2nd November 2026
Britain's growing number of data centres is attracting increasing criticism. Concerns range from the amount of electricity they consume and the water needed to cool their computer systems to the amount of land they occupy and the relatively small number of permanent jobs they create.
These are legitimate questions, particularly at a time when households and businesses are facing high energy costs and the country is struggling to expand its electricity network.
However, there is another side to the debate which deserves considerably more attention. Data centres are becoming essential infrastructure for modern economies, and Britain has good reasons for wanting to maintain its own capacity rather than becoming entirely dependent on computing facilities located overseas.
The challenge is not simply deciding whether Britain should have more data centres. It is deciding where they should be located, how they should operate and what benefits they should bring to the communities hosting them.
There is also a need to look more carefully at some of the claims being made about their environmental impact, particularly water consumption.
Why Britain needs its own data centres
Most people rarely think about where the information they use every day is actually stored or processed.
Online banking, government services, hospital records, insurance systems, business transactions, online shopping and increasingly sophisticated artificial intelligence all depend on substantial computing infrastructure.
Behind many of these services are large buildings containing thousands of computer servers, supported by sophisticated cooling, electricity and communications systems.
The information may appear on a computer or mobile phone thousands of miles away, but the processing and storage often take place in a data centre.
This raises an important national security question. Should Britain depend on other countries to provide the infrastructure supporting essential parts of its economy and public services?
There are already considerable international dependencies in technology. Many of the largest cloud computing companies are American, while important components of the technology supply chain are manufactured in Asia.
There is nothing inherently wrong with international cooperation, but excessive dependence on a small number of overseas providers creates potential vulnerabilities.
A foreign government could impose restrictions on technology exports or services. International disputes could affect access to essential systems, while major technical failures or cyberattacks could disrupt services used by millions of people.
Having data centres in Britain does not eliminate these risks, but it gives the country greater control over an important part of its digital infrastructure.
It is also important to understand that simply locating a data centre in Britain does not automatically make information secure. Ownership, management, encryption, access controls and the laws governing the company operating the facility all matter.
Nor does every piece of information need to be stored domestically. Some information can be stored securely overseas under appropriate legal and technical arrangements.
The sensible approach is to retain sufficient British capacity for essential services and sensitive information while continuing to benefit from international technology and investment.
Digital infrastructure is increasingly as important to resilience as electricity networks, telecommunications and transport.
Water consumption: are the concerns justified?
One of the most frequently raised objections to data centres is the amount of water they use.
It is an understandable concern. Some of the largest facilities require sophisticated cooling systems to prevent thousands of computer servers overheating. In countries with hot climates, water-intensive cooling can place considerable demands on local supplies.
However, it is misleading to assume that every data centre operates in the same way.
There are several different cooling technologies, and the differences between them are significant.
Some facilities use evaporative cooling, where water is deliberately evaporated to remove heat. This requires a continuing supply of replacement water because some is lost into the atmosphere.
Others use closed-loop cooling systems. These circulate water or another coolant around the equipment, transferring heat away from the servers and through a cooling system. Much of the liquid is reused rather than discharged or continually replaced.
This is an important distinction. A closed-loop system can be filled initially and then operate for extended periods with relatively small additions to replace losses or accommodate maintenance. It does not necessarily require a continuous supply of fresh water.
There is, however, a qualification. Some systems use a closed loop around the servers but still rely on evaporative cooling towers elsewhere in the installation. In those circumstances, water consumption continues even though the liquid circulating around the servers is reused.
Other facilities use air cooling or dry cooling, which can avoid water consumption for cooling altogether, although these systems can require more electricity under certain conditions.
The choice of cooling technology therefore matters enormously.
Research published by technology industry organisation techUK in August 2025, based on a survey conducted with the Environment Agency, found that 51% of the 73 commercial data centre sites surveyed used waterless cooling. Another 64% used less than 10,000 cubic metres of water annually, while 89% either measured their water use or did not need to because of their closed-loop systems.
These figures challenge the assumption that every data centre is consuming enormous quantities of water. However, the survey was voluntary and covered only a section of the industry, so it should not be treated as a complete picture of every facility in Britain.
More recent research published in 2026 by the Water Research Centre, reported by techUK, estimated that data centres in England use approximately 1.879 million cubic metres of drinking water annually. That represents around 0.2% of the non-household water market.
