Data centres on the Moon and other tales: a volumetric and elemental analysis of the coloniality of digital infrastructures

ABSTRACT

Recent years have seen the expansion of data centres in mountains, deep oceans and outer space. These infrastructures claim, adapt and interrupt environments for data needs – at times prioritizing these over human and ecological needs, and often to the benefit of only a limited number of stakeholders. From this departure point, this paper explores how volumetric and elementally attuned analytical frameworks can make explicit some of the ways that colonial relations extend to digital infrastructures. The paper explores two vignettes of data centre infrastructures – in the Atacama Desert and in outer space – looking at a range of pre-existing, in-progress and speculative projects. Through this examination, it considers how these frameworks, which stem from Indigenous epistemologies, help to foreground certain processes. First, the obvious and non-obvious ways that colonial logics continue in these infrastructures. Second, understandings of how the volumetric and elemental are differently embodied – such as how thin air mediates labour – which helps to counter flattening discourses that serve to benefit certain actors. Third, the colonial futurities that are embedded in data projects such as those related to radio astronomy, the distribution of orbital flight paths and the demarcation of oceanic graveyards as well as the plans for data centres on the Moon. Finally, this paper also attends to alternative worldviews that can help to unsettle colonialist trajectories.

KEYWORDS:

• coloniality
• data centres
• outer space
• elemental
• volumetric
• infrastructure
• digital geography
• data

1. INTRODUCTION

Recent years have seen the expansion of a massive variety of infrastructures relating to the storage, processing and transfer of data. A notable part of this growth involves claiming, adapting and interrupting environments for data needs, such as through the building of data centres (Amoore, 2018; Hogan, 2018; Hu, 2015). As data centres creep into ever higher mountains, deeper seas and even outer space, they are opening up volumetric geographies, bringing with them competing jurisdictional claims, geopolitical influences and socio-environmental legacies, all of which are also influenced by localized elemental considerations. To critically examine the expansionist trajectories of data centres and their related infrastructures, this paper examines two vignettes of digital data assemblages, one situated at some of the highest altitudes on Earth and the other in outer space.

Critical scholarship argues that digital infrastructures present yet another sphere of life where coloniality and imperialism live on (Abdilla et al., 2020; Aouragh & Chakravartty, 2016; Tawil-Souri, 2012). Building on this, this paper investigates how foregrounding the elemental and volumetric dimensions of digital assemblages reveals different ways in which spaces are occupied, colonized, inhabited and experienced. The focus here is on digital data assemblages, or the socio-political, technical and economic apparatuses that constitute the production, circulation, and deployment of digital data in complex and intertwined ways (Iliadis & Russo, 2016; Kitchin, 2014). This includes the physical and material underpinning of computing systems such as cables and servers, but also other elements including protocols and standards, norms, governing institutions, labour, and marketplaces. As such, examining data infrastructure involves studying the ‘things and also the relation between the things’ (Larkin, 2013, p. 329). While digital data assemblages understood in this way present an unruly subject of study, this paper argues that volumetric and elemental geographies offer a useful conceptual toolkit to excavate different dimensions of such infrastructures.

The past decades have seen vibrant conversation among geographers on how best to approach the analysis of space. The tendency for earlier scholarship to emphasize categories such as surface area has prompted a focus on verticalities as a way of contesting ‘flat’ traditions, by paying attention to the ways that territory and organization are constituted through height, stacking and other vertical dimensions. However, as Graham and Hewitt (2013) argue, instead of attending to the vertical as a reflexive counterbalance to pre-existing biases towards horizontalism, there is a need to better understand how horizontal and vertical extensions, materialities, lived realities, and imaginaries intersect and constitute each other. The subsequent ‘volumetric turn’ in geography responds particularly to this concern, and this body of literature contends that verticality alone is an insufficient analytical approach to comprehensively understanding the three-dimensional complexities of the world.

The volumetric turn has instead emphasized the stacking and moving of people, things and ways of being above, across and below, paying particular attention to dimensions such as density and depth, and to the interactions that occur within. As McNeill (2020) describes, part of the volumetric agenda involves addressing how horizontal and vertical dimensions co-constitute each other through volume. A rich body of work has opened up in response to these calls, with this paper building on two particular streams in volumetric literature. First, it builds on scholarship excavating the embedding of digital flows in the volumetric world and how they continue the colonial logics of the non-digital world (Amoore, 2018; Hogan, 2018; Pickren, 2018; Tawil-Souri, 2012; Weizman, 2004). Second, it builds on scholarship that seeks to excavate elemental and sensory qualities of volume in order to foreground different material registers of volume (Adey, 2013; Billé, 2018; Jackman & Squire, 2021).

This elemental entanglement involves moving beyond a static understanding of the elements and, instead, foregrounding the ways in which natural systems, such as air, intersect and interact with the socio-political, the human, and the non-human to ‘form parts of assemblages but also proving agential in their own right’ (Barry, 2013, p. 34). For instance, Squire (2016) examines the way elements function within the contested territory of Gibraltar, where she argues that elements such as rock and air are entities in their own right in the ongoing spatial dispute. Elements have their:

own specificities, volumes and properties that interact with the body in a co-constitutive relationship that shapes and critically underpins geopolitical constructions of the territorial volume. The molecular structure of limestone, for example, has enabled and disabled certain geopolitical realities throughout Gibraltar’s history; the mobilities and power of the sea and its waves have ignited tensions as power is exerted by and over this elemental substance; air both contained within a balloon and in the atmosphere becomes the driver of affect. (p. 559)

Given their entanglement in a variety of volumetric and elemental complexities, data centres are a fertile ground of study. This paper asks: How can a more volumetric and elementally attuned understanding of digital infrastructures, such as data centres, reveal how colonial relations persist in these processes? It dissects this question by examining two vignettes about pre-existing, in-progress, and speculative projects of data centre assemblages on Earth and in outer space. Through compiling a variety of primary source material (such as advertisements, company speeches, legislation/protocols and patents) and secondary sources (such as news articles, market trend reports, policy documents and law reviews), these vignettes tackle this question in four ways: first, they examine how colonial logics are perpetuated in obvious and non-obvious ways; second, they emphasize how elemental systems mediate digital data assemblages, including processes of labour and pollution; third, they probe how colonial futurities are ultimately being embedded in data assemblages; and lastly, they suggest alternative worldviews that can help counter colonialist trajectories.

