exploring these questions, we aim to make three contributions. First, our "whole systems" approach enables us to examine not only impacts onsite at existing or proposed datacenters, but their entire suite of lifecycle impacts, including mineral extraction and waste. Secondly, our mixed methods research design enables us to capture and humanize risks and benefits across scales and types, including local impacts such as boom-and-bust cycles, the exclusion of indigenous groups, or impacts on employment, especially after datacenters may close. Thirdly, our focus on the Arctic brings into focus various salient geopolitical issues including autonomy and independence for Greenland, regional plans for Nordic industrial strategy, and even proposals to invest in infrastructure to counter major shifts in global power such as a current concentration of data infrastructure among China and the United States. Connected to this theme is an investigation of how blockchain technology and the growth of cryptocurrencies can expand in contexts of crisis in peripheral societies, as they have in rural areas of the United States or Venezuela [20][21][22].

Mixed methods research design
Unlike other areas of energy consumption or technology adoption prone to rich, qualitative assessments (such as electric vehicles, household solar panels or heat pumps), there is less close, qualitative work on datacenters involving local stakeholders and environmental concerns, at least at a larger scale. To improve both validity and triangulation, and to address the fact that limited qualitative data was available on data center development in the Arctic to begin with, we employed a mixed methods research design centered on original data collection. Our four primary sources of data were corporate benchmarking information, semi-structured research interviews, focus groups, and site visits. A supplemental method was documentary photography. More details on our methods are offered in our Supplementary Online Materials (see Appendix A).
Firstly, we relied on corporate benchmarking data to identify the number of datacenters (and other digital infrastructure such as colocation centers) in the Arctic. This resulted in a list of 31 data centers in Norway (as of late 2021), 7 in Iceland, and 1 in Greenland (see Table 1). We utilized this list to identify particular communities or locations experiencing datacenter development, and also to identify the names of key firms that we could approach for our second method of data collection, the interviews. We also utilized these sources to publish a second study that benchmarks datacenter performance across key quantitative metrics (see [13]).
In terms of data collection via interviews, we approached, largely by email, some 100 potential organizations and individuals considered to have a stake in datacenter development in Iceland and Norway. These include some of the companies in Table 1, but also ministries, state agencies, datacenter developers, municipal authority planning and economic departments, environmental NGOs and others (see Table 2). The conversion rate through to actual interviews was around 20%. Although we have captured a variety of views, we do not claim wide  Thor ICE-01  Iceland  2  atNorth ICE02  Iceland  3 Etix Fitjar #1 Iceland 4 Opin Kerfi Korputorg Iceland 5 Verne Global Iceland 6 Vodafone Reykjavik Iceland 7 Etix Blönduós #1 Iceland 1 Tele Greenland Nuuk Greenland representativeness from the interviews alone. The interviews were conducted in English from March to May 2021 over Zoom video-call, made necessary by the Covid-19 pandemic, followed by selected inperson interviews in June and July 2021. We asked respondents the following, among other questions: 1. What are the positive impacts, if any, of datacenters on local communities in Greenland, Iceland, and Norway? 2. What are the negative impacts, if any, of datacenters on local communities in Greenland, Iceland, and Norway? 3. Who may be impacted by datacenters in Europe or even across the whole sociotechnical system?
All interviews were recorded and transcribed for anonymity, with interviewees giving informed consent. Coding was assisted by Nvivo, with high level codes informed by the interview question themes and sub-codes added as more specific themes became evident [23]. The quotations cited here have been given light editing to aid comprehension, clarity, or readability.
In terms of focus groups, we used both in-person and online workshops. Two online focus groups of eight people each were held in Iceland and Norway at the end of June 2021 (i.e. four groups and 32 people in total). All participants were recruited by market research firms (Norstat in Norway, Maskina in Iceland). Facilitation was by Maskina in Iceland and the researchers in Norway; the online platform was Brainstork. Participant characteristics were broadly matched to country demographics in terms of age, gender and socio-economic class, with the proviso that selection filtered for the ability to speak and understand fluently in English and for those who owned a computer with a camera. The choice of online rather than physical focus groups was partly determined by the social distancing requirements of Covid-19; however online groups are also more amenable to achieving a rural/urban split, given recruitment patterns and the particular geographic distribution of citizens in these countries. Two in-person focus groups were conducted in Greenland, one urban one in the capital city of Nuuk, and one rural one in the fishing village of Sisimiut, also with a total of 16 participants (to match the size of the online focus groups). Focus groups per se can also compare well with individual interviews in terms of eliciting a wide range of comments on a topic [24]. Similar to the interviews, all focus groups were transcribed and allocated to pre-determined themes. Those themes reflect a wider project theme of "energy justice", taking into account the life cycles and spatial impacts of the processes involved [25].
Fourthly, in terms of site visits, members of the research team visited Akureyri and Reykjavík, Iceland; Oslo, Bergin, Stavanger and Sogndal in Norway; and Kangerlussuaq, Nuuk, and Sisimiut, Greenland. Again, a mix of urban and rural locations was chosen to maximize diversity. The site visits were conducted to both offer context and background, and enable some site visits and in-person interviews. These site visits enabled the "unstructured observation" of datacenter or community activity and enabled an examination of events in real-time in the real-world [26]. Strengths to this approach include stronger validity, with more authentic actions captured by researchers, including spontaneous ones; minimal influence of respondent or researcher bias; and witnessing events in their total complexity and context.
Finally, and as is partly obvious by Fig. 1, we present in this paper a large number of relevant (and original) documentary photographs as a supplemental method, to both further enhance our analysis and also present visual arguments that many readers may find helpful, interesting, or illustrative.
Despite these strengths, our study does have some shortcomings. For our expert interviews, we utilized a critical stakeholder approach to select participants, meaning we wanted those not only "for" or "against" internet datacenter infrastructure, but also respondents from a mix of the private sector, civil society, government, and academia. Thus, our final sample has only a small number (N = 4) of respondents from within the actual industry of datacenter operators and firms. This number is counterbalanced by insights from other key stakeholders. Given the richness of the original data we were working with, we elected to execute a narrative structure to the paper, so that it tells a story, rather than a more inductive structure organized only around academic themes, or a deductive structure organized around a particular concept or theory. Additionally, we present our data below anonymously, to protect the identity of our respondents. In terms of site visits and research interviews, we had to collect data in Iceland and Norway during the peak of the Covid-19 pandemic, meaning our access was more circumscribed and many interviews and visits were done virtually over Zoom. However, for Greenland, we had the benefit of traveling during a window when their economy opened up and had fewer restrictions on travel, making our interviews and site visits richer. This creates a slight imbalance in our data in favour of Greenland. Finally, we took an ethnographic approach that did not correct or problematize responses, so we present the data unfiltered, even if our respondents may have had misperceptions on specific points. In simpler terms, we dispassionately collected our data and present our results without "taking sides," aiming for a more neutral and balanced positionality as researchers.

