The race to lead the global artificial intelligence industry is creating a new challenge for the United Kingdom: how can the country build the massive computing infrastructure needed for AI without increasing its dependence on fossil fuels?
Behind every AI chatbot, cloud service, online payment and digital public service is a network of powerful computers operating inside large facilities known as data centres. As demand for artificial intelligence grows, these facilities are becoming bigger, more energy-intensive and increasingly important to national infrastructure.
The UK government wants more of this computing capacity to be located inside the country. Domestic data centres could support sensitive services such as the National Health Service, defence systems and other critical public infrastructure. They could also attract investment, create jobs and strengthen Britain’s position in the global technology economy.
But a major environmental question is emerging. Some proposed large-scale data centres are considering using gas-fired power stations because they cannot obtain immediate connections to the national electricity grid. If these facilities depend on natural gas for years, their emissions could complicate Britain’s legally binding commitment to reach net zero greenhouse gas emissions by 2050.
The issue is no longer simply about whether Britain should invest in artificial intelligence. It is about how that investment will be powered, how much water it will consume and whether economic growth can be achieved without undermining climate commitments.
What Are AI Data Centres and Why Does the UK Need Them?
Data centres are large, highly secure buildings containing computer servers, storage systems and networking equipment. They process, store and transfer the information used by digital services.
Whenever people use online banking, stream films, access cloud storage, play internet-based games or interact with AI applications, data centres may be involved in handling the computing work behind those services.
Why AI Is Increasing Demand for Data Centres
Traditional data centres already support a wide range of digital activities. However, modern artificial intelligence systems can require much more computing power, particularly when they are trained or used by millions of people at the same time.
AI workloads often rely on specialised processors, including graphics processing units, which can consume substantial amounts of electricity and generate considerable heat.
As companies develop more advanced AI models and services, they need additional servers, faster networking equipment and more sophisticated cooling systems. This is driving interest in a new generation of very large facilities, often described as hyperscale data centres.
These facilities can operate around the clock and require reliable electricity supplies. Their size and continuous demand can place significant pressure on local electricity networks.
Why the UK Wants More Domestic AI Infrastructure
The British government’s ambition is to make the country a major global centre for artificial intelligence. One part of that strategy is to ensure that important computing capacity is available within the UK rather than depending entirely on facilities overseas.
This is particularly relevant for sensitive public services. Healthcare organisations may use AI to support medical research, administrative work and clinical services. Defence organisations may also rely on advanced computing for analysis, planning and other security-related applications.
Keeping critical data and computing infrastructure within the country could improve resilience and give public institutions greater control over sensitive systems.
The government’s Data Centre Covenant estimates that the UK will need around 6 gigawatts of AI data centre capacity by 2030 to support its strategic ambitions.
However, building that capacity will require much more than constructing large buildings. It will also require electricity connections, water supplies, skilled workers and suitable locations.
The UK’s Electricity Grid Faces Growing Pressure
One of the biggest challenges facing the proposed expansion is access to electricity.
Data centres need dependable power because their servers must continue operating even when demand is high. An interruption can affect online services, business operations and critical digital infrastructure.
However, connecting large new facilities to the national grid is not always straightforward. Electricity networks need sufficient capacity, and new connections may require substantial investment in transmission infrastructure.
Data Centre Developers Are Seeking More Power Than Current Peak Demand
The scale of demand is particularly striking.
According to figures cited in the investigation, data centre developers were seeking electricity grid connections totalling approximately 72 gigawatts. For comparison, the UK’s peak electricity demand in the previous year was around 45 gigawatts.
These figures refer to requested connection capacity, not electricity that data centres are already consuming. Not every proposed project will necessarily be built, receive a connection or operate at its requested maximum capacity.
Nevertheless, the scale of the applications highlights the pressure that the sector could place on future electricity infrastructure.
Ofgem, the energy regulator, has also warned about the rapid growth of data centre connection requests. In July 2026, it said demand connection applications had increased sharply and that data centres accounted for a substantial proportion of the queue.
Why Some Projects Could Wait Years for Grid Connections
Large electricity connections require planning, engineering work and, in some cases, new network infrastructure.
As a result, some developers have been told that they may have to wait several years before receiving the electricity capacity they need. The investigation reported that some projects could face waiting periods of up to a decade.
For companies investing billions of pounds in new infrastructure, such delays create a major commercial problem.
The government, meanwhile, wants to attract investment and expand domestic AI capacity quickly. This creates pressure to find temporary power solutions while permanent grid connections are developed.
Gas-Fired Power Could Create a New Climate Challenge
One of the most significant findings in the investigation is that at least four proposed hyperscale data centre projects are considering gas-fired power generation to meet their electricity requirements.
Using on-site power generation could allow developers to begin operations before receiving full connections to the national grid. However, if natural gas is used continuously, the facilities could produce substantial greenhouse gas emissions.