A more comprehensive answer to the broader question of how attending to the volumetric and elemental can help excavate the way colonial relations are embedded in data infrastructures involves continuing to unsettle commonly held notions, such as how we understand infrastructure (Star & Bowker, 2006), our relationship to nature (Smiles, 2020) and even how we approach pollution (Liboiron, 2021). One particularly stubborn orientation that has been critiqued across vertical and volumetric literature is how research tends to skew towards Western and state/military-centric frameworks that privilege territorialization, control and conflict (Jackman & Squire, 2021; McNeill, 2020; Squire, 2016). While this paper still largely traverses these topics, it also considers what alternative imaginaries of the volumetric could be: what it might look like to heed the calls by feminist scholars to destabilize disembodied discourses, or to be informed by Indigenous scholarship, which attended to voluminous worlds long before the volumetric turn. This includes going beyond the idea of calculable, measurable and controllable volumes – whether it is the calculable sky, a neatly laid out equation of environmental risk, or an outer space that can be perfectly divided and managed. It involves embracing the embodied messiness of reality and the unpredictability of the elements.

The paper is structured as follows. First, it considers previous scholarship of elemental, digital volumes and how colonial logics persist in these spheres. It then sets out how the volumetric and elemental lens might be useful in exploring coloniality in infrastructures and explains why data centres are a productive site of study. Following this, it dives into two vignettes of data centres: at high altitude and in space. The paper then concludes with a discussion of how an elemental, volumetric lens can help reveal the messy colonial realities and futurities entangled in the expansion of digital life as well as with a consideration of alternative imaginaries of the volume.

2. LITERATURE REVIEW

2.1. Coloniality in digital infrastructures

Digital data and the volumetric spheres of life are differentially hidden, obfuscated and imperceptible to human senses and daily life. Various bodies of literature have emerged to examine exactly how these processes, infrastructures and power relations are imbued in these assemblages. This includes scholarship on the materialities of digital geography, which seeks to push against the inclination to think of digital technologies as placeless, invisible and immaterial, and recent scholarship on complicating volumetric territories, which seeks to push against the inclination to fixate on the horizontal and on the visible as well as the tendency to see natural elements as static. Both bodies of work also draw on the implications of how digital volumes are arenas that extend the colonial logics and processes of claiming and occupation.

The scholarship on the materialities of digital geographies (Amoore, 2018; Kinsley, 2014; Pickren, 2018) has grown alongside the infrastructural turn in science and technology studies which has prompted many incisive studies viewing the internet and digital technologies as infrastructures and considering the implications of this (Aouragh & Chakravartty, 2016; Parks & Starosielski, 2015; Plantin, 2017; Star & Bowker, 2006). Much of this literature discusses the invisibility of dataspaces, where examinations have foregrounded what has been rendered out of sight to the general public, including the mapping out of internet nodes and networks (Parks & Starosielski, 2015; Tawil-Souri, 2012). Work in this area also investigates the tangible spaces of internet circulation such as data centres and internet cafés (Harris, 2015; Hogan, 2018; Hu, 2015; Mosco, 2014; Munn, 2020). Others have focused on untangling the effects of what is less possible to visually render, including protocols and standards in the fields of science and technology (Cath, 2021; Star & Bowker, 2006; Star & Ruhleder, 1996), the data within cloud computing (Amoore, 2018), and the algorithms that stratify social life (Noble, 2018; Plantin & Punathambekar, 2019).

This unearthing of how digital technologies are tethered to geographies and power relations has also become an important space for the conversation about how colonial relationships – hierarchical, extractive and exploitative relationships (Amrute, 2020) – continue in these spheres. Critical scholars have carefully traced the myriad ways in which infrastructures are inextricably intertwined with colonial and imperial relations (Abdilla et al., 2020; Amrute & Murillo, 2020; Aouragh & Chakravartty, 2016; Kukutai & Taylor, 2016). For instance, Tawil-Souri (2012) examines how Israel’s territorial control over Gaza is perpetuated through the high-tech realm, including through shaping and limiting the digital and telecommunication flows within the Gaza strip. Tawil-Souri reminds us that while border technologies such as communication wires, radio-wave signals and internet switches may be less visible than the physical walls, iron gates and imposing checkpoints of the physical world, their effects are no less tangible.

Indigenous scholars have also studied the physical claiming of territorial volumes in digital infrastructure projects and the deep ties of this practice to settler colonialism, a form of colonialism that is based on the permanent presence of colonists upon land, which is distinct from forms of colonialism based on resource and labour extraction (Smiles, 2020; Veracini, 2014; Wolfe, 2006). This permanent presence by settlers is at the expense of Indigenous or local people in a given space, where the ultimate aim is the occupation and remaking of space, which is often justified by conceptions of terra nullius (Latin, ‘nobody’s land’) – that is, that before Western presence, the space belonged to no one. The coloniality of such infrastructures is perpetuated by governments and corporations alike. For instance, Hogan (2018) examines this in terms of how companies that own data centres not only claim land for these physical infrastructures but also embed a type of ‘settler futurity’ – that is, the configuration of settler colonialism towards a certain futurity, or the ways in which the future is rendered knowable through practices such as calculation and imagination, which in turn intervene on present time through anticipatory logics such as precaution and preparedness (Tuck & Gaztambide-Fernández, 2013, p. 173). Thus, a diverse set of critical works on the coloniality of digital technologies exists. A commonality in this myriad of examinations is how colonial processes are often not easily visible or obvious.

