The Hardware of the Software

Why we cannot implement unlimited digital solutions in a limited physical environment

Camilla Siggaard Andersen
13 min readNov 17, 2023

By Camilla Siggaard Andersen (Hassell) and Ana Matic (Scott Brownrigg), based on conversations from the NLA Built Environment Technology Expert Panel

‘Connectivity Matters’ mural curated by Global Street Art for Colt Technology. Photo by Samuel Regan-Asante on

Digital technology is often presented as the counterpart to building technology; one moves fast, scales freely, and exists primarily in a non-physical realm, while the other moves slowly, scales painfully, and is decidedly physical.

Where digital technology is usually celebrated for its ability to seamlessly augment our lives and work, building technology is more likely to be seen to cause disruption. Just think how much quieter and quicker it is to construct an entire world in Minecraft than it is to add a single house extension, for example.

Perhaps because of this apparent ‘non-physicality’, the discussions about digital technology in the built environment tend to be focused on how the opportunities presented by the former might be wielded to better mitigate the challenges created by the latter. As a result, now a whole range of solutions, ranging from online engagement tools to autonomous vehicles, and from BIM models to digital twins, permeate our industry, offering new efficiencies and enhanced productivity.

But what about the physical disruption caused by the increasing permeation of digital technology itself?

Today, when we interact with digital technology, it is easy to overlook the physical impacts of our activities. And yet digital infrastructure is starting to make its mark known on our built environment, communities, and resources in a very real way. On a global scale, the International Energy Agency estimates that data centres and data transmission already account for 1% of greenhouse gas emissions. With new AI-powered digital applications becoming mainstream, this environmental impact is only expected to increase further.

Whilst actively promoting the digitisation of urban environments to provide better resilience to future climate and energy supply challenges, we must also consider better awareness of the overall physical impact of digital networks and develop relevant urban planning requirements to keep up with the scale of the development.

In this article, written as part of the Built Environment Technology Expert Panel’s work for the organisation New London Architecture, we interrogate the physical requirements of three fundamental components of the digital realm: 1) the Cloud, 2) its networks, and 3) their interfaces.

The Cloud (aka buildings & storage facilities)

“(…) the internet isn’t a ‘cloud’ — it is a network of cables, crossing countries and continents and enveloping the world, joining servers that hold our data.” — James Ball, “The System”, 2020

Most people who have worked in an office sometime between the 1980s and 2000s will remember a door marked ‘server room’, behind which loomed a hot, humming, blinking beast of a digital machine, barely constrained by a mess of colourful cables and hardworking zip ties. However, over the past decade, these once-familiar spaces have steadily vanished as companies and businesses have shifted their data storage and computing needs to ‘the cloud.’

Remember server rooms? Photo by Massimo Botturi on

Cloud-based services became widely available in the 2010s, offering off-site data storage and processing to alleviate the pressure on in-house IT infrastructure. The term ‘cloud computing’ was coined in 2006 by Google’s CEO Eric Schmidt, indicating that the physical requirements of digital tools would somehow evaporate from the face of the Earth. In actuality, the server rooms merely moved in together in dedicated warehouse-like buildings, also known as data centres, managed by global technology giants like Amazon, Microsoft, IBM, and Google. With their square feet firmly planted on the ground, these facilities were every bit as physical as the server rooms they replaced, only with more centralised operating requirements.

According to the Data Center Journal, there are currently 104 of such data centres distributed across London, with an estimated total building footprint of around 45 hectares[1]. The largest facility, operated by Sungard Availability Services, is located in Hounslow and covers circa 293,000 sq. ft., while the three largest providers, Digital Realty, Equinix, and NTT Ltd., occupy a combined 1 million sq. ft. (9.2 ha) across 26 sites. These companies are all privately owned and headquartered outside of the UK.

By comparison, the capital is served by 325 public libraries with an estimated total building footprint of circa 20 hectares[2]. Both building types may be categorised as information management facilities, albeit with vastly different service propositions. While the library is traditionally an open-access centralised hub of knowledge exchange by physical books and conversation, data centres exist purely as closed storage facilities enabling the distributed access to knowledge via computers and other personal devices. As such, their presence in the built environment often goes unnoticed, even as their numbers and prevalence continue to spread.

