For years, the public was told that moving from paper to digital was an obvious environmental gain. Less printing, fewer trees cut, less waste. That argument now looks too simple. As Scotland courts green data centres and faces new concern over hyperscale AI infrastructure, the comparison between paper and digital systems has changed. Paper uses trees, water, energy and chemicals. Data centres use land, steel, concrete, water, rare minerals and electricity every hour of every day. The honest answer is no longer that one is always better than the other. The honest answer is that Scotland must ask what is being produced, how long it is kept, who powers it, who benefits, and what local damage is being accepted in the name of efficiency.
The environmental case against paper once seemed easy to understand.
A tree was cut. Wood was processed. Pulp became paper. Paper was printed, posted, filed, read, discarded and perhaps recycled. The harm was visible enough for a child to grasp.
Digital technology seemed cleaner because it removed the object from the hand. A document became a file. A newspaper became a webpage. A book became an an epub. A photograph became cloud storage. The waste bin stayed empty, and many assumed the environmental debt had disappeared with it.
It had not disappeared. It had moved.
The servers, networks, cables, cooling systems, backup generators, batteries, devices and power stations behind digital life are not imaginary. They exist somewhere. Increasingly, Scotland is being asked to host more of them.
The Scottish Government has promoted Scotland as a potential home for “green data centres”, arguing that the country’s renewable energy, cooler climate and fibre connectivity could support a low-carbon digital economy. That ambition was set out in the 2021 Green Datacentres and Digital Connectivity Vision and Action Plan, which aimed to position Scotland as a zero-carbon, cost-competitive data-hosting location.
The idea has an obvious attraction. Scotland has renewable energy, cooler temperatures, land, industrial sites and political appetite for green investment. But the data-centre debate has shifted sharply with the rise of artificial intelligence and hyperscale computing. A normal digital service and an AI-scale data-centre campus are not the same environmental proposition.
In the United Kingdom, data centres have already been designated as Critical National Infrastructure. They now sit in the same broad national-security category as energy and water systems. The UK Government made that designation in 2024, arguing that data centres underpin services such as health records, finance, communications and public administration.
That changes the weight of the industry. Data centres are no longer treated merely as private buildings full of servers. They are increasingly described as strategic infrastructure.
For Scotland, this raises a direct question. If digital services are now essential, and if data centres demand rising amounts of electricity, water, land and grid reinforcement, is the digital alternative still environmentally lighter than paper?
The answer depends on what is being compared.
A single email usually has a much smaller material footprint than a printed letter. A digital document shared once may avoid printing, paper, delivery and storage. But that small gain becomes less clear when documents are stored forever, duplicated across systems, backed up repeatedly, searched by AI, attached to multiple accounts, and kept in data centres that require constant power.
Paper has a lifecycle. Digital infrastructure has a lifecycle as well.
Paper begins with fibre. That fibre may come from virgin wood, recycled paper, agricultural residue or a blend of sources. Virgin wood pulp has land, biodiversity, carbon, water and chemical impacts. Recycled paper reduces pressure on virgin fibre but still needs collection, sorting, transport, de-inking, cleaning, heat, water and energy.
The UK Government’s 2024 waste statistics recorded paper and cardboard as the packaging material with the highest recycling rate, at 74.3% under one methodology. That does not mean all paper is harmless. It means paper and cardboard have an established recovery system that can work reasonably well when materials are clean and collected properly.
Scotland’s forestry position is also complicated. Forest Research’s 2025 statistics estimate that almost half of the UK’s forest carbon stock is in Scotland, around 1.9 billion tonnes of carbon dioxide equivalent. Trees and soils therefore hold a very large carbon store. Harvesting that store for short-lived products such as paper is different from using timber in long-lived construction, where carbon may remain locked up for longer.
Forestry and Land Scotland says sustainable forestry can support climate action when woodland is well managed, restocked and used in ways that retain carbon value. Forest Research also notes that harvested wood products, including paper, can contribute to climate mitigation by storing carbon and by substituting for more greenhouse-gas-intensive materials. But paper is usually a short-life product. It stores carbon only briefly compared with a timber frame in a building.
So paper is neither innocent nor uniquely guilty. It can be wasteful if used once and thrown away. It can be part of a circular system if recycled well. It can damage forests if fibre is poorly sourced. It can support managed forestry if fibre comes from certified, restocked, well-regulated woodland. Much of the public argument goes wrong because “paper” is treated as one thing.