2.2. Elemental and volumetric perspectives

One way of extending the work on the coloniality of digital infrastructures is to shift perspectives to foreground the different ways these relationships exist and are mediated. The literature that seeks to complicate the volumetric aspects of life is particularly helpful towards this end. This stream of research asks what it means to have a politics of verticality and volumetrics, including what the power negotiations are that occur above our heads and beneath our feet, what flat orientations of the world might omit, and what it means to more comprehensively understand what ‘fills’ the volumes of the world (Elden, 2020; Graham, 2016; Graham & Hewitt, 2013; Harris, 2015; Squire & Dodds, 2020; Weizman, 2004). Recent work has sought to revisit the elemental entanglements of the volume and how we sense the multitude of spaces we are embedded in. This includes works that tease out the ontological, epistemological and socio-political implications of encounters with air (Adey, 2013) and mountainous altitudes (Gordillo, 2018), of watery volumes (Starosielski, 2015), the sun (Engelmann & McCormack, 2018), and the subterrain (Nolte & Özdemir, 2018), and of ice (Dodds, 2021) and fire (Squire, 2016).

However, the work of examining the elemental entanglements of the volume goes beyond merely acknowledging the existence of the elements. As Lehman (2013) writes, while the natural environment is often integrated in research as an important factor of socio-political processes, it is often treated as having passive features or providing resources and constraints. This strand of work moves beyond a static understanding of the elements, seeking instead to foreground the way systems such as air and water interact with the political, the human and non-human, and the material and immaterial to ‘form parts of assemblages but also proving agential in their own right as they enter into relationships with the social and political’ (Barry, 2013, p. 34).

An important part of this agenda is thus being more expansive in how we come to understand, sense, and embody the elemental volume. For instance, in examining border spaces, Billé (2018) argues that the predominance of sight can come at the expense of other understandings of the border, such as tactile ways of knowing and embodying our surroundings. Similarly, Jackman and Squire (2021) point out how researchers have been operating on a limited methodological toolkit in the exploration of voluminous complexities, where they advocate for more explorations into immersive and multisensory methodologies – such as using sound to highlight different ways of experiencing the volume. Here, they point to Indigenous and feminist scholarship being formative where in particular, Indigenous scholars have long-noted the inherently ‘voluminous’ ways in which a range of actors live through every day. For instance, how Hau‘ofa (1998) talks about the significance of the sea to Pacific Islanders by foregrounding the sea’s shoals, depths, and mobilities, and the profound and mundane ways in which they shape Pacific community lives.

3. THE COLONIAL PERSISTS: AN ELEMENTAL AND VOLUMETRIC TOOLKIT

Taken together then, these wide-spanning bodies of work on coloniality, volumes, and the elements ask: what are we overlooking, and why does it matter? This corpus of scholarship urges for the three-dimensional qualities of Earth – even ones that are less obvious and undetectable by human senses – to be made more concrete, more explicitly theorized and excavated. Even if certain processes have tendencies to be less tangible, their effects can still be felt strongly by those affected – particularly by those harmed, excluded and displaced from the systems. Furthermore, despite how the work on the colonial entanglements of digital technologies spans vast geographies, a recurring theme is how colonial processes and legacies are not always obvious but imprint a certain spatiality and temporality in important ways.

A lens that is more attuned to elemental, volumetric and Indigenous epistemologies can help extend work on the coloniality of infrastructure (digital or not) by shifting perspectives to foreground the different ways these relationships exist and are mediated. The vocabulary and concepts stemming from this diverse set of work help to excavate the dimensions of socio-political life that are overlooked, obscured, and often imperceptible to various human senses. It can help reveal the ways that digital systems co-constitute how spaces are colonized, divided, claimed, inhabited, and experienced. This includes the active ways the elemental and the volumetric form parts of assemblages and the ‘fragile, minute, world-specific localization’ (McCormack, 2015, p. 87) of these processes. These close inspections help to push back against disembodied and flattened discourse on digital technologies. Specifically, it includes exploring:

• The obvious and non-obvious ways that colonial logics and the claiming of territory continue in such infrastructures (including beyond land), and the elemental specificities that complicate this. This might include, for instance, the claiming of orbital pathways in low Earth orbit, where physical airspaces are a limited resource; the authoritative distribution of the radio-wave spectrum, where limited frequencies are available; the claiming of crucial strategic locations; or the claiming of light space in the night sky. A more expansive view of the elemental volume helps us to grasp the breadth of colonial processes along multiple axes.

• How the elemental volume is differently embodied and experienced. Emphasizing the elemental volume is one way of contending with different lived realities and how digital expansion affects processes in labour, maintenance and pollution in ways that simplified equations of risk leave out. In asking ‘what else fills volumes’, Adey (2013, p. 52) encourages the examination of objects, bodies, atmospheres, lived experiences, and other things that are less calculable – what McCormack highlights as how human or non-human mediates affect one another: ‘how wind becomes shivering, sun becomes warmth, gas becomes death’ (McCormack, 2015, p. 87). This can help uncover the political work of flattening and equalizing discourses that paves the way to colonial processes; whether it is the flattening of labour realities, Indigenous epistemologies, or myths about outer space being inherently egalitarian. Attending to elemental specificities counters the passive understanding of natural elements as merely providing either resources or constraints.

• The colonial futurities that are being embedded in a variety of data projects, even well-meaning ones. As Hogan argues, ‘data centres play a hand at furthering settler futurity in at least two ways: they stake a claim to land and its proximity to water, and they further locate humanity outside of human bodies, and into machines’ (Hogan, 2018, p. 640). Similarly, as Smiles (2020) contends, fantasies of terra nullius frame territories on Earth and in outer space as claimable and for the taking. Excavating the coloniality of infrastructure thus also requires the examination of the assumptions of the future that underlie our infrastructures.

• The alternative worldviews and imaginaries that can contribute to alternative experiences, ways of knowing, and potential futurities of the messy world – including the idea that the volumetric and the elemental are not always calculable, predictable, and separatable. While the volumetric – particularly as implied by ‘metric’ – is calculable (Elden, 2013, p. 35) and can be measured, surveyed, managed, ordered, weighed, and controlled, volumes can also produce much more than calculable, scientific, and political–technical insights (Adey, 2013; McCormack, 2015; Squire, 2017).