Case study selection and background: grappling with datacenter development in the Arctic
Although definitions vary, we conceptualize a "datacenter" as "a structure, or group of structures, dedicated to the centralized accommodation, interconnection and operation of IT and network telecommunications equipment providing data storage, processing and transport services, together with all the support facilities for power supply and environmental control with the necessary levels of resilience and security required to provide the desired service availability" [27]. The industry relies generally on three different business models for datacenters: • Enterprise: ownership of the facility, IT equipment and software systems is common. Typical example is a bank, university or hospital data center. • Co-location: ownership of the facility is separate from the one of IT equipment, software systems and their immediate accommodation. Thus, the owner of the data center rents the infrastructure to allocate IT equipment. • Hosting: ownership of the facility and the IT equipment is common, but the software systems are dedicated by others. Thus, the owner rents both the infrastructure and the IT equipment to host information, servers, etc. We see all three business models employed in our three Arctic countries.
From energy security (availability, accessibility, and affordability), geographical, climatic, and political perspectives, the Nordic region presents an attractive location for datacenters. Since all Nordic countries, especially Finland, Sweden, and Norway offer these incentives to a considerable degree, one can expect some intra-region competition among these countries in attracting favorable datacenter owners/operators. For instance, both Iceland and Norway claim that they are the only countries where DCs can operate with "100%" renewable power [28,29].
In 2018, a joint study by Nordic Council Ministers explored the C. Kangerlussuaq, Nuuk, and Sisimiut Greenland competitive advantages of the Nordics for hosting a major share of the global datacenter installation [30]. These competitive advantages include abundance of cost-efficient renewable energy sources; cold climate, low energy prices; political stability and policy boosts including tax cuts for datacenters, environmental stabilitylow risk of natural disasters; and securitylow risk of malicious or terrorist attacks. The study acknowledged more commitment from Nordic country governments to further sharpen these competitive advantages by improving international connectivity (connection latencies); improving workforce competency in ICT; harmonizing national policies, processes, and planning to attract datacenter owners including earmarking specific zones for datacenter deployment with requisite enhancements of multiple grid connections and accessibility to district heating networks. Currently, there are plans to run undersea fiber optic cables through the North Pole and Artic to enhance East-West connectivity, driven by several attractive features of the Arctic, including: minimal crowding with cables, less risk of damage from stray underwater items, like a fishing net or boat anchor [31,32]. The datacenter industry is gaining ground in the Nordics, with colocation providers like Equinix, Interxion, and Digiplex expanding their footprint there. American big technology firms such as Microsoft, Google, Facebook, Apple, Amazon, are also looking there for more favorable datacenter sites; each of them now have at least one large datacenter in region as well [33].
Norway has a connectivity ecosystem made up of 37 colocation datacenters and 73 cloud service providers [34] and is home to the large datacentres of Big ICT and automobile players (e.g. Microsoft, Google, and Volkswagen), among major recent entrants into Norway's datacenter market, which continues to grow [35]. With some 18 datacenters in the wider Oslo region alone and another eight elsewhere in the country, Norway has the most developed datacenter scene of the Nordics to date. In 2018 Norway published its strategy on datacenters as part of a plan to make Norway a key player in the datacenter scene globally, particularly in terms of attracting larger sized facilities [33]. Associated with this drive, the Norwegian state-owned power company Statkraft has itself sought to attract large datacenters, after the country exempted such facilities from paying property taxes in the same year [36].
Iceland has a connectivity ecosystem made up of 7 colocation datacenters and 10 cloud service providers [34]. Iceland has promoted itself for datacenter investment in terms of being a favorable business environment, including low corporate tax; the availability of land; low carbon energy at competitive prices; and regulatory efficiency within a European legislative framework and tax incentives [37]. Regarding the latter, regional incentives, according to Act 41/2015, apply to the whole of Iceland outside the capital area. Such investment agreements with the government are valid for up to 10 years and include a low rate of taxation, favorable asset depreciation rules and site subsidy. A variety of conditions are set, for example that the annual turnover of the investment project is at least 300 million ISK (€1.92 m), or 20 permanent jobs are created during the first two years; a minimum operational period in Iceland of 10 years and advance information on whether the project will legally require an environmental impact assessment.
In Greenland, there is only one colocation datacenter in Nuuk. Nuuk hosts the telecommunications facility of Telepost, run by Tele Greenland A/S, a company wholly owned by the Government of Greenland and responsible for telecommunications, IT and postal services. All settlements in Greenland with a population of more than 70 people have broadband Internet and GSM mobile phone service. Greenland also has 5 cable landfall points (Aasiaat, Maniitsoq, Nuuk, Sisimiut, Qaqortoq) connected directly or indirectly into the Canada-Iceland submarine cable and hence is connected to both North America and Europe [38]. Greenland however is in the middle of modernizing and updating its IT infrastructure, seeking to diversify away from 50 radio transmitting stations on the west coast and satellite connections for the eastern and the most northern regions to a more nationalized and reliable network [39].

"A Ferrari in the fast-lane" and "the new oil": local economic benefits and favorable business models
Across our interviews, focus groups, and site visits, datacenters were generally seen as a necessary and important part of modern, digitallymediated life. There was general approval of their employment potential, but also an expectation that this would be modest to low relative to, for example, aluminum processing plants or mines. Corporate clients of datacenters are seen as the main beneficiaries, but also publics in various roles as consumers and users of e.g. government services. This section highlights three core positive themes emerging from our data that emphasize mostly the Nordic scale: economic investment and development, capacity building and employment, and environmental sustainability. The next section juxtaposes these with three negative themes at the Nordic scale, and the final section on whole systems impacts discusses both positive and negative themes at the global or macro scale.