Digital systems suffer from the same problem.
A low-energy Scottish archive hosted in a modest, efficient data centre using additional renewable power is one thing. A hyperscale AI campus drawing hundreds of megawatts is another.
Recent public concern in Scotland has focused on the scale of proposed new data centres. Campaigners have warned that planned hyperscale developments could create very large new electricity demand. Some estimates from campaign groups place the Scottish pipeline in the gigawatt range. Those estimates should be read as campaign evidence, not official confirmation of built demand. But even official and parliamentary material now recognises that data-centre demand is becoming a strategic energy-planning issue.
The UK Parliament’s POST briefing in 2026 reported that, in July 2025, the National Energy System Operator estimated data centres used about 2% of UK electricity. It also noted that most data-centre energy powers servers directly, with cooling taking around a third, and that AI servers typically draw more power and fluctuate more than ordinary data storage or general computing.
A National Grid data-centre impact study also cited projections suggesting UK data-centre electricity demand could rise sharply by 2035. In one cited projection, National Energy System Operator modelling showed a rise from 7.6 TWh in 2024 to between 20 and 41 TWh by 2035.
That is the point at which the old “paperless office” claim begins to crack.
If digital systems merely replace unnecessary printing, they may reduce material use. If they create permanent, high-energy computing demand, they become a new industrial load. Scotland cannot judge them by the absence of paper alone.
Water must also be considered. Paper production can use significant water in pulping, washing, bleaching and processing, though modern mills can reduce and recycle water. Data centres may use water directly for cooling in some systems and indirectly through the water used in electricity generation. In Scotland’s cooler climate, some data centres may need less cooling than they would in hotter countries. That is a genuine advantage. It is not a free pass.
Land and landscape also matter. A managed forest is not automatically an ecological good. Scotland has experienced criticism of dense commercial plantations, especially where non-native conifers replaced more diverse habitats or affected peatland. But woodland can also provide biodiversity, flood protection, soil stability, shade, habitat and recreation if planned well.
Data centres occupy land differently. They may be built on former industrial sites, which can reduce pressure on greenfield land. That is the strongest siting argument for some Scottish proposals. But large campuses also need grid connections, substations, access roads, backup power, water infrastructure and sometimes new transmission reinforcement. Their physical footprint extends beyond the perimeter fence.
Energy is the core issue.
A paper document has much of its environmental cost up front: forestry, processing, transport, printing and disposal. A digital file has a small visible cost to the user, but the infrastructure behind it consumes power continuously. The longer the file is stored, duplicated and processed, the more the digital footprint grows.
That does not mean paper is better. It means digital storage is not weightless.
The fairest conclusion is this: for short documents read once, digital is usually likely to be better than printing and distributing paper copies, especially if the digital infrastructure is efficient and powered by low-carbon electricity. For documents that must be preserved for decades, printed, archived and recycled responsibly, paper may sometimes be more durable, transparent and less dependent on permanent energy systems than endless digital storage.
For books, newspapers, public records and civic archives, the answer is particularly complicated. A printed book may be read for decades without electricity. A digital article may be accessible instantly but depends on devices, networks, servers, formats, passwords, companies and energy. Paper can burn, rot or be pulped. Digital files can vanish when a platform closes, a subscription ends, a format dies, a server fails or a company decides the archive is not profitable.
Scotland should therefore stop treating the question as paper versus digital in moral terms.
The better question is what each system is for.
If paper is used for wasteful duplication, junk mail, unnecessary forms or single-use bureaucracy, it should be reduced. If digital systems are used for endless duplication, surveillance, AI processing, low-value data retention and permanent storage of material nobody needs, they should also be reduced.
There is a discipline that applies to both: use less, keep what matters, recycle what can be recycled, power what must be powered cleanly, and do not pretend that hidden infrastructure has no environmental cost. The danger for Scotland is the arrival of hyperscale data centres that run constantly and demand enormous power connections.
For Scotland, three policy questions follow.
First, any major data-centre proposal should publish clear energy-demand figures, water-demand figures, backup-power plans, grid-connection requirements, expected carbon emissions, employment numbers and waste-heat use. Claims of “green” operation should be tested against additional renewable generation, not merely the ability to buy renewable certificates or consume power that other users also need.