These interrelated points help to unpack, contest, and unsettle disembodied and flattened discourses of digital expansionism. In de-naturalizing these processes of coloniality, the active work to displace territory, prioritize certain needs and claim space is made more explicit. In particular, it is important to have a more proactive approach to how settler-colonial and other imperial futurities are being embedded in our collective schematics for the future: such as a future where the night sky can be claimed or where global commons such as the Moon are sold as real estate.

4. THE CASE OF DATA CENTRES

One way of exploring elemental digital volumes is by examining data centres, which are diverse in use, size, stakeholder arrangements, architecture, and how they use, change and disrupt the environment they are built in. In their most basic sense, data centres host buildings of servers (computer hard drives usually stacked in rows and generating significant amounts of heat). They are often outfitted with cooling and heat-recovery systems as well as backup equipment in case of power failure (Diguet & Lopez, 2019). These data infrastructures are foundational to a whole host of socio-economic and political operations, from stock market exchanges to mass surveillance. In recent years, data centres have become even more embedded in our environment. They are nestled in dense cities, urban peripheries and in uninhabited lands.

The geographical sprawl of data centres is shaped, in part, by their operating requirements. Data centres demand significant amounts of energy to power their cooling systems, which counteract the heat generated by the servers and computing equipment. Cooling systems often represent 50% of a data centre’s electricity consumption (Diguet & Lopez, 2019). Related to this is the need for reliable, inexpensive electricity sources. Whether data centres tap into existing electricity networks or invest in new infrastructures or collaborations, they interrupt the energy projects of the localities they are built in due to their high electricity consumption. Generally, they also need to be in a secure place, ideally one that is out of the way of natural disasters and that is challenging for outsiders to penetrate. Proximity to renewable energy, skilled labour, agreeable tax incentives and governance structures are also all crucial factors. Furthermore, there is a need for affordable land. Similar to how data centres impact the energy demands of the territories they are in, they also affect the local market value of land as well as various other aspects of life – in some cases displacing people and ecologies in favour of meeting the data needs of certain actors.

The development of innovative data centre architecture has paved the way for these facilities to be built in and around an incredible variety of environments. This includes data centres clustered around urban hubs (e.g., to serve edge computing needs where servers need to be close to the site of service) as well as data centres built into natural environments that can help meet the cooling needs of these infrastructures. In particular, data centres are increasingly making use of what the industry calls ‘free cooling’ – that is, using natural cooling systems. For instance, the Swiss Fort Knox facility is built in an old Cold War bunker and is able to withstand a nuclear blast. It is also cooled by its location deep under the mountain, using glacial water from an underground lake in its cooling systems. The Room 48 data centre facility in the United States is similarly built in a repurposed mine, where limestone walls help to absorb heat (Iron Mountain’s Energy Efficient Bunker, 2010). Likewise, Microsoft has been testing the viability of aquatic data centres through the Northern Isles data centre: a cylindrical structure that has been plunged 117 feet underwater, on the seafloor off Scotland’s Orkney Islands, and that uses free cooling from the water (Innovation Stories, 2020). On the other end of the vertical axis, data centres are being developed to operate in outer space and even on the Moon.

To explore this further, the next section examines two vignettes. The first looks at the data infrastructure landscape in the Atacama Desert and illustrates: (1) the coloniality that continues through the claiming of crucial mountainous locations, the night sky and a type of spectral occupation; (2) the embodied experiences of maintenance labour at high altitudes and flattening rhetoric that validates digital expansions; (3) the futurities embedded by the astronomy data landscape; and (4) the discontinuities between Indigenous astronomy and radio astronomy that perceive the elemental volumes in different ways. The second vignette looks at the speculative data centre projects in outer space and illustrates: (1) the coloniality that continues through the claiming of orbital paths, radio-wave spectrums, airspace, the ocean and the Moon; (2) the elemental entanglements of outer space, the sky and the oceanic volumes which are flattened in calculations of risk; (3) the colonial futurities assumed, for instance, in the division and distribution of limited resources related to outer space; and (4) the perspectives that view nature as part of a global commons.

5. DATA CENTRES IN THE ATACAMA DESERT

Currently, the data centre located at the highest altitude is in the driest non-polar desert in the world: the Atacama Desert. What is more obscured in a flat map is the vertical and volumetric variation in this environment. The desert’s mountainous regions are characterized by high winds and snowfall, low humidity, dust accumulation, low temperatures, and a thin atmosphere of oxygen that makes it tricky for staff and equipment to work in. The Atacama Desert also has some of the most conducive environments for night sky observations with minimal cloud cover – experiencing over 200 cloudless nights every year – and lack of light and radio interference from urban hubs. For this reason, it is the location of almost two-thirds of the world’s infrastructure for astronomical digital data production (Detwiler, 2021). The data, optical and radio astronomy projects in the region are all expected to expand in the coming decade. As of 2019, Amazon has partnered with the Chilean government to host, consolidate and analyse astronomy data across all Chilean observatories in the Atacama Desert. Google is also expanding its only data centre in South America, which relies solely on solar power generated in the Atacama Desert (Hall, 2018; Sherwood, 2019). Here, global capitalist powers shape the contours of digital infrastructure in the region.

One of the data centres located here is part of the Atacama Large Millimeter Array (ALMA) astronomy observatory facility, where air, in particular, plays multiple dynamic roles. Sitting atop the Chajnantor Plateau in Chile, Cisco maintains that this data centre facility is adapted for the high-altitude environment at 5000 masl, which enables the free cooling of data servers (Cisco Newsroom, 2016). The data centre serves computing processes at the astronomy facility and a supercomputer that is adjusted for operations at high altitudes. The thin air here means that almost twice the normal airflow is necessary to cool machinery. Furthermore, computer disk drives do not work reliably in this air so computers must operate diskless. Computing parts are also made to withstand seismic activity common in the area (Miller, 2013).

The thin atmosphere and mountainous terrain also means a difficult environment for workers to operate in. This became evident in the largest labour dispute that occurred at the ALMA facility where in 2013 80% of the workforce went on strike for better working conditions and pay. The 194 striking workers, who were mostly Chilean engineers, technicians, administrative staff and maintenance crews (but not foreign scientists), cited issues including the harsh working conditions at high altitudes, which were further exacerbated by how isolated the remote area is, the long hours (including 12-hour shifts to keep the facility running 24/7), and inadequate sanitation at the facility (Hurtado, 2013; Witze, 2013; Young, 2013). The risks of working at such heights include altitude sickness and hypoxia, as well as being far away from larger medical facilities in cases of emergency (Jamieson, 2011).

