Rundale

Deanna Horton

It’s seven o’clock on a Monday morning at the College Green bus stop in front of Trinity College Dublin. Like thousands of other Dubliners, I’m starting my own commute. Destination; a business park in South Dublin County. People cluster along the pavement, dodging tourists and street sweepers, checking TFI Live, and defending their breakfast rolls from opportunistic gulls. 

On one of these mornings, I noticed two men in heavy work boots and black technical uniforms. The multicoloured Microsoft logo on their sleeves made their destination unmistakable: the Grange Castle data centre campus, part of a much larger concentration of digital infrastructure in West Dublin.

That was the moment I realised I wanted to write my Environmental History thesis on data centres. Not because they were new or my class loved to debate them, but because they were suddenly visible. Realising I worked so close to so many made me wonder how this massive infrastructure had reshaped Dublin’s suburbs, and what history lay behind it.

Over the past decade, data centres have become lightning rods in Irish public debate. They’re blamed for rising electricity prices, resource strain, and the country’s uneasy position in the global tech economy. These arguments treat data centres as purely 21stcentury problems, modern crises to be solved alongside the housing shortage, rising cost of living and the climate emergency.  

But what if I told you that Ireland’s data centres are not new? What if their mostly invisible expansion is rooted not in Silicon Valley, but possibly in Ireland’s own planning system — in decisions made long before anyone had heard the term “the cloud”? 

When I began researching, Patrick Bresnihan and Patrick Brodie’s From the Bog to the Cloud was foundational to how I understood Ireland’s data centre industry and its relationship with Big Tech capitalism, energy, and the digital economy. In another life I was trained as a legal researcher, so to me the obvious first question was: what laws had permitted all this? And what could those laws, planning codes, and land use classifications tell me about how digital infrastructure came to occupy Ireland’s physical landscape?

To understand how Ireland became a major European data centre hub, and how hyperscale infrastructure became central to debates over energy and climate, we have to go back further: to the planning laws of 1963 and 2000 and the early data-processing facilities of the 1980s and 1990s. Because long before hyperscale campuses arrived, Ireland built a planning system that misclassified digital infrastructure, treating it as office work rather than energy intensive industrial activity. That misclassification created a governance blind spot — one that hyperscale operators later exploited.  

The Planning Act That Built the Blind Spot 

On paper, the Local Government (Planning and Development) 1963 Act has nothing to do with computers. It was designed for a country modernising its towns and industries, not for one preparing to host hyperscale cloud campuses. But it quietly built the planning regime that would later shape how digital infrastructure was seen—or more importantly, not seen. 

What the Act did was simple but powerful: it created a centralised framework which required local authorities to prepare development plans and gave these local actors the legal tools to regulate land use, control new buildings, and balance economic growth with amenity and landscape protection. Those development plans contained the actual zoning categories and permissions tables that decided what could be built, where, and under what conditions. 

Those categories reflected the world of the 1960s and 1970s: housing estates, factories, warehouses, offices, shops, parks, and farmland. They were built around broad land use types like residential, industrial, agricultural, and amenity. There was no concept of “digital infrastructure” because, at the time, it didn’t exist as a recognisable category. 

When early data‑processing facilities began to appear in Ireland in the late 1970s and early 1980s—mainframes for banks, insurance companies, health services, and public administration—planners had to fit them somewhere into this mid‑century framework. The only available boxes that made sense were office and light industry.  

So, that’s where they went.

In the 1983 Dublin County Development Plan, zoning matrices defined what was “Permitted,” “Open for Consideration,” or “Not Permitted” in each land use zone. Office uses were treated as service oriented administrative activity, while light industry referred to relatively clean, low impact manufacturing. Early data-processing facilities had no dedicated category of their own. Planners therefore slotted them into these inherited midcentury classifications. Legally speaking, they were not infrastructures; they were simply another form of office‑based work. This misrecognition obscured their material footprint and set the conditions under which later, far larger digital facilities would be approved through the same planning regime. 

And this is where the governance blind spot begins. 

Because these facilities were understood as office or light industrial uses, planning files focused on the things those categories were designed to evaluate: site layout, parking, landscaping, and visual impact. The infrastructural characteristics of digital work—electricity demand, cooling requirements, and network connectivity—were not part of the assessment. Nothing in the planning system asked whether these facilities consumed more power than a small town. 

National policy debates reinforced this framing. Throughout the 1980s and 1990s, digital technologies were discussed as tools of modernisation and economic competitiveness. Parliamentary debates emphasised telecommunications upgrades, foreign direct investment, and Ireland’s place in the emerging “information society.” Data processing was imagined as an administrative necessity, not as an energy intensive industrial activity with environmental consequences. 

However, this was not a single legislative failure. Planning law and digital development evolved together, normalising computation as office work. By the 1990s, digital facilities were routinely sited in industrial estates and business parks on Dublin’s periphery. They looked like ordinary commercial buildings, and the planning system treated them accordingly. Their infrastructural demands remained invisible because the categories governing them were never designed to see them.

So, when the cloud era arrived in the 2000s, Ireland was not starting from scratch. It was building hyperscale infrastructure on top of decades of misrecognition—on a planning system that had never been asked to understand digital infrastructure as spatially significant.  