Economic investment, diversification and development
Participants spoke strongly about the actual and prospective economic benefits of datacenter development, for reasons including investment and diversification to greater tax revenue. In the expert interviews, A1 for example noted that datacenters were a way to diversify Arctic economies away from "oil and gas" or "mining, fishing and agriculture" to a "more sustainable and responsible form of economic development." IS2 agreed when they stated that "yes, we simply want to diversify our industries," IS4 stated that datacenters were a strong way to diversify away from "steel plants and aluminum facilities." IS2 went on to speak about how local municipalities have been "quite supportive" of datacenter development both in terms of "political action" and "financing," making it also "more attractive to international investors" who could rest assured with the high levels of Nordic product management and competence. G1 spoke about how datacenter development could lead to "cryptocurrency mining" and that via this "some politicians have mentioned the idea of using this as a source of tax revenue." A1 spoke further about some of the cultural elements of Arctic governance that would further be an asset to datacenter development, noting that "the Arctic is a special area" and "there is an Arctic way of doing things and we will ensure our local, cultural principles are respected." NO4 added that given available financing and policy support within the Arctic countries, they were "quite optimistic when it comes to funding of datacenters in Norway, actually." N04 elaborated on how datacenters could serve as anchors for future industries to come to the Nordic region, similar to the strategy Ireland used to lure American companies to set up facilities there, a strategy already being utilized by some Nordic municipalities to attract Facebook and Google datacenters to Haminia (Finland) and Luleå (Sweden). As they stated: "we hope to copy that kind of effect across the region." G1 lastly spoke more about diversification for Greenland, noting that "datacenters would enable Greenland to pivot away from mining, fisheries, and extraction to a new, environmentally benign form of economic activity … This is a great opportunity for Greenland." G8 added that "right now Greenland's economy is dangerously dependent on only four sectors or companies-Telepost for post and ICT and phones, Royal Arctic Line for sea freight, Air Greenland for air freight and travel, and Royal Greenland for fishing-anything that diversifies beyond these sectors is a very good thing … anything that could push income from export or energy generated here would contribute towards more self-rule, and Greenlandic autonomy." Such diversification would be of prime importance for Greenland, given that recent attempts to invest in mining activities in Greenland have only lowered resilience and resulted in more bankruptcies; many investors have "lost their money" and in 2020, one out of two active mines in Greenland was about to go bankrupt and needed financial aid [40].
Datacenters were also stated to offer economic resilience and cultural protection for Inuit communities in Greenland. The reasoning is that Greenland is facing a period of social change and instability, not only due to climate change but also because of its aspirations towards greater economic and political independence from Denmark [41]. Most of the country's income comes from the fishing industry or a large block grant from Denmark, which accounted for almost 40% of national GDP in 2019. Given these challenges, Greenland Group 2 spoke about how "investment in crypto-mining or datacenters could be the solution they need," and that "datacenters could bring in tax revenue and investment that would enable the national government to keep supporting rural villages (or bygd) that have immense social and cultural value." A woman in Greenland Group 2 added that "datacenter development could be a lifeline for rural, Intuit communities, not directly with jobs, but indirectly by generating revenues the government needs to keep them operating or existing, a counter to fears that industrialization will erase their cultural heritage." Indeed, one mural in Nuuk seemed to visually depict this erasure (see Fig. 2).
Another positive theme from our material related to digitization and how datacenters form the backbone of modern lifestyles in all three countries. G6 put it best when he noted: Themes of investment, development and diversification also emerged in the focus groups, although there more focus was put on tax revenues. Iceland Group 2 spoke about how datacenter development would create "more tax money for the Icelandic government, to be used for healthcare or whatever, more tax income for us would be good." Similarly, a middle-aged man in Iceland Group 1 spoke about how "If Iceland located a larger datacenter around the country and they would expedite fiber optic around the country, that would be a huge gain for us, because when those small villages around the country have access to high speed internet, they can form all sorts of companies. Being an internet company located in Iceland is probably going to produce some new, beautiful ideas".
In Norway Group 1, a middle-aged, male participant said that "datacenters also bring money into the community. People will use money in Norway. For food, gas, electricity, they will bring a lot of money into Norway, even if they are only there for half the time." Another participant from Norway Group 1 agreed and said that "I think that having a datacenter in Norway is probably better than having it somewhere else... If it's going to be the new oil, then the tax income would be a really important thing." Community development and diversification were seen as particularly important for Greenland. Greenland is the world's largest island, larger geographically than Western Europe, and yet it is covered in ice and has only about 56,000 inhabitants clustered along more than 44,000 km of coastline. The challenge, as Wilson [42] puts it, is that "businesses (both local and international) need to understand and respond to local expectations around industrial development and optimize existing local skills and knowledge for mutual benefit." Datacenters represent a critical way to diversify the economy away from mineral extraction, hunting, and fishing; and also to offer a sign of hope for Greenlandic communities suffering from unemployment, high levels of alcoholism, and rising rates of HIV/AIDS. For these reasons, Leerberg [43] writes that "Greenland urgently needs to develop new growth industries to consolidate future income," and that the "datacenter industry could step in" to provide an ideal cooling environment for servers with plenty of space.
Greenland Group 1 lastly picked up on this theme of living modern digital lifestyles. As one older man put it: When the data cable was cut a few years ago from a fishing trawler, we didn't have internet or mobile phone coverage for weeks. Nobody could use a credit card to pay for anything, couldn't send email, couldn't do work. Nobody could do telemedicine, in some places they couldn't call 911 for emergencies. The economy shut down. We need more bandwidth capacity and more datacenters in Greenland to live a comfortable, digitally connected. Modern life.
A young woman in Greenland Group 2 was even pithier: "Of course more datacenters would be good for Greenland, who wouldn't want more internet, faster Facebook, better Netflix?" Datacenter investment and development has become a necessity to life in the 21st century in Greenland, as elsewhere.