Another consideration of air and sky space is its use by Indigenous and local populations in the Atacama Desert. The Indigenous communities in the Atacama Desert – collectively known as the Atacameños or Likan Antai – have practiced astronomy since long before the region became a popular hub for modern telescopes (Detwiler, 2021; Storey-Fisher, 2019). Chajnantor is also known in the Kunza language of Likan Antai collectives as Thaknatur, or ‘the place of departure/flight’ (Detwiler, 2021), and is involved in an Andean ecology defined partially by the ring of stratovolcanos that surrounds ALMA. These ecologies are organized through Likan Antai relations of ayllu, or ‘collectives of human and non-human actors that include plants, animals, the elements, landscape features, and people. Here, mountains are not only mountains but are agencies in ayllu kinship networks’ (Detwiler, 2021; see also Herrera, 2019, for further discussions of thinking through ayllu frameworks). This is a paradigm that is perhaps similar to how data assemblages are coming to be understood, where components are not individual entities that layer onto each other in an additive way but, instead, are inextricably entangled.

Beyond ALMA, other foreign radio astronomy projects, including ones from Europe and the United States, have been operating in the region since the 1960s, most developing with minimal Indigenous participation and some with active disengagement from local populations. ALMA itself started its operations in 2011 as an international partnership; unlike previous data facilities in the region, the ALMA project included the involvement of Indigenous people at various stages, including negotiations on how archaeological sites and the high-altitude ecosystems in the area could be preserved and investments in the local community, such as involving local students in research projects and projects to preserve Likan Antai knowledge. Despite this, various issues remain, including the power asymmetries between local populations and the predominantly Western actors in the region. In particular, Detwiler (2021) argues that while ALMA sponsors what they call ‘ethno-astronomy’ projects with the Likan Antai community, these efforts are ultimately used to substantiate the observatory’s claim that Likan Antai cosmology is continuous with optical/radio astronomy projects – that these projects are essentially the same. In flattening the differences between Indigenous astronomy and optical/radio astronomy, Indigenous claims to ayllu hybrid kinship are undercut and nature is separated from the wider ecologies of human and non-human actors. Detwiler contends that for the radio astronomers, the ‘Chilean sky is cultivated as a pristine, media-rich ecology and is simultaneously produced as a technical environment to be measured, shaped, and optimized via the regulation of light pollution’ (Detwiler, 2021). The assertion that Indigenous Andean cosmovision is the same as radio astronomy is a way to justify the datafication of the sky and to accrue legitimacy through multicultural respect, ‘while also [being] a way of garnering innocence to its spectral and territorial occupation’ (Detwiler, 2021).

As seen here, separating infrastructures into their physical and more amorphous parts involves a false binary, as they bleed into each other. Illustrated here are also some of the ways that colonial logics, projects, and futurities are embedded and entangled in volumetric and elemental complexities. The optical and radio astronomy data factories here are at once taking up physical space on the land but also taking up space through means such as spectral occupation, which perpetuates a type of settler futurity in which the ‘calculable sky’ is platformed above all. The elemental specificities, including the thin air at high-altitudes, the harsh wind, and the cloudless skies, are not just resources to be extracted but interplays with a variety of processes that affect embodied experiences including labour, cultural practices, and Indigenous astronomy. To take seriously the discontinuities between Indigenous and optical/radio astronomy in the Atacama Desert, then, involves disentangling the elemental and volumetric epistemologies underpinning these projects and futurities.

6. DATA CENTRES IN OUTER SPACE

Satellites serve a wide array of functions. Rather than transferring data through physical cables, satellites beam information through the vacuum of space, where they are intertwined with elements within and outside of the Earth’s atmosphere. The latter being a place that makes explicit how our Earth is a sphere suspended in space, where there is no objective ‘up’, ‘down’ or ‘centre’, and where the horizontal/vertical binary starts to break down further. Like other geographies in which dataspaces are embedded, there are ongoing claims in extra-terrestrial space. This claiming of outer space, Smiles (2020) argues, is a continuation of the ongoing settler colonialism on Earth (see also Gorman, 2005; Klinger, 2018; Maile, 2018).

According to the United Nations Office for Outer Space Affairs, there are currently 7500 active satellites as of late 2021, with many more inactive satellites still in orbit. A confluence of factors has led to this increase, including the growing demand and dependence on satellite data and connectivity by a number of industries and governments. The development of even lighter and even smaller satellites has also facilitated more economically viable satellite projects. A total of 94% of all spacecraft launched in 2020 were ‘smallsats’, satellites that weigh fewer than 600 kg (Chakrabarti, 2021). The list of actors who are able to make a physical claim to space is highly stratified. The countries that take the lead in terms of number of active satellites are the United States with 1897 satellites, China with 412 satellites and Russia with 176 satellites. While more than 105 countries have at least one satellite in space, some countries have never launched their own satellites (Salas, n.d.). In contrast, the US-based company SpaceX has launched over 1700 satellites as of 2021, and OneWeb, a company that is partially UK government-owned, has around 352 satellites in orbit. Both companies aim to launch more than 40,000 satellites between them in the coming years (Chakrabarti, 2021).

Data centres in space are not necessarily new. If data centres are defined as facilities that host servers and/or computer hard drives, many mini-data centres are already in space. Satellites are often built to be multifunctional; this can include acting as small-scale data centres, especially if a satellite has a central processing unit (CPU) and can process, store, transmit and receive data. However, what is relatively new is the development of more dedicated and larger scale data centre facilities for outer space. This can, for instance, look more like the distributed system of the Internet of Things (a network of physical objects that are embedded with sensors, software, and other technologies for the purpose of collecting and sharing data), where mini-data centres are integrated into the ecosystem of devices and man-made objects in space (Mohney, 2020). Others are more closely aligned with the idea of edge computing in space, such as with US firm OrbitsEdge, which aims to offer services from a proprietary ‘satellite bus’ that contains server racks bolted to the bus floor. Here, high-performing computing infrastructure would be launched into orbit to collect, clean and aggregate data from satellites, where the satellite bus would be in much closer proximity than land-based data centres. Similarly, Japanese firm Nippon Telegraph and Telephone is planning to launch a data centre into space as early as 2025, where it aims to locally process satellite data in order to only beam select, useful information down to Earth, which would reduce the time and cost spent in transferring large amounts of data (Menear, 2021; NTT Press Release, 2021).