It is exactly this quiet expansion of privately-owned physical data infrastructure, driven by the increasing demands for cloud-based services, which must be addressed. For while our appetite for data and data-driven services may be limitless, the city’s supply of land, energy, and water is not.

This is the Cloud. Photo from inside a data centre by Taylor Vick on

This reality is acutely felt in west London, where developers have been struggling to obtain planning permission for new housing on account of the region’s limited electrical capacity, which is largely used up by local data centres. Some estimates suggest that an average 50-megawatt data centre uses as much energy as 5,000 homes, while the cooling requirements of a 15 megawatt facility uses as much water as 2,500 homes.

In a city as dynamic and innovative as London, it is imperative that we address the tangible consequences posed by data centres in conjunction with the capital’s ambitious commitments to achieving net-zero carbon emissions and meeting housing targets. These data hubs, while crucial for our digital age, cast significant shadows over the urban landscape.

Moreover, we should be delving deeper into the economic and political risks associated with data breaches and losses, which may arise due to the limited oversight of data centre facilities by local authorities. The potential fallouts from such incidents extend beyond mere data security, encompassing both financial stability and political trust. As our world becomes more interconnected and reliant on data, the governance of these critical facilities requires meticulous attention. Balancing innovation with responsible management becomes paramount in safeguarding both the city’s digital future and the welfare of its residents.

The Networks (aka cables & linear infrastructure)

“Layered atop the fragile power grid, already prone to overload during crises and open to sabotage, the communications networks that patch the smart city together are as brittle an infrastructure as we’ve ever had.” — Anthony Townsend, “Smart Cities”, 2013

In the digital age, the flow of information across the internet relies heavily on a vast network of ‘data highways’, where data in various forms, from text to multimedia, traverses through intricate cable systems right under our feet. One of the key enablers of the digitally connected world is a 13,000 km transatlantic submarine cable linking the United States and the United Kingdom. Presently, this critical infrastructure is under the stewardship of Tata Communications, a private Indian company with ownership of over 500,000 kilometres of subsea fibre cabling worldwide.

In the United Kingdom, the digital landscape is largely run by Openreach, a subsidiary of BT Group, which oversees a web of 192 million kilometres of cables weaving through urban centres and countryside. The city of London, at the heart of this digital network, relies on the services of several prominent firms, including BT, Vodafone, Plusnet, Community Fibre, and G.Network.

Since 2023, about half of London’s homes have been connected to full-fibre broadband, delivering a swifter and more dependable experience compared to traditional copper wires. In a push to further expand this network, Transport for London has recently awarded a 20-year telecommunication concession to the Canadian-owned BAI Communications, which will see the installation of a high-capacity fibre optic network alongside the existing tube network.

Meanwhile, at street level, G.Network has embarked on a monumental task since 2020, involving the excavation of asphalt to roll out full-fibre broadband across approximately 4,500 kilometres of roadways, with the ambitious target of connecting 1.4 million homes within five years. For anyone living or working in London’s central boroughs, the extensive amount of road construction works that this project has set in motion, have been virtually impossible to miss.

Presently, an illustrative case underscores the far-reaching consequences of such infrastructure developments. The construction of a new data centre hub located in North London, details of which remain confidential, necessitates the creation of a new 132kV Electricity Substation to meet its energy demands. This level of power delivery, in turn, requires trenches of up to 3 meters in width and 2 meters in depth along key roads to accommodate cable installation.

Such large-scale construction and infrastructure projects are not uncommon for standard-capacity data centres and leave enduring effects on local transportation, public services, sustainable rainwater drainage, and, critically, energy supply and consumption.

Physical cable networks for digital communication. Photo by Scott Brownrigg.

Much like the 19th-century urban landscapes were moulded by extensive railway networks, and the 20th century ushered in an era of car-centric development, the 21st-century city finds its opportunities defined by a complex web of data transmission cables.