The national figure is relatively small, but it does not tell the whole story.
Water consumption is concentrated in a relatively small number of large facilities, and demand can increase during hot weather when water resources are already under pressure. A data centre using substantial quantities of water in an area experiencing shortages could create a serious local problem even if the national percentage appears modest.
The Government's position, set out in a parliamentary statement on 28 October 2026, is that developers should first minimise water requirements, then investigate recycled water and alternative supplies, before considering additional demands on public drinking water supplies.
That is a sensible hierarchy. Rainwater harvesting, treated wastewater and other non-potable sources should be considered wherever practical.
There is also a case for requiring developers to publish actual consumption figures rather than relying on general statements about the efficiency of their equipment.
The debate should therefore be about the technology being installed, the amount of water actually consumed and the availability of local supplies, rather than simply assuming that all data centres are major water users.
Electricity: the bigger challenge?
Water attracts considerable attention, but electricity could prove to be the more difficult long-term challenge.
Data centres require continuous electricity to operate their servers, cooling equipment, security systems and supporting infrastructure. Unlike many businesses, they cannot simply switch off their operations when electricity demand is high.
The Government estimated in its October 2026 parliamentary statement that current UK data centre electricity demand is around 2 gigawatts.
For comparison, that is a substantial continuous demand on the national electricity system. As artificial intelligence develops and computing requirements increase, further growth is expected.
The Office for National Statistics has also highlighted the expansion of the sector. Its analysis estimates that data centres in Great Britain consumed 4.5 terawatt-hours of electricity in 2024, equivalent to approximately 2% of electricity consumed from the grid. Consumption had increased by 41% between 2020 and 2024.
There are two sides to this argument.
On one hand, Britain needs sufficient electricity to support new industries, digital services and technological development. If we want to become a serious player in artificial intelligence, we cannot ignore the infrastructure required to support it.
On the other hand, electricity supplies are not unlimited. New data centres can require major grid connections, and their demands may compete with housing developments, manufacturing businesses and other proposed investments.
There is also the question of who pays for the infrastructure.
If a large technology company requires a new electricity connection, substations and improvements to the local network, the costs should be properly assessed. Existing households and smaller businesses should not find themselves carrying an unfair share of the expense.
Developers should also be encouraged to locate facilities where electricity generation and network capacity are available, rather than simply choosing sites where land is cheapest.
Renewable electricity has an important role to play. Britain has invested heavily in offshore wind, onshore wind and other low-carbon generation. Data centres could help provide a substantial and predictable customer for that electricity.
However, signing a contract to purchase renewable electricity is not necessarily the same as ensuring that the facility is powered by renewable energy every hour of the day. The availability of generation, storage and grid capacity all matter.
There may also be opportunities to make better use of electricity that would otherwise be constrained because the network cannot transport it to where it is needed.
This requires careful planning between developers, electricity generators and network operators.
The waste heat opportunity
There is another aspect of data centre development that deserves more attention: the heat produced by the servers.
A substantial proportion of the electricity entering a data centre eventually becomes heat, which has to be removed to protect the equipment.
Yet that heat need not always be wasted.
In suitable locations, heat recovery systems could transfer it to nearby buildings, industrial processes or district heating networks.
This is particularly interesting in Scotland, where heating requirements are substantial for much of the year.
Imagine a data centre located close to a town with public buildings, housing or commercial premises that could benefit from a reliable source of recovered heat.
The economics would depend on the distance involved, the temperature of the recovered heat, the cost of connecting buildings and the availability of customers. It would not make sense everywhere.
Nevertheless, developers should be expected to investigate the possibility before construction begins.
The Government's October 2026 statement specifically encourages assessments of waste heat recovery and reuse as part of the planning process.
There is a wider principle here. If Britain is going to build energy-intensive infrastructure, we should look for ways to make productive use of the resources it generates rather than treating every environmental impact as unavoidable.
Jobs and the economic benefits: are communities getting enough in return?
Another criticism of data centres is that they can involve investments running into hundreds of millions or even billions of pounds but create relatively few permanent jobs.
There is some truth in this.
Construction can provide substantial temporary employment for engineers, electricians, construction workers and specialist contractors. Once operational, however, a highly automated data centre may employ considerably fewer people than a traditional industrial facility of comparable investment value.