Second, Scotland should protect and improve its forestry and paper systems rather than lazily treating paper as the old enemy. Public bodies should use recycled and certified paper where paper is necessary, reduce wasteful printing, and preserve important civic records in formats that will survive corporate platforms.
Third, public-sector digital expansion should be judged by its full infrastructure cost. Moving a form, archive, health record, school file or council service online does not remove environmental impact. It moves the impact into data centres, networks, devices and electricity systems.
The Scottish environmental choice is between two material systems, both with costs. Paper asks questions of woodland, water, recycling and consumption. Digital asks questions of electricity, cooling, land, minerals, hardware, cyber-resilience and permanent dependence on infrastructure few people can see.
If Scotland wants an honest answer, it should reject both myths: the myth that paper is always destruction, and the myth that digital is automatically clean.
Modern Scot has not treated industry-funded environmental claims from either the paper sector or the data-centre sector as settled fact. Developer, trade-body and campaign sources are used only where clearly identified as such. Official statistics, parliamentary material and regulator sources have been preferred where available.
Sources and further reading
Scottish Government, Green Datacentres and Digital Connectivity: Vision and Action Plan for Scotland, 2021. Used for Scotland’s policy ambition to position itself as a zero-carbon, cost-competitive data-hosting location.
Scottish Government Digital Connectivity, Green data centres. Used for the public-facing summary of Scotland’s green data-centre strategy.
UK Government, Data centres to be designated Critical National Infrastructure, 2024. Used for the official UK designation of data centres as Critical National Infrastructure.
House of Commons Library, Data centres: planning policy, sustainability and resilience, 2026. Used for UK policy context on data centres, planning, sustainability, energy and water concerns.
UK Parliament POST, What are data centres and how sustainable are they?, 2026. Used for the National Energy System Operator estimate that data centres used about 2% of UK electricity in July 2025 and for background on data-centre energy, cooling and AI workloads.
National Grid Electricity Distribution, Data Centre Impact Study, 2025. Used for UK data-centre demand projections and modelling references, including projected growth in UK data-centre electricity demand by 2035.
Scottish Parliament Written Question S6W-41364, 2025. Used for Scottish Government acknowledgement that strategic planning is being undertaken for AI data-centre infrastructure demand and the Strategic Spatial Energy Plan.
Foxglove, Data centres planned for Scotland could use three quarters of country’s current electricity demand, 2025. Used as campaign evidence on the possible scale of Scotland’s proposed hyperscale data-centre pipeline. Treated as campaign research, not official confirmation of built demand.
Action to Protect Rural Scotland and Environmental Rights Centre for Scotland, data-centre press releases and resources, 2025 and 2026. Used as campaign and planning-context material on proposed Scottish data-centre electricity demand.
Killellan AI Growth Zone application material, 2025. Used as developer material on the proposed Argyll data-centre and renewable-energy campus. Treated as project claims, not independent proof of environmental performance.
Forest Research, Forestry Statistics 2025. Used for official UK woodland statistics and UK roundwood production data.
Forest Research, Forestry Statistics 2025: Carbon. Used for official estimate that Scotland holds almost half of the UK’s forest carbon stock, around 1.9 billion tonnes of carbon dioxide equivalent.
Forestry and Land Scotland, Timber and the Climate. Used for Scotland-specific background on sustainable forestry, timber and climate action.
Forestry and Land Scotland, Climate Change Plan. Used for background on timber harvesting, peatland restoration and Scotland’s public forestry role.
Forest Research, Wood Products. Used for official explanation of harvested wood products, including paper, and their potential role in carbon storage and material substitution.
UK Government, UK Statistics on Waste, 2024 data. Used for official UK packaging waste recycling statistics, including paper and cardboard recycling rates.
Joint Research Centre, Energy efficiency and GHG emissions: prospective scenarios for the pulp and paper industry. Used for technical background on energy use and greenhouse-gas emissions in the pulp and paper sector.
European Paper Recycling Council, Monitoring Report 2024. Used for European paper and board recycling context. Treated as sector reporting.
Confederation of Paper Industries, recycling and industry material. Used only as industry background on UK paper recycling and recovered fibre use, not as independent environmental proof.
Woodland Trust, State of the UK’s Woods and Trees 2025: Scotland summary. Used for conservation context on Scottish woods, biodiversity, climate, flood protection and soil benefits.
NatureScot, Evidence on carbon and nature. Used for Scotland-specific context on native woodland, peatland and nature recovery.
