Fish Farming in the Digital Cloud

In winter 2022, I drove through blanket peatlands in the shadow of the Seven Sisters mountain chain in Donegal, Ireland, down a one-lane gravel road where local communities expressed turbary rights. The Irish bogs have long been a source of fuel for household heating in the form of peat, or “turf,” as the carbon-dense material of dying plants and mosses accumulates over centuries in this anaerobic, semi-aqueous soil formation.

Stacks of dried and drying peat cut from the bog—along with residual machinery and empty beer cans—sat alongside grazing and wayward sheep along the road and in the distance, giving a visible sense of the active traditional energy and farming cultures in the area. In a Gaeltacht, or Irish-speaking region, these activities were tied up in the protection of local knowledge and culture, long threatened by colonialism, postcolonial under-investment, and contemporary erosion of viable rural livelihoods.

As I came around a bend, I was suddenly confronted with an industrial installation, protected by a fence and security gate. Several people in hazmat suits oversaw a couple dozen circular tanks in this open-air facility. I could hear the buzz of electrical equipment through my open windows.

I had happened across one of the more marginal solutions for bog life in the midst of Ireland’s “just transition” away from carbon-intensive energy and farming: aquaculture. Posed as one option amongst many to locate environmentally “sustainable,” green development activities on the peat bogs, these fish tanks would keep bog carbon locked beneath, while still promising employment for local people and providing a “greener” source of protein for a growing global population. In these rosy imaginaries, multinational industry and research & development can exist side by side with traditional knowledges and lifeways, cultivated in and through “the cloud.”

Perhaps surprisingly, I became interested in fish farming as an offshoot of my research into the application and integration of digital technologies in peatland science. Investigating the interconnections of fish, bogs, and clouds tells us a story about the unfolding injustices of “sustainable growth” in Ireland and beyond. It also enables us to think ecologically about how digital data interfaces with and profoundly shapes environments. A political ecology of data allows us to follow this “fish story” across multiple sites and regimes of knowing and valuing, following a story of mastery, devastation, and possibility.

Fishing for Clouds or Fishing in the Cloud?

The fish farm in Donegal belongs to the Norwegian multinational company Mowi, who produces one-fifth of the world’s farmed Atlantic salmon. Even cursory research on the company reveals its substantial (and heavily promoted) investment in cloud and digital technologies. A 2020 press release about its “Blue Revolution” of sustainability and digitalization from its partner Cognizant promises to “modernize the core IT systems” of the company to support an emerging “cloud first strategy.” By these logics, only through the cloud can Mowi deliver truly sustainable fish farming to a “growing” global population.

Given the description of migrating its IT services from internal servers in Scotland to the “cloud” (specifically Microsoft Azure), one would be forgiven for initially thinking Mowi’s branding of the “cloud first” strategy wasn’t particularly innovative for a multinational enterprise. However, in the sphere of food production and fish farming specifically, the increasingly sophisticated use of digital technologies is becoming a core principle of industry growth.

From the cloud to artificial intelligence (AI), aquaculture companies and their digital partners are producing more efficient yields within what even they acknowledge to be planetary limits. Most AI companies and applications, at whatever size or scale, rely on major cloud providers (think Microsoft, Amazon Web Services [AWS], or Google) for computing infrastructure. Those companies are in turn becoming increasingly interested in the AI aspect of their data-driven supply chains.

Whether autonomous agriculture or environmental monitoring, AI has been pitched as a silver bullet for sustainability. The application of these technologies in food production has provoked frenzied public attention and a surge in research funding into AI applications. If AI can more efficiently manage resources, squeezing out waste and/or making use of it along the supply chain, who says endless growth isn’t possible? Making fish farming digital becomes an existential imperative—the sustenance of humans across the planet relies on the optimized management of fish swimming in the cloud.

In Ireland, this approach also ties the rural bog facility in Donegal to cloud operations in Dublin—specifically the extensive Microsoft data center installations in the western suburbs, connected hundreds of kilometers away by underground fiber optic cables. Cloud computing is primarily hosted in large-scale data centers and already hogs energy, utilizing 18% of Ireland’s electricity on a daily basis in 2022. Recent International Energy Association estimates suggest this will nearly double to 32% by 2026.