Four Proposed Data Centres Could Produce Millions of Tonnes of CO2
The investigation’s calculations suggest that the four proposed facilities would require approximately 1.3 gigawatts of electricity if they operated continuously at full capacity.
Under that assumption, their combined annual carbon emissions could reach around 4.1 million tonnes.
The estimate was compared with the combined emissions of Manchester, Liverpool and Southampton.
This figure needs to be understood carefully. It is a calculated scenario based on the facilities operating at full capacity throughout the year. It is not a record of emissions already produced, nor does it establish that all four projects will ultimately use gas-fired power.
Even with that qualification, the estimate illustrates why the proposed use of natural gas has attracted attention from environmental campaigners and policymakers.
Why Natural Gas Could Undermine Net Zero Progress
Natural gas produces carbon dioxide when it is burned. Although it generally emits less carbon dioxide than coal for an equivalent amount of electricity, it is still a fossil fuel.
If new data centres rely on gas generation for extended periods, their emissions could make it harder to reduce the UK’s overall greenhouse gas output.
There is also a longer-term concern. Electricity generation from renewable sources such as wind and solar can become cleaner as additional capacity is installed. Gas-fired generation, by contrast, continues to produce direct emissions whenever the fuel is burned.
The climate impact will therefore depend not only on how much electricity data centres consume, but also on the source of that electricity and how quickly temporary gas generation is replaced.
What Is the UK’s 2050 Net Zero Target?
Net zero means reducing greenhouse gas emissions as far as possible and balancing the remaining emissions through removals from the atmosphere.
The UK has a legally binding commitment under the Climate Change Act 2008, amended in 2019, to reduce its net greenhouse gas emissions by 100 per cent compared with the 1990 baseline by 2050.
This target applies across the economy, including electricity generation, transport, industry, buildings and other sectors.
Why Data Centre Expansion Matters for Climate Policy
The expansion of AI infrastructure does not automatically mean that the UK will miss its climate target.
Data centres can operate using low-carbon electricity, improve energy efficiency and adopt technologies that reduce their environmental impact. The challenge arises when additional demand is met through new fossil-fuel generation rather than cleaner sources.
The government therefore faces a difficult policy decision: how to accelerate an industry that could contribute to economic growth and national security while ensuring that its energy requirements remain compatible with the country’s climate commitments.
The outcome will depend on the electricity mix, the speed of grid upgrades, the duration of temporary power arrangements and the environmental standards applied to new projects.
Water Consumption Is Another Concern for Local Communities
Electricity is not the only resource required by large data centres. Water can also play an important role, particularly in cooling systems.
Servers generate heat while processing information. If temperatures become too high, equipment can become less efficient or suffer damage. Data centres therefore use ventilation, cooling equipment and, in some cases, water-based systems to control temperatures.
Why AI Data Centres Need Cooling Systems
Some facilities use water directly in cooling processes. Water may also be consumed indirectly during the production of electricity used by the data centre.
The amount required varies significantly depending on the facility’s design, cooling technology, local climate and energy source.
As AI workloads become more demanding, operators may need more advanced cooling arrangements to manage heat generated by high-performance processors.
This has raised questions about whether large facilities could compete with households, agriculture and other businesses for water in areas already facing supply pressures.
Concerns Around the Proposed Microsoft Data Centre in Leeds
One example discussed in the investigation is Microsoft’s proposed data centre campus at Skelton Grange in Leeds, on the site of a former power station.
The development plans include multiple data centre buildings and associated infrastructure. Local campaigners have raised concerns about whether existing water infrastructure would be able to support the project’s requirements.
Microsoft, in response, said that water would only be required to support cooling when outside air temperatures rise above 29 degrees Celsius. The company also said rainwater harvesting would be installed to help support water use.
These are competing perspectives that need to be assessed through the planning process and technical evidence. The final impact will depend on the facility’s actual design, operating conditions and the availability of sustainable water supplies.
Why Water Planning Will Become More Important
Water demand is a particularly important issue for regions that already experience seasonal shortages or are expected to face greater pressure as temperatures rise.
A data centre may bring investment and employment to a community, but local residents also need confidence that the development will not put essential resources under additional strain.
That makes water efficiency, rainwater collection, alternative cooling systems and transparent reporting important parts of responsible data centre development.
The UK’s New Data Centre Covenant: What Has the Government Announced?
On 28 October 2026, the UK government published its Data Centre Covenant as part of its wider effort to accelerate the development of strategic data centres.
The policy is intended to balance the country’s need for computing infrastructure with responsibilities towards local communities, energy consumers and the environment.
Five Main Expectations for Data Centre Developers
The covenant sets out expectations covering several areas:
- Fair energy costs: Data centres should pay their fair share of the costs they create for the electricity system rather than shifting those costs onto households and other businesses.
- Local community investment: Strategic projects should develop local benefit plans, with investment directed towards priorities identified with communities.
- Environmental responsibility: Operators are expected to protect nature, minimise water use, improve energy efficiency and publish annual information about energy and water consumption.