Capacity building and employment
A second core benefit arising from our data capacity relate to building, skills development, training and employment. A1 expanded further on the strong governance principles that Arctic datacenters would be encouraged to follow would lead to skills development and education for the local community. As they said: We promote the Arctic investment protocol. It's six simple principles that we asked people that develop industry in the Arctic to follow. So, three out of those six principles are related to datacenters. Datacenters have to have local support for any economic activity. You have to have local participation. You also have to have the transfer of knowledge. NO1 stated that datacenters offer "a platform for establishing new types of jobs, new type of technology, and to be attractive to other people and also we engage people not only in in Norway, but also all over the world." G1 also elaborated on this theme of how datacenters could improve training and retention for local workers in Greenland: Within Nuuk, there is an increasing disparity in wages and inequality in terms of unskilled labor working in construction and fisheries, whereas white collar managers and executives make good money. An incoming datacenter company that offers training and better wages or education for unskilled labor would bridge the gap and provide wellbeing.
In place such as Greenland, G1 articulated that datacenters could even enhance worker retention and the training of youth: Datacenters offer an exciting way of keeping people in Greenland, youth is very important for the country, programs that keep young adults in Greenland and generate job satisfaction, retention of labor, and stopping migration is critical. Promoting job creation through new sectors like datacenter is a key political strategy. N02 added given electricity or energy losses grow with distance, many datacenters need to be built close to sources of energy supply, which would invariably mean the growth in rural areas and a growth in rural skills and employment. As they explained: You have … a high voltage distribution system that … you have to transform down to a usable level. The availability of those levels where we want the power is quite limiting in some regions, meaning that we have to locate our data centers where the power is available, or will become available, if we're more or less building on a unique site with a lot of space where everything is flat. And if the power is available maybe 100 km away, the cost of bringing that amount of power is too expensive, so we have to locate where the power is produced. That will build skills and employment across many rural areas.
UK1 also brought up the point that due to concerns over latency, datacenters will always need to be distributed across broader geographic spaces, including rural areas, soat least for functions requiring speed rather than storage -datacenters "can't be very far away from their populations because latency matters, and we're limited by the speed of light. So, you need datacenters close to the populations using them." The potential employment and labor benefits also arose in our focus groups. In Norway Group 1, a young man said that "I was thinking about the local communities... Some of these datacenters are not located in big cities and they may have a lot of impact on the development of local communities… so I put employment high [in the ranking exercise] because this may have a big impact on local communities." A woman in Norway Group 2 added that "I understand that there aren't going to be a lot of direct employment jobs maybe around just working there, but I feel if it's some, then it's good. Because there are a lot of competent people in Norway who need jobs and who might benefit from working at those datacenters."

Environmental sustainability and green energy
A final benefit from our data concerned energy use and ample supplies of "low-carbon," "green" or "clean" sources of electricity and heat. Multiple respondents spoke about how Norway has "one of the lowest energy carbon footprints in the world" (NO3) due to their reliance on hydroelectricity and a "super clean grid." Iceland is a world leader in geothermal energy production, Greenland is what G4 called "a world leader for hydroelectricity" but is also diversifying into solar energy, wind power, and waste-to-energy (see Fig. 3).
Greenland in particular could have what G3 termed "immense potential" to power datacenters with clean electricity. The country already receives a huge majority (76%) of its electricity supply from hydropower. Strong investments in a network of hydropower facilities (see Fig. 4) mean that most of the larger towns and even rural settlements and smaller towns have access to hydroelectricity networks. Greenland also has some progressive energy utilities such as Nukissiorfiit that already provide district heating or tri-generation (simultaneous provision of electricity, steam, and heat), or are switching out remote oilbased generators with hybrid wind-solar PV facilities with battery banks. The Energy Sector Plan for Greenland published in 2017 also states that "The energy supply shall sustain the climate and the environment" and it stipulates that by 2024, 100% of the electricity and heat supply of the national energy company Nukissiorfiit will be based on renewable energy [44]. By 2030, Greenland as a whole has a stated goal of making sure that all public energy supply must be only from renewable energy sources, to the fullest extent possible [39]. Other low-carbon energy options that could also power datacenters include the use of fishwaste and residue into combined heat and power in Sisimiut, the use of glacial meltwater for small hydropower supply, or hybrid solar-windheat pump energy systems in rural villages such as Saarloq [45]. As Leerberg [43] surmised, "the small population [of Greenland] means that power could be largely uncontested for investors [for datacenters]." Thus, G2 concluded that Greenland's energy credentials make it an extremely attractive place for datacenters with very cheap and green energy. As they noted: I see the capacity for up to at least 15 MW of datacenters here in Greenland. The first wave of investment would have to come to Nuuk, where we have the existing capacity of hydro plant, an international port, and availability of technicians. We would need to upgrade the hydropower plant with another turbine, to provide the extra MW in one shot, but this is easy. Once done, we can provide almost free clean electricity to anchor the datacenter.
G7 spoke about the attractiveness of datacenters the following way, adding that land would also be free: It would be a good idea to bring datacenters here. At some places Nukissiorfiit (the national energy company) could provide hydroelectric power for them. Make sure they got a green energy solution, and that cost prices would be very low, like a special commercial tariff already given for the fishing industry and fish processing stations … I am confident Greenland could give them really cheap energy and an abundance of cold air. Also in Greenland the government owns all land, one cannot buy land but also does not need to purchase land. The government could ensure datacenter developers there are no costs in buying land.
G8 added that in some of Greenland's microgrids, "small settlements sometimes have 600% excess capacity, in terms of peak, it would be great to put that wasted energy to use via something like a datacenter, that could absorb it." Clean energy arose in the focus groups as a theme as well. In Iceland Group 1, one older male respondent spoke about how "we can build a chain in several areas where they have both the power sources and the land and the buildings already and we can use the excess energy from the datacenters as a form of heat recycling; and maybe in the cellars in the abandoned farms they could have this bitcoin center down there so it looks like a money center." This statement also supports the potential coupling of clean energy to bitcoin mining for sustainable development. Greenland Group 1 added "we have a super-abundance of clean energy here and could grow the national economy tenfold if needed, we have so much resource potential." 5. "A digital resource curse" and "bomb ready to explode": community risks and prospective injustices The perceived and prospective benefits to Arctic datacenters, while real, do not exist in isolation. Our data did also suggest at least three intersecting risks or injustices that could also emerge. These include the embedding of monopoly markets for data or energy; limited employment benefits but graver risks of boom and bust cycles; and waste and environmental degradation.

Monopoly markets, digital authoritarianism and data cable issues
Our data firstly raised concerns about unfair concentration of corporate power in the datacenter market, uncompetitive practices as well as monopoly issues over infrastructure such as data cables.
For example, UK1 remarked that "The bad news is that the datacenter market is not very diversified, about 60% of them are owned by two companies. Okay, one of them is called Google and the other one's called Facebook. These are not companies known for their sound management practices." IS1 added that "the prices for data cables have grown significantly" since datacenters were established, price increases that nobody could protest because they didn't own the cables. G1 noted in Greenland, that datacenters would have to face "a true monopoly on telecoms and internet provision: Telepost exists and it charges unfair rates." Issues of corporate consolidation and control and anticompetitive practices also emerged in the focus groups. In Norway Group 1, respondents said that "Facebook and Google own cables and datacenters because God forbid Facebook to be slowed down by anything human: they want to know everything about your life." A man in Norway Group 2 said "I think it's more of a risk if Facebook and Google start actually owning their own lines. Because they have a lot more power over them. That doesn't sound very good… I think I'd prefer if it was government owned or something like that… they own the information, they're transporting the information and if they're also owning the infrastructure, then hmm, I don't know." Greenland Group 1 came out the strongest against datacenter development, warning against "digital authoritarianism," with one man going on to say:

Datacenters are a way for Facebook and Twitter to decide who is President, to manipulate us and control us, datacenters are a tool they use to enslave us. Every datacenter further entrenches these companies corporate control into our lives.
Respondents in this group felt that investing in such centers was "ethically questionable" given these connections to "social control," and worried about their expansion into the Nordic market.

