There is a particular kind of invisibility that comes with digital infrastructure. It sits just beyond everyday awareness, humming behind fences, tucked into industrial estates, or buried inside planning documents that few people read. Data centres fall squarely into that category. They are the quiet engines of modern life, powering everything from streaming and cloud storage to artificial intelligence and cryptocurrency.
What they are not widely recognised as is a growing and significant consumer of water.
That matters. Because while Britain debates reservoirs, leakage, and drought planning, a new and largely unaccounted demand is rising in parallel, driven not by households or agriculture, but by the digital economy.
This is not a hypothetical risk. It is already happening.
The hidden cost of keeping servers cool
At the heart of the issue is heat.
Data centres generate enormous amounts of it. Every query, every streamed video, every AI model training cycle produces thermal energy that must be removed to keep systems operational. Cooling is not optional. Without it, servers fail.
There are several ways to cool a data centre, but one of the most common and efficient methods is evaporative cooling. This process uses water to absorb and dissipate heat, often through cooling towers that release water vapour into the atmosphere.
It works well. It is also water-intensive.
A large hyperscale data centre can consume millions of litres of water per day during peak operation. Even smaller facilities can rival the water use of small towns. In regions where water stress is already an issue, that consumption becomes more than a technical detail. It becomes a strategic concern.
The challenge is compounded by the fact that water use is often indirect. Some data centres use on-site water systems. Others rely on electricity generated elsewhere, particularly from thermal power stations that themselves require water for cooling. In both cases, the digital activity translates into physical water demand.
The connection between a smartphone and a river basin is not obvious. But it is very real.
AI and crypto are accelerating the trend
Not all digital activity is equal in its water footprint.
Artificial intelligence, particularly large language models and machine learning systems, requires vast computational power. Training a single advanced AI model can involve weeks or months of continuous processing across thousands of servers. That translates directly into sustained cooling demand.
Cryptocurrency mining presents a similar challenge. Although less prominent in the UK than in some other countries, it remains a relevant factor globally and influences infrastructure design and investment decisions. Mining operations are energy intensive and, by extension, water-intensive when linked to water-cooled systems.
The result is a surge in high-density data centre development, designed to handle these workloads.
In practical terms, this means more facilities, higher energy consumption, and greater cooling requirements. It also means that water demand is no longer a marginal issue. It is becoming central to the operational model of digital infrastructure.
Britain’s water context makes this more urgent
The UK is not a water-rich country in the way that perception sometimes suggests.
Parts of the South East already experience water stress comparable to Mediterranean regions. Population growth, climate variability, and ageing infrastructure are placing increasing pressure on supply systems. At the same time, regulatory frameworks are pushing water companies to reduce abstraction and improve environmental performance.
Into this landscape steps a new class of industrial water user.
Unlike traditional industries, data centres are often perceived as clean and low-impact. They do not produce visible pollution or heavy traffic. They can even be framed as part of the green economy, particularly when powered by renewable energy.
But water does not feature prominently in that narrative.
This creates a gap between perception and reality. Planning decisions may focus on energy efficiency and carbon emissions while overlooking water consumption. Communities may welcome investment without fully understanding the resource implications.
That gap needs to close.
Transparency is the first missing piece
One of the most striking aspects of data centre water use is how little publicly available data exists.
Unlike energy consumption, which is often reported or estimated, water usage is rarely disclosed in detail. Where figures are published, they are often aggregated, estimated, or presented in ways that make comparison difficult.
This lack of transparency makes it challenging for policymakers, regulators, and the public to assess the true scale of the issue.
It also creates an uneven playing field. Some operators are investing in water efficient technologies and alternative cooling methods, while others continue to rely on more traditional, water-intensive systems. Without clear reporting, there is little visibility into who is doing what.
For a sector that prides itself on data, the absence of data on water is notable.
Technology offers solutions, but not silver bullets
It would be wrong to present this as an unsolvable problem. There are already several approaches that can significantly reduce water use in data centres.
Air-cooled systems, for example, eliminate the need for water in cooling processes, although they can be less efficient in certain climates. Closed-loop systems reduce water loss by recirculating it within the facility. Liquid immersion cooling, where servers are submerged in specialised fluids, offers another promising route with lower water requirements.
There is also growing interest in siting data centres in cooler climates or near abundant water sources, and in using non-potable or recycled water for cooling.
Each of these approaches has merits. None are without trade-offs.
Air cooling can increase energy consumption. Immersion cooling requires new infrastructure and operational models. Relocation raises questions about data sovereignty and network latency.
The key point is that solutions exist, but they require deliberate choice, investment, and in some cases, regulatory encouragement.
Planning and policy need to catch up
At present, water is not consistently embedded into data centre planning frameworks in the UK.
Planning applications may include environmental assessments, but water consumption is not always treated as a primary constraint or decision factor. This contrasts with sectors such as agriculture or heavy industry, where water use is closely scrutinised.
There is an opportunity here to evolve policy in line with emerging realities.
This could include clearer requirements for water usage reporting, integration of water efficiency standards into planning approvals, and alignment with broader water resource management strategies. It could also involve incentivising the use of lower water technologies or alternative water sources.
Importantly, this is not about restricting digital growth. It is about ensuring that growth is sustainable and aligned with national resource constraints.
The local dimension cannot be ignored
While the issue is national in scope, its impacts are often local.
A single data centre development can place additional demand on local water infrastructure, particularly in areas already under stress. This can have knock-on effects for households, agriculture, and ecosystems.
Local authorities and water companies therefore play a critical role. They need visibility of proposed developments, access to robust data, and the ability to assess cumulative impacts.
Community engagement is also important. Residents are more likely to support development when they understand both the benefits and the implications. Transparency builds trust. Silence tends to do the opposite.
A data centre in Slough, England, is colocated with residential buildings
A familiar pattern in a new sector
There is something familiar about this story.
For years, water has been treated as a secondary consideration in many sectors, addressed only when it becomes a constraint or a crisis. The result has often been reactive policy, rushed infrastructure investment, and public frustration.
The rise of data centre water demand risks following the same path.
The difference this time is that the issue is emerging early enough to act. The growth trajectory of AI, cloud computing, and digital services is clear. The associated resource demands are increasingly understood. The question is whether policy and industry move quickly enough to address them.
From invisible to accountable
The digital economy is not weightless. It has a physical footprint that includes land, energy, and water.
Bringing water into the conversation does not diminish the value of digital infrastructure. It strengthens it. It ensures that the systems we rely on are resilient, responsible, and aligned with the realities of a changing climate.
For WaterMatters readers, the implications are clear. This is not just a technology story. It is a water story, a planning story, and ultimately a governance story.
The crypto mine next door may not look like a traditional industrial facility. It may not even be visible at all.
But its water footprint is real.
And it is time we started counting it.