As a researcher of data centers and digital infrastructures more broadly, I was used to wandering off and fishing for “clouds” where most would only see air. However, this was the first time I went scanning the cloud for fish. Digital processes in fish farming provide key ecological insights into how “nature” is valued within planetary transitions that increasingly pair decarbonization and digital growth.

The Values (and Dangers) of the Peat Bog

In the bogs, Ireland sees its most profound contestation surrounding the “just transition” and, specifically, the wages of eco-modernity. The Irish state has been phasing out industrial-scale extraction and burning of peat since the 1990s, including closing Electricity Supply Board’s (ESB) turf-powered Gweedore power station nearby the same Donegal fish farm in 1995. Bord na Móna (literally “Turf Board,” the semi-state company that industrialized the bogs), once an image of Ireland’s carbon-intensive modernity, re-branded in 2020 as a “climate solutions” company. The same year, the fruits of an experimental aquaculture project with Bord Iascaigh Mhara (the semi-state seafood agency) on a former cutaway site in Mount Lucas were reported as having “done very well, especially the trout.”

However, in spite of such “just transition” planning, objections and indignation surrounding the vilification of turf-cutting arise from justified resentments about the place of the rural in Ireland’s energy transition, especially the cultural relations and livelihoods that are being sidelined in favor of large-scale climate action. In places where peat is still widely depended upon as a household fuel source, and in the midst of extortionately high imported solid fuel prices, the contradictions of climate action amidst widespread fuel poverty and lack of direct state investment in initiatives such as household retrofitting are stark.

As extremely productive carbon sinks that are otherwise difficult to make “valuable” for capital, research into bog carbon cycles is increasingly tied up in big tech platforms and companies, such as the Microsoft-funded Terrain AI project or the use of Google Earth Engine in digital mapping.

Both post-extractive cutaway and active blanket peatlands are also sites for the development of large-scale renewable energy such as wind farms, frequently contracted to provide energy for data centers in Dublin. An event elsewhere in Donegal in 2020 stands out. In a mountainous part of the border region near the Barnesmore Gap, the construction of the Meenbog wind farm caused a catastrophic peatslide, polluting waterways north and south of the border and destroying a vast swathe of active peat bog.

The power from the planned nineteen turbines on the development had in 2019 been promised to AWS’s Dublin data centers through a direct purchasing contract. The lack of consultation surrounding AWS’s involvement angered existing local objectors to the wind farm, whose correct concerns about the integrity of the bog, its hydrology, and its ecology formed their principle disagreements with the plans.

Much of the dislodged bog found its way into the Mourne Beg River, where local fishermen reported finding fish with gills clogged black with peat. Concerns over damage to a northern trout farm as well as a “total fishkill” in one of the EU-protected Foyle catchment’s most important spawning rivers for Atlantic salmon caused disquiet among agencies and communities on both sides of the border, sparking state investigations into responsibility for the catastrophe.

AWS’s name was absent from most news reports, let alone rulings, on the issue. As they were not legally involved in the development directly, they were spared responsibility for damages. When and if the wind farm becomes operational, however, the contract will undoubtedly be a part of their corporate environmental responsibility portfolio.

“Smart” Fish Farming as Energy Justice?

Like Microsoft Azure, AWS is not just a cloud provider, but a cloud and AI services company. Within this aspect of their operations, they provide sponsorship, support, promotion, and infrastructural provision for a wide range of “sustainability” projects—many of which are geared toward corporate social responsibility as much as potentially profitable research & development. Microsoft, for example, produced a “Changemakers” documentary about its partnership with Spanish multinational Pescanova in constructing the “Smartfarm” platform and producing the “world’s first intelligent aquafarm” in its facilities in South America.

