Unsettling the Cloud: On Data Centers and Counter-Narratives

As I gaze out of my window, I am met with a totem. This totem is gray and windowless, nestled in between offices and academic buildings. Behind it is a park, and the longer I stare, the deeper it becomes embedded in the natural landscape, after a bit I forget it’s there. But in the corner of my eye I can see another one; another totem. This one intimidates me with its red glow.  

These buildings came to serve as mystical pillars of data flows to me, they became sites of reification, sites where the cloud finally condensed and data rained down. They assumed a posthuman status; high-tech facilities where humans are only needed to keep other humans out.

 I always imagined data as something abstract, as a floating entity, but as my encounters with these pillars started a process of materialization, it simultaneously sparked a desire to interrogate and to demystify.

Luke van Wijk, 2023

Data centers and information infrastructure present the complex materialism of what is imagined as a dematerialized industry. The data industry is comprised of a myriad of large corporations that mainly rent out servers for data processing and storage. They sell the dream of the digital, which is freed from the constraints of material capitalism and can therefore grow infinitely. In dematerialized finance, the physical is omitted and an abstract idea of the cloud takes form. The cloud is the dominant imaginary that constitutes a large part of our understanding of data. Nonetheless the physical remains. Not only does it remain, it requires energy and labor, it is privately owned, it is opaque; often hidden in plain sight.

Acid Clouds is a book by Niels Schrader and Jorinde Seijdel in collaboration with photographer Roel Backaert wherein Dutch data centers have been systematically photographed and catalogued. The book, which will be published in April 2024 by NAI010  is described as ‘a research project initiated by graphic designer Niels Schrader and photographer Roel Backaert that maps the hidden cloud infrastructure in the Netherlands’[1]. These photographs are paired with critical essays and other relevant information to form a critical intervention in the data industry and the ubiquity of data infrastructure. In the book, they have attempted to establish a link between dominant narratives and their material realities. It also seeks to expose the entanglement of colonialism and capitalism as it exists in the data industry. Readers not only become familiar with the infrastructure that facilitates the flow of data but gain a sort of material literacy of these spaces, allowing them to recognize and critically reflect on them. Nonetheless, the representational logics and imaginaries of the data industry continue to dominate the collective understanding (or lack thereof) of these spaces, which is why works like Acid Clouds are so important. In my analysis of the data industry and of the book, I was informed by conversations both with Niels Schrader and with a former security employee of a prominent Dutch data center.

Luke van Wijk, 2023

In the current neoliberal paradigm, cognitive labor and dematerialized finance are defining features of the socioeconomic landscape. Precarity and abstracted forms of labor dominate the mind of the contemporary proletariat. We live and work in the digital; our labor and the cloud are deeply entangled. As was aptly stated by Franco ‘Bifo’ Berardi in Heroes: Mass Murder and Suicide: “As we move into the age of info-labour, there is no longer a need to invest in the availability of a person for eight hours a day throughout the duration of his or her life. Capital no longer recruits people but buys packets of time, separated from their interchangeable and occasional bearers. In the internet economy, flexibility has evolved into a form of fractalization of work.”[2]. My generation was born into this paradigm, and therefore largely accepts the collective imaginary of the cloud and the structures of dematerialized finance as a given. Nonetheless, physical infrastructure is required for this digital laborscape to function. This infrastructure remains largely out of sight and therefore unacknowledged, meanwhile, the companies that own and maintain the infrastructure are moving freely and uncriticized as they continue to present narratives of a green and immaterial digital world; the lack of visibility allows them to govern the facilities without scrutiny. These privately owned spaces facilitate the flow and storage of data through thousands of servers, using significant amounts of power. Furthermore, the internet as well as fundamental government services rely on these facilities to function, granting the companies that own them a consequential amount of political leverage and power. As I explored these spaces and their materiality, several geopolitical and environmental complexities started to emerge.

I lived in Amsterdam Science Park for three years, sharing my surroundings with two major data centers. These windowless buildings left a deep impression on me and were an important point of departure for my research. They are owned by big players in the data industry: the Data Tower (AMS9) is owned by Digital Realty, which is an American company specializing in data infrastructure. The other center (AM3) is owned by Equinix, also American. Amsterdam Science Park also houses the Nikhef, an academic organization that specializes in data and internet technologies. Science Park is therefore a significant site of data technology and infrastructure. Both data centers are colocation or multi-tenant spaces, meaning that they rent out servers and data storage to independent clients. These include both commercial clients and government services. They simply provide the space and technology to facilitate internet and data services. This contrasts with single-tenant data centers that house servers used by a single company, enterprises like Google, Amazon, and Facebook come to mind here.