However, a significant distinction arises when comparing these historical transportation networks to today’s digital communication infrastructure. While railways and roads are integral parts of the public estate, accessible to all and governed by public interests, our contemporary digital communication networks predominantly exist in private hands, with limited mechanisms in place to oversee their rightful use.

Ofcom, the government’s regulatory body for telecommunications, has devised a strategy aimed at fostering competition among broadband providers as a means to protect consumer interests. Nevertheless, there remains a notable absence of comprehensive strategies designed to coordinate the physical integration of these networks within the public sphere. Furthermore, there is a conspicuous lack of public discourse surrounding the private ownership of what can be deemed as vital data superhighways. In this era, where information flows shape the very fabric of our cities, it is essential to consider how we balance private interests with the broader needs of society.

The Interfaces (aka smart devices, screens & digital gadgets)

“… the programmability of platforms by virtue of their API deliberately decentralise and extend conditions of data production while simultaneously recentralising methods of data collection.” — Sarah Barns, “Re-engineering the City”, 2020

How are you reading this article? Unless it was printed out for your convenience, it is highly probable that you are accessing this information through a physical device, such as a computer or smartphone. These devices serve as intermediaries connecting your physical location with the server housing our article in the realm of cyberspace. So far, our discussion has revolved around the role of data centres as physical storage facilities for information, and the function of cable networks as physical links between servers and end-use devices. However, we have yet to delve into the essential aspects of data ingestion and extraction, particularly within the context of built environment technology.

There are, on average, nine connected devices in every household in the United Kingdom, while the global count of internet-connected devices is expected to double from 15.1 billion in 2020 to more than 29 billion by 2030. It is through these devices that the world’s exponential growth of data is both produced and consumed. Global technology companies such as Alphabet, Meta, and Apple capitalise on this data by using it to generate valuable consumer insights which can be used in targeted advertising campaigns.

In the context of the built environment, it has been suggested, and sometimes demonstrated, that large quantities of data can be used to derive insights about the intelligent distribution of resources, smart management of utility and transport systems, and evidence-led decision-making. Visions of holistic ‘digital ecosystems’ and inter-connected ‘digital twins’ of the urban reality frequently excite and inspire city leaders and make for captivating headlines at conferences. While there are many aspects of these concepts to scrutinise (including the transparent, democratic, and equitable use of algorithms and the lack of data standards), there are also very real, physical limitations to contend with. Namely, the sheer number of physical devices required for consistent data extraction to fuel such systems effectively.

The big data enterprises of Alphabet, Meta, and Apple rely on billions of data points harvested from billions of privately owned and maintained devices. In stark contrast, cities are typically limited to ingesting data from publicly-owned sensors and cameras situated in the public domain. These installations demand investment for procurement, installation, and upkeep, either directly or indirectly facilitated by local authorities. Alternatively, cities may choose to purchase or license data from private software companies. While this model presents some opportunities, it also incurs costs through the acquisition and processing of vast, unfiltered datasets. Additionally, it restricts cities to making decisions based on the data deemed commercially viable by private providers.

Creating a built environment capable of consistently harvesting data to enhance its own functionality would necessitate a substantial initial investment in digital sensors and devices, coupled with ongoing maintenance expenditures. In the realm of smart city transformation, London already stands out, particularly in the domain of security. Clarion Security Systems estimates that there is approximately one CCTV camera for every ten citizens, with individuals potentially being captured on camera up to 70 times daily. In 2021, Transport for London awarded a contract for the installation of a minimum of 50 new smart cameras at junctions to help with the enforcement of traffic rules, with each installation costing an estimated £40,000.

Data capture devices in the public realm. Photo by authors.

Beyond the financial expenditures, the installation of physical sensors, cameras, and gadgets for capturing smart city data also carries a financial burden. On average, digital technology devices utilise 67 stable elements, among which 45 have been classified as ‘critical minerals’. Moreover, we have to consider the inherent connection between the scale of device deployment, the magnitude of data points they amass, and the ensuring demand for data storage and processing facilities.