Limited employment, conflict and boom and bust cycles
Some of our respondents questioned the presumed economic benefits suggested in Section 4, and countered as well that investment could be prone to dangerous boom and bust cycles for communities, and the risk of community conflict.
In terms of employment, NO3 believed that opportunities would be "limited" because most datacenters are "strictly slim organizations, just running administration and sales. All facility management and technical staff are hired only, IT professionals from a company at site that is also one of our  main shareholders, not a local partner." G6 went even further, and said that datacenters provide "practically no jobs, a bitcoin mining operation doesn't even need an empty storage facility or warehouse, it's all just machines spinning and consuming power in the dark somewhere, there may not even be a janitor to sweep the floor."

C Solar energy in Nuuk, Greenland
G2 added that although even small datacenters would be considered "huge projects for Greenland," a significant barrier is "lack of technical skills and staff" and "lack of capacity to do it ourselves, we would be dependent on global actors." G7 expressed doubts Greenland even could follow in the footsteps of Norway and Iceland's datacenter investment, let alone become world leaders. As they said: Datacenters arguably "work" in Iceland    Greenland to make datacenters a possibility, commenting that "neither hydropower nor clean energy in Greenland are cheap" and cautioning that "the unit cost per MW is very high." G4 added that datacenters need not only proximity to energy sources, but also to fiber networks too:

E. The Sisimiut Waste-to-Energy Facility in Greenland F. Wind turbines outside of Sisimiut Greenland
Here in Greenland, we can do cabling with entities between cities, via our own submarine cables, but these are expensive and we have only five cable landfall nodes for the whole country. I do not see a business case for building a datacenter anywhere outside of these five nodes. Even the largest new datacenters with the most resources always place themselves near what is already a fiber highway, not trying to connect a new one. Any sort of a datacenter would have to be in the vicinity of the submarine cable structure, it wouldn't make any sense to do it anywhere else.
The viability of datacenters was also questioned by respondents on economic grounds. G6 even described energy tariffs as a financial "bomb" waiting to explode, given that they leave "no room for upgrades, large investments, or improvements" and that "debt and market distortion are sitting like a bomb ready to explode, sitting there and ticking for years and years, the whole business model will likely collapse. Because datacenters would consume so much energy, they could accelerate that collapse." Ultimately, this meant G6 felt datacenters were no more than "a political ploy" and would "not be feasible" practically.
Others spoke about boom-and-bust cycles, bubbles, and volatility. G1 mentioned how the volatility of cryptocurrency markets could enhance economic insecurity noting that: I have an inherent skepticism over cryptocurrency, so many currencies out there, no fundamental stability behind them. Investment in these things is risky, there is a real risk of losing all of your money, of adding to instability in financial markets. Datacenters could create further investment risk, more boom-and-bust cycles, which is the last thing Greenland needs.
G2 went even further, and put these concerns into the context of a "digital resource curse," expressing concern that "datacenters will be foreign owned, with foreign workers sending money home, and spending very little here, with no value add other than some small taxes to Greenland." The risk of boom-and-bust cycles is particularly apt in Greenland given the economy suffered a major one in the 1990s, when the fishing industry collapsed. NO2 stated that competition over land, resources, heat and energy could even lead to potential conflict with farmers.
Participants in Norway Group 2 referred to a specific Norwegian example of land ownership problems arising from a datacenter investment process involving HIVE Blockchain Technologies, who specialize in cryptocurrency mining. The experience had influenced their expectations of datacenters vis-à-vis sustainability practices and other environmental benefits, which the marketing material of the datacenter promises. A male participant offered "I live quite near and have read everything about it [Kolos]

in the newspaper and the story is quite ridiculous because the community literally gave away their land. The Kolos company sold their shares to some other company, now it is an unknown entity that owns all this land in Bellagen."
The focus groups also questioned community benefits, and even spoke about possible risks of conflict in the sense of being a target for terrorist or military attack. Regarding the first, a man in Norway Group 2 said that "We don't have much industry making things for the datacenters in Norway… so we have electricians… and pipeline workers and of course construction companies, but for the computer components, nothing. You buy them from China, Asia, somewhere." Continuing in this vein, a young woman working in healthcare explains why she ranked tax income last in the ranking exercise in Norway Group 1, expressing worry it could be a net "drain on taxes and social resources," and create (via migrant workers involved in construction) "conflicts over healthcare." In Greenland Group 1, a man mentioned he had visited a very large datacenter in Denmark associated with Facebook, and there, he stated there were not so many employed in this very big plant, he said it had "minimal jobs" and "most employment was during building it, not much after that." As he went on to say:

This Facebook datacenter was felt like entering a prison or a national bank, it was very secure, but it had surprisingly few people. I was of an opinion it would bring a huge number of jobs, but it did not create many. I saw two long buildings, at least 800 m long, 30 m wide, no windows, with endless corridors with racks of servers behind glass screens, but no people.
Our focus group participants were even more skeptical about the utility of cryptocurrency mining, which most people saw as "fast money", ethically dubious and unreliable (although one participant had benefitted financially from the sector and another had a friend who had bought a flat in Stockholm from the proceeds). As a man in Iceland Group 2 remarked, "… it's almost like the power companies are mining money, quick money. But if the Bitcoin system collapses, all of a sudden you have a lot of power investments in Iceland, and you need to get that back from somewhere else."