- Jobs and skills: Developers should create employment opportunities, support apprenticeships and invest in training and UK supply chains.
- National security: Facilities should provide secure and reliable computing capacity for critical services and work with relevant authorities to strengthen cyber resilience.
The covenant also encourages efficient cooling systems and the reuse of waste heat where practical.
Could Data Centres Receive Faster Grid Connections?
The government says it will use new powers and work with energy network organisations to help strategically important projects access electricity capacity more quickly.
This could include reserving or reallocating grid capacity, identifying additional capacity and facilitating earlier connections where possible.
The government also plans to develop a tightly controlled framework for temporary on-site power generation.
This option is intended for exceptionally strategic projects where other measures cannot provide the required capacity in time. It is not a blanket guarantee that every data centre will be allowed to build and operate its own power station.
The precise conditions and eligibility requirements will be important in determining how this policy works in practice.
Can Renewable Energy Provide a Cleaner Solution?
Renewable electricity could help reduce the climate impact of the UK’s expanding data centre sector.
Wind and solar power are particularly important because they can generate electricity without the direct carbon emissions associated with burning natural gas.
However, supplying large data centres with reliable electricity around the clock is more complicated than simply installing renewable generation.
Wind and Solar Need Support From the Wider Energy System
Wind and solar output changes according to weather conditions and the time of day. Data centres, by contrast, generally require continuous electricity.
A cleaner power system therefore needs a combination of renewable generation, electricity storage, stronger transmission networks, flexible demand and other low-carbon sources.
Developers can also enter into long-term electricity agreements with renewable energy producers. Such arrangements may support new clean-energy projects, although the actual impact depends on when and where electricity is generated and consumed.
Energy Efficiency Could Reduce the Pressure
Data centre operators can reduce electricity requirements through more efficient servers, improved cooling systems and better facility design.
Some technologies can reduce the amount of energy required to remove heat from computing equipment. Operators can also improve efficiency by managing workloads and using computing resources more effectively.
The government’s covenant encourages high efficiency standards and the reuse of waste heat where a practical local demand exists.
These measures will not remove the need for additional electricity, but they could reduce the amount of new generation and infrastructure required.
Economic Growth Versus Environmental Protection
The debate over AI data centres reflects a wider challenge facing governments around the world.
Artificial intelligence is increasingly viewed as an important source of future economic growth. Countries want to attract technology investment, develop domestic expertise and avoid becoming overly dependent on foreign computing infrastructure.
For the UK, data centres could support research, public services, defence capabilities, business productivity and skilled employment.
However, the benefits need to be weighed against the environmental costs of construction, electricity consumption, water use and emissions.
Why Delaying Development Also Has Economic Risks
The government argues that delaying data centre construction could discourage investment and encourage companies to build facilities in other countries.
If that happens, the UK could lose opportunities for jobs, research and domestic technological development.
But moving infrastructure overseas does not automatically eliminate its environmental impact. The electricity used by overseas data centres may also come from fossil fuels, depending on the country and its energy system.
This makes the issue more complicated than a simple choice between building data centres or refusing them.
The central question is whether the UK can attract investment while setting credible standards for clean energy, resource efficiency and local accountability.
What Happens Next for the UK’s AI and Net Zero Plans?
The debate is likely to continue as developers seek planning permission, electricity connections and approval for new projects.
Several developments will be important to watch.
First, the government will need to clarify how its temporary on-site power framework will operate and what restrictions will apply to projects using fossil fuels.
Second, energy network upgrades and reforms to the electricity connection queue will determine how quickly viable data centres can access the grid.
Third, local planning decisions will help establish whether proposed developments can secure sustainable water supplies and deliver meaningful benefits to the communities hosting them.
Finally, annual reporting on electricity use, water consumption, emissions and efficiency will be essential for judging whether the industry’s environmental commitments are producing measurable results.
The UK does not necessarily have to choose between becoming an AI leader and pursuing net zero. But achieving both ambitions will require careful planning, credible environmental safeguards and a much cleaner supply of electricity.
Conclusion: The Future of AI Must Also Be Sustainable
The expansion of artificial intelligence is creating a new kind of infrastructure challenge for the United Kingdom. Data centres are becoming essential to the digital economy, but their growing demand for electricity and water is raising serious questions about how that infrastructure should be developed.
The possibility of gas-fired power at some proposed sites highlights the tension between rapid technological growth and the country’s long-term climate commitments.
The government’s Data Centre Covenant is an attempt to address these concerns by linking support for strategic projects with expectations on energy costs, water efficiency, local investment, jobs and national security.
Whether that approach succeeds will depend on its implementation. Temporary gas generation must not become a long-term substitute for clean electricity, and communities must be able to see clear evidence that environmental promises are being met.
Britain’s ambition to become an AI superpower will be measured not only by how many data centres it builds, but also by whether it can power them responsibly. The real test is whether technological progress can strengthen the economy without creating a lasting cost for the climate and future generations.
According to Reuters