The hardware considerations of these data projects in outer space must account for the elemental environment. For instance, the satellite bus makes use of solar panels to gather energy from the sun but also batteries to keep systems running when the bus is in the shadow of the planet. Likewise, systems and servers must be able to work in the vacuum of space, they must be protected from cosmic radiation and other space events.

Other projects include the ambitious space-based content-delivery network (CDN) patented by Amazon in 2019 (Amazon, 2019). The patent outlines the plans for a geographically distributed network of data centres in an extra-territorial environment and, eventually, a lunar data centre as part of the Amazon Web Services to be located in Mare Tranquillitatis, a basin on the Moon. The patent states that the satellite-based CDN aims to support a range of applications including voice calls and streaming services, but also time-sensitive applications that require near real-time responses ‘such as professional gaming in selected areas covered by the high-altitude airships’ (Amazon, 2019). Thales, a French company, is also conducting feasibility studies to support human living quarters and computing infrastructures on the Moon as part of a collaboration with the Italian Space Agency (ISA) and National Aeronautics and Space Administration (NASA); it aims to establish a permanent lunar base as early as 2024 (Menear, 2021; Moss, 2021).

Amongst these speculative data projects, several intersecting grand narratives about technological infrastructures in space are noticeable. This includes, first, considerations of outer space as an egalitarian and placeless space, which ties into, second, claiming outer space volumes through logics of terra nullius. With regard to the former, for many businesses, data centres outside of Earth offer an attractive investment prospect with free cooling from the frigid vacuum of space and accessible renewable energy from the sun as well as being in a territory that is away from most human interference and has relatively nascent regulation along many axes. For instance, ConnectX – a start-up that plans to launch a satellite constellation for data storage that targets crypto-currency and stores digital wallets off-planet – notes:

space gives ConnectX an unfair advantage. Economically, we do not have to pay for real estate, electricity, cooling, staff or security. From a security perspective, no one can physically access our system and no government or entity can force the exposure of your information. (ConnectX Company Statement, quoted in Donoghue, 2018)

The sentiments expressed by ConnectX perhaps represent the more extreme end of utopian ideals of data centres in space. Nevertheless, the idea that outer space is a blank slate with no implications for human labour or environmental impacts is a continuation of the colonial discourses on Earth where realities of digital infrastructures are flattened and equalized. In the larger picture of the privatization of space, especially by large technology companies, Cath and Lewis (2021) tease out how the rhetoric used to justify these projects mirrors the discourse of the internet being a placeless and egalitarian space, where everyone has equal participation. The implication of this being that regulation would only restrict the possibilities and progress of this supposedly new frontier.

Indeed, the regulations and international agreements pertaining to space exploration appear to support the claiming of territory in space along multiple axes, which a more expansive view of the volume helps to reveal. This includes the Outer Space Treaty of 1967 by the United Nations Office for Outer Space Affairs, which states that the exploration and use of space should be open to all of humankind (United Nations Office for Outer Space Affairs, 1967). Its related committees coordinate and demarcate how space is used.

There are multiple dimensions to the claiming of physical and elemental territory in outer space. Here, as Smiles (2020) argues, fantasies of terra nullius frame how space represents yet another frontier to be claimed both for the possibility of further settlement in space and also for the preservation of the existing status quo on Earth. The more traditional considerations in terms of occupying space include the way that the physical territory of outer space is divided and distributed. For instance, there is a limit to physical airspace, particularly in the high altitudes of Earth and in near-Earth orbits. The claiming of these territories is coordinated by various international bodies, which often operate on a much disputed ‘first arrival’ basis (Purity, 2020; Rothblatt, 1982; Thompson, 1996) that is reminiscent of settler-colonial logics.

Beyond the claiming of physical space, satellite constellations also claim space in the radio-wave spectrum landscape. Wireless data are transmitted by sending signals over certain frequencies, where limited frequencies are available. If two signals occupy the exact same frequency, they may interfere with each other; thus, frequencies have been treated as a limited resource in which equitable access must, in principle, be protected. The allocation of orbits and spectrum space are both coordinated by a specialized United Nations agency, the International Telecommunication Union (ITU), which is responsible for information and communication technology matters (Boley & Byers, 2021; Voelsen, 2021). Both orbital positions and frequency plans operate largely on the basis of allocating a frequency or orbital path to whichever eligible actor is first to put in a request. In theory, the first occupant does not own these spaces but instead occupies them for a certain amount of time. Furthermore, certain ‘slots’ are reserved for countries with less developed space capabilities. Nevertheless, this allocation system still perpetuates a particular settler futurity where space is and will continue to be divided and claimed by nation-states (more specifically, ITU member countries). This is a futurity that perhaps appears counter to the spirit of the Outer Space Treaty of 1967.

Finally, these settler-colonial logics persist into the afterlife of space expansion projects, where data centres will inevitably incur pollution. The current spacecraft afterlife care processes fit what Liboiron (2021) calls the dominant paradigms of today’s pollution science, the ‘permission-to-pollute’ system that is built on top of colonial land relations, where land is seen as a resource where pollution is allowed to happen – where the right to waste is parcelled out to certain actors who are given permission. This involves managing rather than eliminating pollution and assuming an entitlement to land, rivers, atmospheres and bodies for settler-colonial goals. As such, it not only paves the way for a settler-colonial futurity that eliminates other types of possible futures and relations with the environment but involves an ad hoc approach to dealing with the myriad of possible pollution legacies rather than refusing them in the first place. This continues in space.