Energy consumption, inefficiency and environmental degradation
Our data was lastly critical of the potential environmental impact of datacenters on the Arctic. At the top of the list was the direct heat and energy consumption of datacenters themselves. IS1 spoke about this in relation to cryptocurrency mining: "Ninety per cent of the energy in the datacentres goes to cryptocurrency. That's a sad waste of our natural resources." Norway Group 1 also pinpointed issues of land clearing and deforestation needed for datacenter construction, noting that near one Norwegian datacenter "It looked like there was forest there that was cut down" and also that "in Norway have a lot of forest and we have a responsibility to be a carbon sink… to keep our trees so they can absorb carbon dioxide, because the rest of Europe don't." An alternative perspective was offered in Norway Group 2 where a participant offered that the datacenter built in one location had improved the area which is adjacent to a waste sorting site. In other words, clearing a few more acres of forest to accommodate a new datacenter adjacent to an existing waste sorting site was a positive environmental action. G1 raised "environmental concerns" over the use of datacenters in Greenland, noting that "energy consumption is very high, and my guess is they achieve a certain agreement with the energy company, based on hydropower, to have a cheap rate for energy consumption, below cost, which would hurt other consumers." G6 added that it gets even worse if datacenters are used for digital currency, given they are "so power inefficient, and a terrible thing from a sustainability perspective." The energy loads for datacenters and/or digital currency becomes particularly pernicious given that new datacenters would require not only electricity for direct consumption (which might be clean), but also at times heating (which would more likely be fossil-fueled) and ancillary fossil fuel consumption for increased jobs, properties, and commuting patterns that datacenters would ostensibly attract.
Rud et al. [44] and the Government of Greenland note that while electricity generation in Greenland is mostly low-carbon, heating is mixed, with three-quarters based on oil, diesel, or solid waste. In some of the larger settlements like the capital Nuuk, waste incineration plants are connected to district heating networks and additional district heating capacity is supplied by electric boilers. Individual heating however remains combination of oil boilers, electric boilers and electric resistance heaters (available only in Nuuk). Most rural settlements remain outside Greenland's network of hydropower stations (mostly in the southwest) and rely on diesel or oil heating for all of their needs. Rud et al. [44] add that oil and diesel prices are heavily regulated in order to provide equal opportunities for all Greenlanders regardless of where they live, with state subsidies "available to all," even though some investments may be environmentally destructive. Transport in Greenland remains heavily fossil-fueled (see Fig. 5).
Other interviewees spoke about the material impact constructing datacenters would have on the Arctic environment, both directly and indirectly. G8 explained it this way: Datacenters need more energy which would mean in Greenland more construction and sand mining, more roads, more hydroelectric plants, more transmission lines and more tunnels. The existing 45 MW Buksefjord hydroelectric dam near Nuuk already set the Guinness Book record for largest overhead transmission line in the world, given it had to span two mountains, and that power plant alone has 20 km of tunnels. Think of the digging needs, the civil engineering, the drilling machines for more dams. Think of the SF6 [a highly potent greenhouse gas] needed for the switchgear, the rare earth minerals, the argonite (argon and nitrogen) for fire suppression, the metal for turbines, the tailings from tunneling. The hydropower behind datacenters doesn't look as clean. And the extra energy from more hydropower stations will only bring more people, more industry, more energy consumption, and travel, and waste, and sewage dumping.
G7 added that the proximity needs of datacenters could exacerbate these risks, because they need be placed very close to sources of energy supply which would be away from population centers. Regarding the types of material reality that underpin and are consequent on the apparent immateriality of datacenter services, as well as other forms of Respondents also spoke about the relocation of communities for energy supply, and dependence on fossil fuels growing in the future, rather than receding. A theme from the Norwegian focus groups was historical displacement of Sami populations who were forcibly resettled during the construction of some of Norway's hydroelectric dams. Greenland Group 2 also raised this concern for current generations of the Inuit in Greenland, indicating that "future datacenter or energy C Diesel generators and fossil fueled district heating in Kangerlussuaq. Source: Authors. development could threaten their land rights." Updated data from Nukissiorfiit, Greenland's national energy company, are shown in Fig. 7, and reveal more fossil fuel consumption than commonly believed. Although more than three-quarters of electricity come from hydroelectricity, when one looks at the national provision of energy and heat as a whole, this value drops to only 66%. And when one includes not only national supply (mostly through islanded micro-grids) but also other private and public forms of energy supply, including industry and remote communities, the value drops to only 36%. So even Greenland, after a closer look, has a predominantly fossil fueled energy economy. And even though the country has stated goals of being carbon neutral, G2 spoke about how politicians in the country are trying to model Norway and their sovereign wealth fund, not only per se, but including through investment in new oil and gas fields. Thus, Greenland could become even more dependent on fossil fuels in the future, rather than less, boding ill for climate change and decarbonized economic development.

C. Diesel generators and fossil fueled district heating in Kangerlussuaq
In Norway, N02 spoke about how most datacenters push old technology and do not quickly upgrade, only exacerbating energy consumption and inefficiencies, given that operators keep old plants going (some built as long ago as the 1990s) to avoid the costs of upgrading or implementing energy efficiency improvements. These themes also emerged in the focus groups. One participant in Iceland Group 1 said that "here in Iceland, we don't have enough energy for electric cars, how are we going to get energy for all the datacenters?" Participants in Norway Group 2 noted that Norway imports more power than they export, due to dry conditions affecting hydropower, and thus were concerned that datacenters would place further pressure on this system. As one man said, "I can foresee the problems that will arise in society when the demands become too heavy." Other issues relating to waste, inefficiency and environmental degradation were primarily anchored to cryptocurrency miningthese are discussed more in Section 6.

From "a digital bloc" to "a sad waste of resources": examining whole systems impacts
So far, the bulk of the paper has focused on benefits or risks within communities or the three Nordic countries we investigated. In this section, we explicitly present results from a more macro "whole systems" perspective that captures sharply impacts outside of the region, at a scale beyond those of Greenland, Iceland or Norway. This is because the direct energy consumption at datacenters tells only "half the story" about the true impacts of digitization [11]. The rest of the story needs to capture impacts across the entire lifecycle, across the "whole system" including mineral extraction, manufacturing, construction, use, and even disposal and waste [46][47][48]. When a whole systems frame is applied to datacentres, this would demand an investigation of frontend aspects such as materials and the manufacturing of computers and servers, to backend aspects such as electronic waste and decommissioning, to unexpected ones arising from our data such as cultural heritage, digital terrorism and illegal spying.
In this section, we discuss three such whole systems impacts in this section in the form of three broadly positive themes (global coupling, protection of culture, and sovereignty and geopolitical strength) and three broadly negative themes (computer manufacturing and disposal, digital terrorism and the dark web, and the global erosion of privacy).

Global interconnection and infrastructural coupling
The first and perhaps most intuitive whole-systems benefit to Nordic datacenters is that they contribute to creating a more efficient global internet network, one that synergizes strengths, improves data speeds, and enhances efficiency. A1 spoke of this in terms of coupled investment between data centers, more efficient electricity grids, and greater community revenues for things like schools or hospitals. G4 articulated Arctic datacenter investment as a way to further strengthen the entire internet, making it "cheaper for us all." They said: "If a mega-data center comes to Greenland, it will help ensure a sufficient need for data for the entire region, with faster speeds, and more reliable internet. This helps everyone, the entire global internet community in fact." Their statement situates Nordic datacenter development as a way to ensure that "everyone benefits by cheaper internet all around the world."