On their website, Digital Realty emphasizes the technological capabilities and reliability of AMS9, as well as mentioning that it runs on “100% renewable energy” and serves as a site for “establishing a European presence and global expansion”[3]. Similarly, Equinix describes their Amsterdam data centers as “energy efficient” and as providing “ample capacity for growing businesses and [enabling] customers to access rich industry ecosystems”[4]. Through these descriptions, the dominant narratives and representational strategies of data centers start to emerge. These narratives are clearly embedded with ideological coding.

The examples from the respective websites hint at the vernacular practices of these companies, but their visual strategies are also relevant to their representational logics. The ideologically layered visual output of these companies presents them as posthuman and entirely technified[5]. This is not actually the case, as these spaces require different kinds of labor to function. This contradictory visual strategy exposes the ideological coding that takes place in their self-representation: data centers are supposed to be imagined as sleek, automated spaces that imply a similar mode of operation as ‘smart’ technology. In reality, these spaces are more like data factories, requiring consistent labor and human intervention to remain operational. The narratives that are ubiquitous across the data industry became glaringly obvious as I started researching them: sustainability, green energy, the cloud, and infinite growth. The contradiction between these narratives and the material realities of these spaces is not at all hard to grasp, nonetheless, our collective imagination of the data industry remains entrenched in narratives put forth by data companies, exposing the need for critical intervention.

In my conversation with a former security employee at a Dutch colocation data center, it was revealed to me that the servers housed in the center not only required frequent maintenance but that the labourers conducting this maintenance were outsourced by the companies that rent the server space. These external labourers work on a freelance basis and are often from India or Eastern Europe. They were described as working under precarious freelancer conditions, arriving at one data center after having just been at another. They would then work for several hours in the cages that house the servers, often in high temperatures due to the heat produced by the servers. Furthermore, the server cages were not maintained by the data center, but rather by the companies that rented them, often resulting in stacks of hard drives and messy cables piling up in them, clearly countering the constructed imagery of the data industry. The conditions of the workers as well as the apparent state of the server cages contrast heavily with the available images of data centers, this contradiction became emblematic of the narratives that are upheld within the data economy. A general feeling of contradiction runs deep in the data industry, as is the case for most industries, but the way that the material realities of data flows are lurking out of sight of our perceptions feels particularly sinister.

Luke van Wijk, 2023

According to Jennifer Holt and Patrick Vonderau, the visual strategies of the data industry have significant ideological implications, they state that “Infrastructural politics [are] not just about what is deliberately hidden from sight or is invisible; [they are] equally about the hypervisibility created around some of an infrastructure’s component parts, all while most of the relations it engenders and the rationality embodied in its overall system sink deeply in obscurity.”[6]. These logics underpin a general strategy of concealment, which makes data infrastructure notoriously difficult to research and scrutinize. In fact, when researching one of the Science Park datacenters on-site, I spoke to a person living across from the building while he was smoking outside of his front door; he told me that he had no idea what the building was even though he saw it every time he left his house. Obscurity is therefore really the perfect word here, and it implies a subsequent need for clarity. This clarity can be achieved through critical counternarratives.

The former employee I spoke to also indicated that one of the floors of the data center was dedicated to an important government department, and (armed) government representatives would frequently travel to the data center with physical hard drives in order to store them on-site. The fact that government data is so sensitive that it can’t be digitally transferred to the servers in fear of interception is stored at a privately owned facility should raise red flags. This dependency on privately owned infrastructure for fundamental bureaucratic services further complexifies the data center as a geopolitically charged space. I found the geopolitical leverage that these companies possess extremely worrying. They came to represent the deterritorialized neoliberal entities that dominate the global financial landscape, however, unlike many other big corporations, they remain largely unscrutinized. The lack of scrutiny is due to a lack of visibility of these spaces and the imaginaries that come to represent them. The more I looked into these spaces, the more concrete their narratives became, alongside this concretization the urgency for critical counternarratives also became increasingly evident.