That does not mean the economic contribution is insignificant.
There are opportunities for electrical engineers, IT specialists, cybersecurity professionals, maintenance technicians and other skilled workers. There are also opportunities for companies supplying equipment, security, maintenance and specialist services.
The important question is whether these benefits extend beyond the initial construction period.
Developers should be encouraged to work with local colleges and universities to provide apprenticeships and technical training. Local businesses should have opportunities to compete for contracts rather than seeing most of the economic benefits flow to companies based elsewhere.
Communities hosting major developments should also receive tangible benefits.
It is not unreasonable for local people to ask what they are getting in return for accommodating a facility that may place considerable demands on electricity networks, roads and other infrastructure.
Investment figures alone do not tell us whether a development represents good value for the area.
Scotland could have an important part to play
Scotland has several characteristics that could make it attractive for future data centre investment.
It has substantial renewable electricity resources, a relatively cool climate and opportunities for further investment in wind generation and electricity infrastructure.
The cooler climate can also help reduce cooling requirements, although modern high-performance AI equipment presents additional challenges.
There is potential to connect data centre development with renewable energy projects, battery storage, engineering skills and research.
For the Highlands, this raises some interesting possibilities.
Areas such as Caithness and Sutherland have considerable renewable energy resources, although grid capacity and the availability of suitable connections remain important constraints.
A major data centre would not automatically solve the region's employment problems. Nor should communities accept a development simply because it promises a large capital investment.
However, there is a case for exploring whether digital infrastructure could become another part of Scotland's economic development, particularly in areas looking for new industries as traditional employment changes.
A well-planned development could bring construction work, technical training, apprenticeships and opportunities for local suppliers.
The challenge would be ensuring that the economic benefits remain in the area rather than disappearing once construction is completed.
Government introduces new expectations
The Government has recognised that the expansion of data centres needs a clearer framework.
Its Data Centre Covenant, published on 28 October 2026, sets out expectations covering national security, resilience, electricity costs, water and energy efficiency, local community benefits and employment.
It also aims to accelerate the development of strategically important facilities by improving access to planning and electricity infrastructure.
Importantly, the Government wants greater transparency over water and electricity consumption. Operators are being encouraged to publish annual environmental performance information, including measures of energy efficiency, water usage and opportunities to reuse heat.
There is a recognition that the country needs more computing capacity but that this must be developed responsibly.
Data centres have also been designated as Critical National Infrastructure, reflecting their importance to essential services and economic activity.
This is an important development because it recognises that digital infrastructure is no longer simply a commercial matter. It has become part of the country's national resilience.
Getting the balance right
There are legitimate reasons for questioning individual data centre proposals.
A development in an area with limited water supplies may be inappropriate if it requires substantial additional consumption. A facility demanding enormous amounts of electricity should not be allowed to proceed without proper consideration of the impact on the local network and other customers.
Developers should explain their cooling systems, water requirements, electricity needs, employment commitments and arrangements for dealing with waste heat.
They should also be transparent about the costs of connecting to essential infrastructure and the contribution they will make towards necessary improvements.
But the debate should not become an argument against data centres altogether.
Britain cannot expect to develop artificial intelligence, modernise public services, maintain competitive businesses and protect sensitive information while relying entirely on computing facilities located in other countries.
Nor should we assume that every new development will necessarily be environmentally damaging. Technology is changing, cooling systems are becoming more efficient and there are opportunities to reduce water consumption and recover waste heat.
The important thing is to insist on good planning and proper evidence.
If a developer proposes a water-intensive facility in an area already struggling with water shortages, that deserves serious scrutiny. If another proposes a closed-loop or waterless cooling system, powered by appropriately sourced electricity and designed to reuse waste heat, it should be assessed on its actual environmental impact rather than treated as identical to every other development.
Britain needs data centres, but it needs to build them intelligently.
The real question is whether we can develop the digital infrastructure required for the future while protecting water supplies, managing electricity demand, creating worthwhile employment and ensuring that local communities benefit.
Refusing to build the infrastructure will not stop the growth of artificial intelligence or the increasing demand for computing power. It could simply mean that more of the investment, expertise and economic opportunities end up overseas.
That would be a rather expensive way of avoiding the environmental challenges.
Data Centre Covenant
Data centres and the UK National Accounts