Protection of data and cultural heritage
A second broad whole-systems benefit relates to enhanced data protection and the ability to store culturally precious information. NO1 discussed how Nordic datacenter investment would also improve programming and enable new business models or platforms that would further drive data protection and creative software. As they said, "with enhanced datacenter investment, we could develop and launch our own cloud platform … [with] people from India, from India, Pakistan all over Southeast Asia, Ukraine, from the US, all the parts of Europe." G2 mentioned how datacenters there could lead to "a digital vault for culturally important information, especially heritage, from around the world." As they went on to explain: The niche that we could go into is long-term storage of huge amounts of data that you don't need to access all of the time. For example, insurance companies, movie or video material that they need to store. Greenland can be a digital cache, an archive of sorts, and it can provide a very price efficient service compared to European datacenters. Investors have been very impressed with prices, low to no costs for cooling, less costs for power. Greenland can offer a green and cheap service for datacenters .. We could keep important data available even should there be a nuclear war, this is the kind of service Greenland can provide. Cultures from around the world can safely store their data here, away from anybody. This safe and secure long-term storage of something important that doesn't crash if Europe and North America crash.
G3 also added that "the High North offers a great place to store data very securely, we can store information for 100 years, 150 years." This would offer human civilization a source of data security believed to be unavailable in most other parts of the world.

Economic sovereignty and geopolitical strength
A final whole systems benefit encapsulated the perceived utility of creating a "digital bloc" that would enhance both sovereignty for places like Greenland and geopolitical strength for all Nordic countries.
To offer some context for this line of reasoning, existing data infrastructure in the form of high-speed subsea cables is shown in Fig. 8. Although some Nordic countries such as Denmark are and Norway have many data cable connections, their connectivity pales in comparison to Ireland or the rest of Europe, while exceeding the small number of connections to Iceland and only two direct connection to Greenland (one via Iceland, one via Canada). Other than these two direct connections in Greenland and three smaller nodes, the rest of the country receives slower internet via radio satellites, one of them displayed in Fig. 9.
Multiple sources of evidence suggested that an expansion of both cables and datacenters would enhance the geopolitical strength of the Nordic region and also serve as a healthy counterbalance to North American dominance of the datacenter market. G2 spoke about how "investment in Arctic datacenters serves the strategic importance of diversifying infrastructure away from the United States or China." G6 articulated how: New datacenters and high-speed subsea cables for Greenland would counterbalance dominance from the Americans. We are actively looking at connecting a data cable to Ireland or the Faroe islands, Ireland because it is the most heavily connected country in Europe. We are really desperate to invest in this area, it would mean if the cable to Canada is cut, we could redirect to Ireland. This would benefit not just Greenland, but the world. It is of geopolitical significance to offer more ways of accessing data and the internet beyond the United States entities that own all of the cables.
G4 also expanded on this theme, arguing that "the stronger presence of Nordic countries is only good for the global datacenter market." The theme was also picked up during our focus groups. Iceland Group 1 participants exclaimed that "I thought we were bigger in this market [for datacenters]" and that "we are missing out, that there should be more [subsea] connections." Greenland Group 2 said "with new datacenters and cables, we could form a digital bloc, a sort of OPEC [Organization of Petroleum Exporting Countries] of the digital world." Greenland Group 1 even spoke about how such investment could inspire all countries, noting that "datacenters would not only help the economy and do little harm to the environment, they could preserve culture and inspire other social movements to self-determination, showing how resource based economies can transition to non-extractive economic activity."

Computer manufacturing, shipping, and environmental degradation
A suite of whole systems negative impacts juxtapose the potentiality of the positive ones. For example, respondents discussed how datacenter investment will depend on material intensive computer manufacturing and materials. G2 in particular worried about "the sustainability of datacenters over the long term, after 10 or 20 or even 40 years." As they went onto explain: You would have all of the environmental impacts of building the IT equipment, shipping it all the way to Greenland or other Nordic countries. Most building materials would need to be imported, along with the technical hardware, and in Greenland we would import everything, from computers, servers, cables, and windows to batteries, desks, and entire offices.
Indeed, Murugesan [15] notes that computer manufacturing does involve electricity, chemicals, raw materials, and hazardous waste. They also calculated that each personal computer in use at that time generated "about a ton of carbon dioxide every year." The total electrical energy for an entire datacenter would be many times this. Rapid turnover of equipment also suggests that organizations dispose of their information technology equipment every 3-5 years, creating a turnover of servers, monitors, data communications equipment, and cooling systems, much of which ends up in landfills.
In addition to these concerns, subsea cables themselves were identified as having whole-systems impacts throughout the ocean floor and marine environment. One aspect of this is accidents; G7 spoke about how fishing trawlers often destroy cables unintentionally, creating extra damage and requiring expensive repairs. (The internet also "goes out" for days to weeks when this happens). G6 expanded on this theme by noting: We have had problems with fishing vessels, with icebergs, with faulty repeaters on the ocean floor, problems others also have, but we are the only cable operator in the area. Building new cables is a resource intensive process that also does have environmental impacts.  Three specific aspects of environmental degradation were the disruption of marine habitats, polar bear attacks while laying cables, and interference with whales. One man involved in cable surveying in Greenland Group 1 stated that: To lay those cables, you have to plow down the ocean floor, and that disrupts marine life. The sea floor isn't straight like a road, so laying a submarine cable is hard, essentially like crossing a high altitude mountain range underwater, and when laying the cable you need to watch out for ice bergs and even polar bear attacks. You need a 50 ton plow to make a trench, which generates dust and sand, then close it again so fishing nets don't catch the cable.
Iceland Group 1 also expressed concern about "so many lines through the sea… seeing it actually makes you wonder about the ocean. About the sea life." Another man in Iceland Group 1 stated that "I think I'm remembering correctly that they were linked to disturbing whales and others that use magnetism to travel around… I don't remember if it was disrupting their communications or their magnetism to find their way." A third whole systems environmental concern was, according to G7, increased shipping activities which could result in a greater risk of shipping or construction accidents. The WWF Arctic Program [49] has also argued that greater rates of shipping not only lead to more shipping accidents; it also leads to more underwater noise that disrupts aquatic life, more greenhouse gas emissions, and more waste streams generated by the shipping industry. G8 added that "the harbors in Greenland are not that large, so any significant expansion of population or economic development would require potentially new, larger ones, and that could in turn create a wave of new shipping." Fig. 6 (above in an earlier section) shows the Sikuki Nuuk Harbor, the largest in the country, but even its container capacity is limited; summer closures are common due to tides; and winter closures common due to ice blockage.
A final whole systems environmental concern relates to waste and electronic waste, or e-waste. As G2 put it succinctly, "on the backend of datacenter management, you have issues of waste, requirements to clean up and decommission if you close, and some very large costs with setting up huge facilities like that and dealing with them when they close down or are abandoned." There was some skepticism in Norwegian group 1 regarding the effectiveness of EU electronic waste recycling regulations: "We know about the toxic waste disposal industry. It's wide open for corruption."