The established representational strategies that have been mentioned rely heavily on the obfuscation of the material and on a rhetoric of the digital that omits physical consequences. The cloud symbolizes this rhetoric. The metaphor of the cloud is a dominant collective imaginary that serves as “[…] a marketing concept that renders the physical, infrastructural realities of remote data storage into a palatable abstraction for those who are using it, consciously or not.”[7]. In my conversation with Niels, he emphasized that this imaginary is most likely derived from geek culture, and doesn’t constitute a malicious media strategy that has been manufactured by the tech industry to mitigate critical engagement with its infrastructure. Nonetheless, data companies benefit from this imaginary, as it essentially renders data immaterial. This imaginary obfuscates the link between the data industry and its physical infrastructure through the abstraction of the metaphor. The imaginary of the cloud is part of a larger narrative of endless growth that stems from a deterritorialized and digitized neoliberal landscape. If data exists only as an abstract, cloud-like entity, then its industry can grow exponentially. This narrative presents the digital as an ever-expanding avenue with seemingly no physical constraints, as was seen in the aforementioned examples taken from the websites of Digital Realty and Equinix. Words like ‘growth’ and ‘expansion’ are ubiquitous in the vernacular practices of these companies.

The environmental impacts of the data industry also become largely invisible through the narratives that naturalize data. The visual representation of data centers often depicts them as embedded in the natural landscape, supplemented by statistics about sustainable energy practices. This strategy obviously serves to divert attention away from the massive amounts of energy consumed by these facilities. I found it noteworthy that the way these companies represent themselves is fundamentally contradictory: on the one hand, they boast about the technological capabilities of their facilities on their website, this information is clearly meant for investors and businesses who are looking to rent server space. On the other hand, they put forth an imaginary of data as something green and immaterial, something that can grow endlessly. This is not only relevant to businesses that seek exponential growth as it promises endless financial growth as well, but also to the general public, who are led to believe that data is something abstract, non-physical, and non-threatening to the environment.

In conversation with Niels, it became clear that he identifies the infinite-growth narrative and its subsequent implications as the central problem and point of critique for data infrastructure. The narrative, which depends heavily on principles like Moore’s Law (which implies that technological innovation leads to exponential growth), is deeply embedded with the ideological underpinnings of Silicon Valley. This ideology not only presents the tech industry as environmentally friendly or neutral through greenwashing practices but also centers white, male voices. Moreover, the practices of big tech companies are inherently colonial and imperial due to their extractivist business models. This also exposes the general entanglement between colonialism and capitalism, as data is not only exploitatively extracted but territory upon which the infrastructure can be built, as well as resources that are required to keep it operational, are imperative to the services they offer. The data industry, which self-represents as immaterial, is actually engaged in extremely territorial and physical practices. Niels also mentioned that the speed of data technology and the lack of tech literacy, even among experts, abstractify the understanding of the limits and constraints of this ‘infinite’ growth narrative. The material consequences are omitted for the sake of a neoliberal imaginary where the data industry can endlessly expand. Of course, the link between the imaginary and the material still exists and needs to be interrogated and scrutinized. This is what Acid Clouds seeks to do: by visualizing and cataloguing the physical infrastructure that maintains these narratives, the link between the two is not only established but can then be criticized. Furthermore, Niels emphasizes a necessity for alternative narratives and critical voices to destabilize dominant narratives, something which Acid Clouds also aims for through the inclusion of critical essays in the project. This kind of critical intervention is necessary because it serves to visualize the “[…] decontextualized technologies, [which remain] largely immaterial, dimensionless, and almost impossible to even imagine”[8].

Acid Clouds is one example of how critical interventions can unsettle dominant narratives and stimulate critical engagement. The practice of investigative aesthetics argues that aesthetics can serve to redistribute sensibilities and ways of knowing the world, doing so according to principles of making knowledge public[9]. There are several examples of artworks that seek to make visible the infrastructure that our data flows through. A well-known example would be Landing Sites (2016) and Undersea Cables (2016) by Trevor Paglen, wherein he attempts to visualize the underground or undersea cables that facilitate the flow of information and are essential infrastructure for the internet. Similarly, Eva and Franco Mattes have several works in which they install yellow cable trays in institutional contexts to materialize the flow of data and data processes that are usually immaterial. These works invite spectators to critically engage with the materiality of data, more broadly they make spectators aware of the existence of the infrastructure in the first place, destabilizing the notion of a dematerialized digital landscape and presenting an alternative visual narrative.