AWS similarly showcased the AI monitoring technologies of a company called Aquabyte in 2019—now widely applied across the industry, especially in offshore farms. According to the Ted Talk-style video, the technology offered facial recognition for fish “from farm to table,” with relatively low-tech photography feeding immense datasets into AI to more efficiently monitor the health (and yield) of the farm. The technology would use this aggregate data to optimize feeding, which, when applied at scale, could eventually operationalize the “holy grail” of fish farming: a fully autonomous farm.

Whether marine or inland, fish farms must navigate the contingencies of biology and ecology in the form of “wild” intrusions such as pests, parasites, and weather. These projects and facilities are also frequently in “remote” regions with reduced signal and internet connectivity. Pescanova, for example, operates “in remote places in Latin America” and claims to “implement connections and artificial intelligence in places where you cannot even use your mobile phone.”

In addition to “wild” contingencies and connectivity, the other major difficulty of locating in such regions is energy supply. After all, like digital systems, electricity is about wires. Laying those wires is expensive, and typically must negotiate with public grids and utilities. Fish farms are themselves relatively energy-intensive, and while producing a fraction of the pollution of beef, tend to release more emissions than similarly scaled pork and poultry producers. In the case of onshore farms, much of this energy is used for thermal management of water and surrounding environs—so much so that in Sweden and other northern countries, data center waste heat has been used to warm indoor fish farms.

These forms of digitalization and automation are especially important when considering the proposed role of fish farming in “just transitions” for communities currently or formerly reliant upon industries such as peat extraction, industrial farming, and commercial fishing. If yields and profitability increase with more automation and fewer workers, the hard developmental promises of multinational industries as job providers fade into the much fuzzier incidental effects of local investment. While companies like Pescanova proclaim that they are transforming their people amidst their technological deployments, Aquabyte is much clearer about the benefits of remote management for planetary-scale aquaculture, promising the eventually automated production of protein as its justification.

In the application of digital technologies and electricity to industrially manage the lifecycles of fish and provide, however chimerically, stable livelihoods for rural communities, we are reminded of the life-sustaining energetic processes from which capital extracts value. Farming is itself an energetic dynamic, harnessing and harvesting the same planetary energies that are increasingly instrumentalized to support decarbonization. Solar and wind, as the After Oil Collective makes clear, are the basis for all energetic processes, from human sustenance through agriculture to the large-scale harnessing of “renewable” atmospheric energies. Within these transformations of the energetic processes sustaining increasingly “green” digital capitalism, the place of “energy justice” becomes one in which the fates and futures of energy, data, and fish are inextricably entangled.

Toward a Political Ecology of Data

This alignment of energy, data, and fish might appear far from where we started. Digital technologies align with durable and familiar frameworks which manufacture territorial encounters between modern technology and remote, irrational ecologies, reproducing particular regimes of marginalization to sustain ongoing economic growth—under the premise of sustainably feeding the planet.

Whether a South American AI fish farm or operations across the Nordics and Ireland, the challenges of “sustainably” producing food within global systems driven by unsustainable growth and accumulation come into friction with the planetary transitions already underway. While sustainability challenges have spurred “eco-innovations” in aquaculture, in Ireland’s peatlands and across the world, it is clear that the ecologies of the digital have been far less considered as a part of these transformations.

This is the eco-modern conundrum: in failing to think ecologically, mastery and devastation marshal the death march of progress, promising the shimmering, cloudy mirages of an equitable modernity while simultaneously closing pathways to the types of energy and environmental justice that would be needed to achieve it. To understand how fish swimming in these clouds can have gills clogged with peat, we have to understand how life is beholden to the energetic faculties of digital capitalism—themselves facilitating expanded forms of extraction resulting in destruction and the enclosure of alternative lifeways.

---

Featured image: The Mowi facility near Falcarragh, Donegal, nestled in the bog. Photo by author, April 2023.

Patrick Brodie is a media scholar and Assistant Professor in the School of Information and Communication Studies at University College Dublin (UCD). His research focuses on the environmental impacts of digital media infrastructures. He is currently writing a book, with Patrick Bresnihan, called From the Bog to the Cloud: Contesting Eco-Modernity in Ireland for Bristol University Press. Contact.