Money laundering, ransomware and the dark web
Another class of whole systems risks related to greater rates of money laundering and cryptocurrency mining, as well as the proliferation of ransomware and the dark web. Firstly, multiple respondents across both the interviews and focus groups expressed concern over greater datacenter deployment and consequent increases in both digital money laundering and connections to ransomware attacks and the dark web. Iceland Group 1 believed that as much as "ninety percent of the energy in the datacenters goes to cryptocurrency. That's a sad waste of our natural resources." It also places Nordic datacenters in a proliferating web of global cryptocurrency traders and markets. NO4 remarked that: The business people behind bitcoin are not coming from the datacenter industry -they are coming from the money industry, and they are buying equipment from China … and the equipment has got the lifetime of let's say one year maximum and then you just have to throw it away and buy new equipment, and then install it for doing mining … And it's anonymous in many ways, out of the bank systems.
NO1 added that this anonymized, location-free form of money makes it "not so easy to control money laundering all the time." UK1 also expressed concerns about money laundering, commenting that: IS3 agreed and noted that such actions have little relevance for domestic markets, but prop up international traders. G1 also said that "in Greenland, it is not really clear about the connections between cryptocurrency mining and taxation as well as financial crime and laundering money. I am concerned datacenters could transfer resources to criminal organizations." Secondly, respondents mentioned how increased datacenter deployment would also prop up or facilitate actions on the dark web. NO1 admitted that "It is very difficult to control the dark web and we all know that." NO2 added that while they don't take any such clients, they do know of some datacenters who are not "socially responsible" and being linked to "gambling and online pornography." G6 added that "the cheaper Nordic datacenters and the internet become, the easier it is to produce bitcoins which many hackers use for illegal activities. Greater datacenter deployment can lead ironically to increases ransomware attacks." Greenland Group 1 also expressed anxiety over this theme, noting that "as datacenter processing power increases, won't that only embed the dark web and all of the evil things online?" Iceland Group 1 raised concerns over "dangerous artificial intelligence misuse" at datacenters.

Digital terrorism, the erosion of privacy and illegal spying
Our final whole systems risk revolved around digital terrorism (referred to briefly above) and online spying. Greenland Groups 1 and 2 mentioned a surprising (but true) connection to data cables and spying. Greenland Group 1 mentioned that "the subsea data cables running to datacenters are very vulnerable to spying, right now Danish intelligence has been cooperating with the United States intelligence to secretly spy on European leaders through the subsea cables." Greenland Group 2 added that "more datacenters means more sensitive data can be stored and sent through subsea cables, and these cables can leak information, intelligence agents can capture a data cable and use it, this is a very big risk, a very big security vulnerability." We checked these claims about spying and indeed the National Security Agency has been accused of collaborating with the Danish Defense Intelligence Service to use data cables to illegally spy on senior officials in France, Germany, Norway, and Sweden from 2012 to 2014 [50]. Spies apparently used the cables to intercept text messages telephone calls, private messages, and internet traffic.

Conclusion and implications
Datacenters have a material and visceral human impact on our communities and the sustainability of the planet well beyond the digital services or internet connections they provide. Such datacenter development in the Nordic countries of Greenland, Iceland and Norway can diversify economies, build local capacity, generate jobs, and rely on cleaner forms of energy than many other parts of the word. But they can also buttress monopoly markets and anticompetitive business behavior, place communities at risk of boom-and-bust cycles and spawn community conflicts over farming or healthcare. They can lead to increased energy consumption for cryptocurrency mining and their own forms of environmental degradation. As a result of these tensions and complexities, Nordic datacenter development can simultaneously reinforce but also erode some of the Sustainable Development Goals shown in Table 3.
A common observation across all our participant groups is the existence of contestations (of varying degrees) between (1) the need to accept the changes that come with datacenter adoption especially in the areas of diversification (for resilience), investments (for job creation and income generation) and development (for improved infrastructure and services); and (2) the need to preserve and sustain their culture (practices, tradition, landmarks etc.) or manage other environmental tradeoffs (energy use, heat, land use, life below water). This contestation for instance exists in the acknowledgment by participants of declining revenue from existing industries and their inability to guarantee sufficient jobs for the younger population on the one hand, and fears over the intrusive and debilitating impact of datacenters development on their environment on the other hand.
Moreover, the impact of Nordic datacenter development is not limited to the countries examined. Fig. 10 shows a host of intricate, internationalized effects arising from our data that range from positive global benefits such as a more efficient and resilient global data network, improved global internet speeds and data center capacity, and upgraded digital connectivity of the Arctic to enhanced data protection and geopolitical shifts in digital power. However, these are contrasted by greater rates of money laundering cryptocurrency mining, proliferation of ransomware and the dark web, an increased susceptibility to digital terrorism or hacking and even the erosion of privacy and spying.  Going forward, we see a number of areas as priorities for research but also public debate in relation to Nordic datacenters. First, it is unclear what the implications of growth in the datacenter sector are for the scale of renewable power supply in the countries examined. While it can be imagined that power supply companies may have undertaken their own modelling on this, we were unable to access corresponding data. Such studies should quantify the economic and environmental costs and benefits of expansion, including the site-specific impacts of additional power supply, and taking account of any climate change impacts on hydropower availability. Secondly, we see a need for wider debate as to the differing economic value of datacenters that operate for different purposes, particularly in the context of any opportunity cost. We are thinking particularly here of cryptocurrency mining. Thirdly, and related to this, is the question of how datacenter host countries can extract more value generally from the sector, given the importation of equipment and relatively light staffing. Of course, there are many more issues from a wide variety of perspectives, but the foregoing are among those that are of applied interest, as datacenters head to the Nordics.
In sum, the digital services so instrumental to the internet and modern everyday life mediate not only work, entertainment, and communication, but an entire constellation of factors spatially extensive and deeply important. Our study thus validates the multi-scalar and whole systems approach to datacenter evaluation and governance, bringing into focus a broader range of issues-and scale and scope of impacts-that are as difficult to manage as they are important to address.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.