Acid Clouds seeks to establish the link between growth narratives and the material infrastructure of the data center. By systematically photographing and cataloguing Dutch data centers, Niels and Roel visualized and concretized the material implications of the cloud and the growth narratives they rely on, providing a counternarrative that unsettles the notions of a dematerialized digital landscape. The lack of alternative visualizations was the original motivation for the Acid Clouds project. Niels states that the objective of the book is not to antagonize companies or data centers, but to provide an accessible and critical counternarrative. Furthermore, Niels refers to the process of the book being one of visual journalism, rather than an aesthetic practice. The book brings attention to the ability of aesthetics to engage in storytelling, particularly in the information age, where individuals are subjected to a myriad of contradictory narratives at all times, which often leads to conspiratorial ways of thinking. This is because the conspiracy theory often simplifies and concretizes a narrative, which allows an individual to anchor themselves in the whirlwind of information. This is obviously problematic because it prevents critical engagement and preconfigures a narrative, thereby negating the critical capabilities of the individual. To Niels, establishing visual narratives through artistic practices has the potential to undermine the conspiratorial tendencies of the information age, as well as democratise access to these narratives and therefore unsettling hegemonic narratives.

The book is structured around the images of the data centers, which have been systematically photographed in order to catalogue the material pillars of the data industry in the Netherlands. This visual aspect of the book is supported by various essays that offer critical analyses of data infrastructure. By presenting the data centers in the format of a catalogue, a network of data infrastructure starts to form. By concretizing this network, the reader is invited to critically reflect on the physical infrastructure that upholds the data industry, therefore disrupting the imaginary of the cloud and immaterial data. Furthermore, readers gain a sort of aesthetic literacy, through which they can recognize and decode sites of data infrastructure, as these sites and their functions often remain hidden in obscurity. Especially when presented as a series, the data centers take on an almost monolithic appearance, and the photographs being taken at night grants them an eeriness that contrasts with the conventional visualizations of these spaces.

Niels Schrader & Roel Backaert, 2023

Niels Schrader & Roel Backaert, 2023

Niels Schrader & Roel Backaert, 2023

Niels Schrader & Roel Backaert, 2023

Throughout this project, the necessity for critical interventions in the data industry became more and more urgent to me. The data industry relies heavily on its constructed imaginaires and narratives, omitting the material. Critical interventions that destabilize and scrutinize these imaginaries by providing counternarratives pose a necessary form of resistance. Green and dematerialized narratives still dominate the collective perception of tech companies, this is why works like Acid Clouds among other artistic interventions are so significant. These works attempt to disrupt those narratives and present alternative ways of knowing and sense-making. Due to the blackboxed and opaque nature of data centers, alternative visualizations are necessary in this disruption. When combined with critical voices that are external to the data industry, works like Acid Clouds have the potential to meaningfully shift the collective understanding of data infrastructure.

Luke van Wijk, 2023

Special thanks to Niels Schrader for taking the time to discuss the project with me and to Luke van Wijk for assisting with the visual material.

More information about the Acid Clouds project can be found here: https://acidclouds.org/

Notes:

[1] “Acid Clouds.” Acid Clouds, acidclouds.org/.

[2] Berardi, Franco. Heroes: Mass Murder and Suicide. Verso, 2015.

[3] “AMS9 Data Center: Science Park 121: Digital Realty.” AMS9 Data Center | Science Park 121 | Digital Realty, www.digitalrealty.com/data-centers/emea/amsterdam/ams9.

[4] “Amsterdam Data Centers.” Equinix, 19 June 2023, www.equinix.com/data-centers/europe-colocation/netherlands-colocation/amsterdam-data-centers.

[5] Taylor, A.R.E. “The Data Center as Technological Wilderness.” Culture Machine, vol. 18, 2019.

[6] Holt, Jennifer, and Patrick Vonderau. “‘Where the internet lives.’” University of Illinois Press, 2017.

[7] Holt, Jennifer, and Patrick Vonderau. “‘Where the internet lives.’” University of Illinois Press, 2017.

[8] Holt, Jennifer, and Patrick Vonderau. “‘Where the internet lives.’” University of Illinois Press, 2017.

[9] Fuller, Matthew, and Eyal Weizman. Investigative Aesthetics: Conflicts and Commons in the Politics of Truth. Verso, 2021.

Share