Integrated & sustainable data centre development

In conjunction with the rapid expansion of data centre infrastructure, there exists significant potential for enhanced integration with their surroundings — both in terms of scale and location, as well as improved energy efficiency. The future of digital data management should be seamlessly woven into the urban fabric, leveraging existing infrastructure, powered exclusively by renewable energy sources, and utilising naturally occurring heat sinks for the transmission and storage of data.

Northern Europe is at the forefront of this revolution, exemplified by the utilisation of post-industrial sites such as the Lefdal Mine in Norway. These sites offer secure data storage with stable temperatures, achieved with up to 60% less energy consumption.

Major data industry players worldwide are actively exploring innovative solutions, including underwater storage and integration with large transportation projects involving tunnelling and landscape reshaping. However, these endeavours currently have a substantial carbon footprint and necessitate significant capital investments. This prompts us to consider that the future of data storage may lean towards network-dispersal, with a noteworthy shift in scale towards more localised, bio-integrated solutions.

Micro Data Centre solutions, strategically located in close proximity to residential or manufacturing urban areas, present an opportunity to harness incoming renewable energy and redistribute any excess energy within local public and private networks. To attract the necessary investment for further integration, urban legislative changes are imperative, supported by increased authority vested in city and local planning bodies.

Building smarter and closer data centres. Illustration by Scott Brownrigg.

Ideas for further discussion

As our reliance on digital tools and services continues to grow, so should our awareness of the underlying physical infrastructure. In our technologically advanced society, data centres remain indispensable for storing and processing information, cable networks serve as the essential conduits for information transmission, and physical devices play a crucial role in collecting and presenting data. At present, these systems remain reliant on physical proximity and capacity, much like any other facet of our interconnected world.

We summarise three key challenges for continued discussion:

  1. The environmental costs of data storage and processing, measured by space, energy, water usage, and GHG emissions, requires us to question the unrestricted growth of ‘cloud-based’ services and expanding data models, particularly within the context of urban decarbonisation goals.
  2. The tangible disturbances caused by physical cable networks, along with the vulnerability inherent in such infrastructure, urge us to re-evaluate the privatisation and oversight of critical communication infrastructure.
  3. The dependence on physical devices for data acquisition from the built environment prompts us to examine the business models that may underpin the city’s pursuit of cost-effective and inclusive ‘smartification.’

In working with the Built Environment Technology Expert Panel, we have also devised four ideas to help London advance its digital infrastructure while concurrently safeguarding its environmental integrity and cybersecurity resilience.

  1. Create a London-wide masterplan for the implementation of digital infrastructure that considers public service requirements, the storage of public data, land use, and energy requirements as one integrated and inter-dependent system.
  2. Apply Section 106 requirements to the construction of data centres and prioritise smaller data hubs located closer to the community over large centres in industrial landscapes. The Local Authority Planning process should also mandate data centres to incorporate heat and energy recycling, specifying clear percentage targets. A recent example from Devon shows the potential of data centres to function as secondary heat sources for community facilities, for example.
  3. Establish a governing body and estate management function like Transport for London (Information for London?) to oversee the city’s long-term digital traffic and infrastructure requirements. “Information for London” would be responsible for the resilience of the city’s digital infrastructure and give ownership of the city’s information flows back to the public.
  4. Implement requirements for privately-owned digital devices which are installed by building owners and landowners to monitor the performance of their real estate assets to also automatically collect and share key data points in service of public realm and urban planning.

Collectively, these challenges and concepts underscore the importance of acknowledging the tangible constraints that ‘smart cities’ and digital technology encounter within the built environment.

In tandem with policymakers’ efforts to promote digital integration within communities and businesses, equal attention must be directed towards the physical integration of data centres, cable networks, and their corresponding physical interfaces.

Cover image: ‘Connectivity Matters’ mural curated by Global Street Art for Colt Technology. Photo by Samuel Regan-Asante on

[1] The Data Journal has mapped the footprint of 71 out of 104 facilities. By this overview, the average size of a single facility is 50,491 sq. ft. We have calculated the total data centre footprint based on these figures.

[2] Authors’ calculation based on a review of 20 libraries distributed across the Greater London region.

Originally published at