For instance, satellite constellation projects have raised concerns over light and sound pollution from the bright light and noise they emit when in orbit – which particularly affects astronomy, scientific, and cultural practices that require clear, dark skies and radio silence (Chakrabarti, 2021; Clery, 2020; Hecht, 2021). Likewise, decommissioned satellites that are left out in orbit as space junk add to the physically congested volumes of space, where there are risks of collision events with the potential to trigger unpredictable chain reactions (Boley & Byers, 2021; Le May et al., 2018; Mann & Pultarova, 2022). While retired satellites can also be ‘designed for demise’, where after their service, they burn up upon re-entering Earth, the environmental considerations include how materials that get burnt up can produce fine aluminium particulates, constituting a form of anthropogenic deposition of aluminium in the atmosphere – something that has been a controversial proposal for geoengineering a climate change intervention (Boley & Byers, 2021).

Space debris can also be disposed of through ‘controlled re-entry’ back into Earth, where risks include potential human casualties,¹ property damage and discard pollution. Currently, the most common practice is for space debris to fall within the South Pacific Ocean Uninhabited Area, the largest unpopulated ocean space that is beyond the legal jurisdiction of any country, into which more than 260 satellites have fallen. This point, also known as Point Nemo (Latin, ‘no one’), stretches over 22 million km² and represents yet another a ‘blank slate’ territory. It is the location in the ocean that is farthest from land, about 2688 km from the nearest land and where the closest people to it at any given point are likely to be along the vertical axis – that is, the astronauts in the International Space Station (ISS) that is orbiting at 400 km above Earth (Sharman, 2022). NASA states:

because of its distance from populated areas, this location is traditionally used for spacecraft disposal and represents an excellent target for ensuring any hardware that survives the heat of re-entry does not land near populated areas on Earth.  …  Once the debris enters the ocean, it would be expected to settle to the ocean floor. Based on the findings of the International Space Station Environmental Impact Statement, no substantial long-term impacts would be expected. (FAQ: The International Space Station 2022 Transition Plan, 2022)

Perhaps a more expansive understanding of earthly volumes would attend to implications of pollution beyond fatalities when, for instance, the 420-tonne ISS (itself a structure fitted with about 100 laptops) is planned to be laid to rest at Point Nemo in 2024.

As Liboiron (2021) argues, in many cases pollution is colonialism, where modern environmental pollution perpetuates ongoing colonial relations to land. This permission-to-pollute and use of harm mitigating, rather than eliminating, assumes a specific settler-colonial futurity rather than the many alternatives that are possible. Responding to the visual pollution of satellite constellations, Venkatesan et al. (2020) argue that the sky should be treated as an ‘ancestral global common’ that contains the heritage and future of humanities’ scientific and cultural practices. They argue that the systemic inequity in the terrestrial world also underlies how ‘space is literally being occupied, with little coordinated international regulation, discussion of ethical considerations, or consensus-building from a shared vision for the future of humanity in space. The rush to claim near-Earth space is also leaving out the world’s most minoritized communities, including Indigenous peoples, who need to be involved as stakeholders in decision-making (p. 1043). They warn that the increasing privatization of near-Earth space risks repeating colonization on a cosmic scale, with effects on the natural sciences and space exploration but also on the human right to dark skies and the rights to cultural traditions of communities across the world, such as the Likan Antai in the Atacama Desert. Similarly, with regard to the sea, while the dumping ground of Point Nemo is lawful under space law, De Lucia and Iavicoli (2018) argue that these processes use the ocean with little consideration of the potentially harmful consequences for marine ecosystems or for the oceans as a global commons. This includes the wider potential problems from leakage of toxic rocket propellant (hydrazine) and the growth of the debris pile of retired satellites. Similarly, potential impacts on socio-cultural life and influence on the ocean as a global commons is little discussed (De Lucia & Iavicoli, 2018). Risk calculus, then, tends to be based on very specific metrics of impact: not endangering human life and not endangering private property. Proposing a different approach, Liboiron (2021) urges us to reconsider how pollution is measured, arguing that pollution is best understood as the violence of colonial land relations rather than environmental damage and harm – land (and space) relations that can be reproduced even through well-intention science and activism.

From the speculation stages to the afterlife of space exploration, settler-colonial logics persist in the form of a sense of entitlement to the physical space of near-Earth orbit and the Moon as well as in the claiming of the sonic landscape, atmospheric landscape and oceanic landscapes. While the volumetric spaces implicated here tend to resist any easy borders or jurisdictional claims, there is good reason for them to be considered as global commons: resources and natural spheres that are to be shared by all rather than the few. Much of these data storage and architecture projects in space are still at proof-of-concept and nascent stages, but they already exemplify some of the potentially perilous pitfalls left in the wake of the race to space. A more expansive understanding of the elemental and volumetric dimensions of such space expansions might contribute to a more comprehensive understanding of their risks, inequities, and enacted violence. There are still many unknowns about future developments beyond Earth, but it is important to extend foresight over these issues. For instance, instead of dismissing the need for manual labour in space, we could attend more carefully to ongoing research about how space travel affects the human body. Many astronauts returning to Earth after long periods of space flight suffer from blurred vision from flattened eyeballs (a condition called visual impairment intracranial pressure) (Alperin et al., 2017), which has implications for space maintenance work. Likewise, instead of defaulting to ‘managing’ pollution, there could be a more concerted effort in refusing certain options and relationships with nature.

7. CONCLUSIONS AND ALTERNATIVE IMAGINARIES OF THE VOLUME

This paper has sought to join other scholarship in using elemental volumes of air, space, and the ocean as environments to unsettle notions of ‘surface’ and to explore the different means through which colonialist logics are deployed in digital volumes. In using vignettes that traverse our planet and beyond, it has explored some of the ways digital elemental volumes are conceived, claimed, sensed, and embodied.