The Year 2000 – The Blind Spot Persists  

By the early 2000s, these pre-cloud data-processing facilities had multiplied across Dublin’s industrial estates or were retrofitted into empty warehouses, yet they were still being treated as office or light industrial uses. The year 2000 offered a chance to rethink this. Instead, two major reforms reshaped Ireland’s digital economy and planning system without changing the categories through which digital infrastructure was understood. 

The Electronic Commerce Act 2000 formalised Ireland’s digital economy, legalising electronic signatures and enabling online contracting. Digital activity was framed primarily as a commercial and administrative domain, with little recognition of the physical infrastructure behind it or its demands on land and energy.

At the same time, the Planning and Development Act 2000 consolidated and reformed Ireland’s planning system. It introduced Strategic Development Zones (SDZs), streamlined development processes, and created new mechanisms for accelerating largescale projects. But while the Act changed how quickly development could occur, it did not change how development was understood. The zoning categories inherited from earlier development plans remained intact. Digital infrastructure still had no category of its own. 

This is where South Dublin County’s new classification — office‑based industry — enters the story. 

Created in the aftermath of the administrative split that formed South Dublin County Council, office‑based industry was designed for clean, service oriented industrial activity: software localization, administrative processing, and back-office operations. It was a category tailored to the late‑1990s economy. But it quickly became the convenient home for early digital infrastructure as it allowed digital facilities to expand within industrial zones without their actual resource load being recognized.  

The Periphery Creep of the “High-Technology Crescent”  

While office‑based industry explains how early digital facilities remained conceptually invisible, it does not account for the physical conditions that enabled large‑scale digital development in the late 2000s and 2010s. That story emerges from a different dynamic: the transformation of West Dublin’s industrial periphery through the introduction of Strategic Development Zones (SDZ). 

Adamstown, designated in 2001 as Ireland’s first SDZ, became a focus for infrastructure investment. Its planned utilities, transport links and high-capacity services allowed development to proceed with unusual speed and certainty. But SDZs also reshape surrounding areas, concentrating infrastructure and investment in ways that can make neighbouring land more attractive for development.

Grange Castle Business Park sits directly within this reconfigured periphery. Although the business park was never an official SDZ, it was treated like one and ultimately benefited from its proximity to Adamstown and from earlier waves of biopharmaceutical development that had already upgraded utilities and road networks. By the mid2000s, South Dublin County’s western periphery was increasingly imagined as a zone for strategic, high tech development — later described as the “high‑technology crescent.” 

This is the context in which Microsoft’s 2007 decision to build at Grange Castle must be understood and why I used their data center campus as my case study. The company did not select an isolated industrial estate; it chose a landscape already structured by Big Pharma investment and positioned along the infrastructural corridor created by Adamstown’s SDZ. Other operators (such as Google, Digital Realty, Edge ConneX, Vantage, Echelon, and CyrusOne) followed, clustering around Adamstown and the Grange Castle corridor. I call this pattern “periphery creep”: data centres locating outside SDZ boundaries while benefiting from the infrastructure and investment those zones helped create.

Microsoft’s arrival was therefore not an anomaly. It was an early expression of a broader spatial logic — one in which digital infrastructure expanded through preexisting industrial frameworks and along SDZadjacent corridors that had been primed for high-capacity development. 

By the mid2000s, Ireland had modernised its digital governance and expanded its planning apparatus, but it had not rethought the planning regime through which digital infrastructure was classified. Facilities like Microsoft’s were approved quickly and confidently, yet their infrastructural characteristics remained invisible within planning files. The blind spot persisted — now reinforced by new legislative tools and new spatial mechanisms — setting the stage for the energy strain that would reemerge in the decade that followed. 

Conclusion: When the Blind Spot Became Visible

The consequences of these earlier decisions did not become obvious immediately. For years, data centres continued to expand through planning categories inherited from a very different era. But as the facilities became larger, their demand for electricity became increasingly difficult to overlook. By the late 2010s, energy authorities were explicitly identifying data centres as a major source of new electricity demand. What had once looked like another form of office or industrial development had become part of the infrastructure shaping Ireland’s electricity system.

That is the larger story behind Ireland’s data centre debate. The problem did not begin with the cloud, and it did not begin with AI. It developed gradually through decisions about land, planning and infrastructure made over decades. Ireland built a planning system for factories, offices and industrial estates, then used those same categories to accommodate a form of infrastructure its designers could never have anticipated.

So, when we debate data centres and AI today, we are also debating the consequences of that earlier history. The cloud may sound weightless, but it has always existed somewhere: on land, connected to water and electricity, alongside communities shaped by planning decisions. Understanding how those decisions were made is important because Ireland is once again being asked to plan for a technology whose future demands are difficult to predict.

And that brings us back to those Microsoft workers I saw waiting for the bus at College Green. What caught my attention that morning was something ordinary: two people commuting to work. But behind that ordinary journey through Dublin’s geography was a much bigger landscape of planning decisions, infrastructure and history.

The cloud was never really invisible. It was being built around us all along.