The first vignette demonstrated the ways in which data centres in the terrestrial world are entangled with the elemental, where data infrastructures cannot easily be separated into their physical and more amorphous elements. Here, colonial logics can take the form of taking up physical territory in key locations and spectral occupation and also through perpetuating a type of colonial futurity where one type of relationship with nature is established above all. Moreover, the different elemental considerations of thin mountainous air, harsh winds, and cloudless night skies mediate computing processes, labour, and cultural practices. The second vignette illustrated some of the ways that colonial logics extend into outer space, from speculative projects of data centres in space to the afterlives of these endeavours. This includes again the physical claiming of territory such as the Moon, which is justified by the terra nullius myth (‘no one’s’ land) or what can be called the Point Nemo myth (‘no one’s’ ocean). It also includes the claiming of the radio-wave and oceanic landscapes as well as schematics such as the ‘first come, first served’ principle and the framing of pollution strategies as ‘managing’ rather than eliminating waste, which ensures the continuity of a certain colonial futurity. Likewise, the different elemental considerations of the cold vacuum of space, solar power, the deep sea, the Earth’s atmosphere, and the other elemental specificities of outer space mediate plans for computing infrastructures, the claiming of resources, risk calculations, and how pollution unfolds.

In foregrounding the volumetric, elemental considerations in both vignettes, various processes are highlighted:

This includes first the obvious and non-obvious ways that colonial logics are extended in volumetric and elementally specific ways. As discussed, there is the more obvious claiming of land or key geographical areas for the purpose of satisfying the data needs of specific actors, but occupation can also take the form of claiming a particular way of harvesting the sky, claiming particular radio-wave spectrums that are invisible to human eyes, the entitlement over what are framed as ‘empty’ oceans and celestial spaces, and the authoritative allocation of orbital paths.

Second, how elemental volumes are differently embodied, sensed, and understood, where the flattening of this reality can perpetuate colonial logics. Emphasizing elemental complexities and how they constitute different lived realities helps to draw attention to the way in which these systems affect processes such as labour, maintenance, and pollution – including how thin air affects workers, how aluminium pollution in the atmosphere affects skies around the world, and how the cold vacuum of space shapes dreams of digital expansionism. The opposite, which is the flattening and equalizing of discourse, can help pave the way for colonial futurities, whether it is the flattening of labour realities and Indigenous epistemologies, or myths about outer space being inherently egalitarian.

Third, how colonial futurities are embedded in a variety of projects, even well-meaning ones. Hogan (2018) warns that the growing size of data centres (pointing to the example of a data centre in Norway which is planned to be the size of 84 football fields) is also an extension of Western colonial thinking. In expanding the analytical scope to the diverse elemental volumes of data centres, it becomes evident that these expansionist trajectories Hogan highlights also extend beyond claiming large swaths of land and water. The settler futurity involved in reserving land for data needs also creeps into very particular ways of claiming and dividing socially, scientifically and culturally significant mountainous landscapes, the visible sky, outer space, the oceans, the atmosphere, sonic landscapes and more.

Many of the data developments in the extreme vertical and volumetric ends of the world – including underground, in the deep sea, at high altitude and in outer space – require extremely specialized labour and equipment and are incredibly expensive projects to start and maintain. The rush to claim these spaces has sidelined many communities, especially marginalized and under-resourced populations. For data centre infrastructures in particular, there will likely be a continued drive to find more free cooling, from the frigid air of high-altitude regions to the chilly depths of subterranean worlds. As developers search for colder homes, data infrastructures will continue to creep into the highest and lowest places of Earth and beyond, bringing forth myriad considerations in these localities. As such, it becomes more important to deepen the understanding of colonialist logics and futurities with regard to digital volumes. Here, it is important to examine speculative digital infrastructures as well as in-progress initiatives to avoid producing only after-the-fact assessments of harm and violence and to exercise more reflexivity about the futures we are signing humanity up to. This line of thought supports the proactive rather than reactive approach to technology that others have argued is necessary for innovation that truly takes into account ethics and minoritized populations (Abdilla et al., 2020; Mohamed et al., 2020).

By taking a more proactive approach in attending to how settler and other futurities are being embedded in our collective schematics for the future, we can more lucidly assess the futurities that humanity is being collectively signed up for. These include a future where the universe is claimable; a future of geoengineered skies; and a future where global commons, such as the Moon, are privately apportioned.

Finally, these lenses attend to alternative worldviews and imaginaries of the elemental volume. Beyond understanding the colonial claiming of digital volumes, how can it be resisted? A more dedicated answer to this important question is beyond the scope of this article, but some important points are worth highlighting from the vignettes outlined here. Jackman and Squire (2021) raise the question of how we can move away from militaristic and state-centric ideas of understanding, living, and embodying the volume. This endeavour of diversifying epistemologies, understandings, and experiences of space is not only important as a way of broadening knowledge but as a way of refusing corporate and settler-colonial trajectories towards dataspaces. The vignettes presented here show how the effects of dataspaces have tended to be measured in a piecemeal way and with a specific idea of harm by various authorities. However, they also showcase the refusal of this by other actors in the ecosystem. What would it be like to understand the volumetric night sky not as calculable, claimed, and divisible but instead more through Likan Antai understandings of ayllu networks? What would it mean to heed Venkatesan et al.’s (2020) call to preserve the night sky as a shared and sacred space and resource and, similarly, De Lucia and Iavicoli’s (2018) and many other community’s call to attend to the ocean as a global commons? What does it mean to take into account the elemental, more than human, more than Western issues of volumes – including a less ‘metric’ centred idea of the volume and a more ‘embodied’ one that attends to the messiness of our elementally entangled spaces? These remain open questions, but one thing is certain: it is becoming increasingly pressing to take the post- and de-colonial perspectives of data and space seriously, with a view to imagining alternative futurities that more comprehensively attend to Earth’s plural communities.

ACKNOWLEDGEMENTS

The author gives their deep thanks to Srujana Katta, Nancy Salem, the special issue editors Dr Mia Bennett and Dr Klaus Dodds, Dr Sneha Krishnan, and the anonymous reviewers of this paper for their kind and thoughtful feedback throughout this process.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author.

Additional information

Funding

No funding was received for this paper, but the author’s doctoral work was generously supported by the Clarendon Scholarship, University of Oxford.

Notes

1 NASA recommends such a risk be below 1:10,000.