As AI companies race to build vast datacentres, residents living beside them are pushing back over noise, size, energy and lost views. A dispute on the edge of Newport in South Wales shows how sharp that tension has become.
Key takeaways
- The Guardian reports that a 23-metre-high hyperscale datacentre now dominates the view from suburban gardens on the western edge of Newport in South Wales.
- Local Labour politicians have promoted the datacentre as a source of jobs, while some neighbours see it mainly as an intrusion on their homes.
- Demand for computing power to train and run artificial intelligence is the main reason datacentres are being built at greater scale and closer to existing communities.
- Datacentres need large and reliable electricity supplies, which pushes developers towards sites with strong grid connections rather than remote locations.
- The Newport case is part of a wider pattern of local resistance to digital infrastructure, in which residents argue that national benefits are bought with local costs.
What is happening on the edge of Newport?
The Guardian reports on a hyperscale datacentre built close to homes on the western fringes of Newport, a city in South Wales. The building is about 23 metres high and is clad in prefabricated grey panels. For at least one household whose garden backs onto the area, it now blocks almost the entire view that used to open out from the lawn.
The report sets this local scene against a much bigger story. Technology companies want more and more computing capacity, and that capacity has to sit somewhere physical. Increasingly, it sits in very large buildings put up on the edges of towns and cities, near the people who already live there.
The Guardian describes the situation as a David and Goliath contest between residents and corporate giants, and says the Welsh city is fighting back. The material available does not explain what form that resistance takes, whether it involves planning objections, legal challenges, council action or community campaigning. Which companies are involved, and how far any dispute has progressed, are also not set out in the summary.
Why is this in the news now?
The immediate reason is the boom in artificial intelligence. Training large AI models and serving them to millions of users takes enormous amounts of processing, and that work runs on specialised chips housed in datacentres. Companies building AI services have been expanding their datacentre capacity fast, and the scale of individual sites has grown with it.
That expansion has moved datacentres from the background into public debate. For years they were mostly invisible to the public, tucked away on business parks. Hyperscale facilities are harder to ignore. They are physically huge, they need substantial power connections, and they are increasingly proposed or built in places where people notice them every day.
In the UK, politicians across government have presented digital infrastructure as part of an economic growth strategy, and datacentres have been given a higher priority in national policy. That makes local disputes politically awkward. The Guardian notes that, in Newport, the area’s Labour politicians have described the datacentre as a creator of jobs, which puts them on a different side from residents who feel it has been imposed on them.
What does a newcomer need to know about datacentres?
A datacentre is a building full of servers: computers that store data and run software for other people over the internet. Every time someone streams a video, uses a cloud service or asks an AI chatbot a question, servers in a datacentre somewhere do the work.
“Hyperscale” refers to the largest kind of facility, usually run by or for major technology firms that need computing capacity on a massive scale. These sites are designed to be expanded in modules, which helps explain the use of large prefabricated panels such as those described in Newport.
Three practical needs shape where datacentres go. The first is electricity. Servers draw a great deal of power continuously, so developers look for locations with strong connections to the grid. The second is cooling. All that computing produces heat, which has to be removed using air systems, water or a mix of both. The third is connectivity. Sites need fast fibre links to the rest of the internet, and some uses benefit from being close to major population centres.
Those needs often point towards the edges of existing towns, where land, power lines and fibre come together. That is one reason datacentres can end up next to housing rather than in remote countryside.
Who is affected, and how?
Nearby residents feel the most direct effects. The Guardian’s account centres on visual impact: a building tall and wide enough to dominate gardens that were previously open to the sky. Residents near large datacentres elsewhere have also raised concerns about noise from cooling equipment, construction traffic, round-the-clock lighting and the effect on property values. How far any of these apply in Newport beyond the loss of view is not detailed in the material available.
Local authorities are affected because they make planning decisions and collect business rates, and because they have to weigh investment against the wishes of residents. Elected representatives face similar pressures, particularly when they have publicly backed a project as good for employment.
Electricity networks and other energy users are affected too. A large datacentre can need as much power as a small town, and connecting it can mean new grid infrastructure and competition for limited connection capacity.
Workers and jobseekers are the group politicians usually point to. Construction brings temporary work, and operating a site needs technicians, security staff and maintenance contractors. The number of permanent jobs a datacentre creates, compared with its physical footprint, is one of the most disputed points in these debates.
Where do informed people disagree?
Supporters argue that datacentres are essential infrastructure for a modern economy, comparable to roads or power stations. On this view, if the UK wants a share of the AI economy, it has to host the computing that supports it, and blocking projects simply pushes investment to other countries. They also point to tax revenue, supply-chain spending and the jobs involved in building and running the sites.
Critics respond that hyperscale datacentres are highly automated and employ relatively few people once they are running, so the jobs argument can be overstated. They argue that the profits mostly go to large multinational companies, while costs such as noise, visual intrusion and pressure on local power and water fall on the surrounding community.
There is also disagreement about energy. Some argue datacentres can drive investment in renewable generation and grid upgrades. Others worry that their demand could slow the decarbonisation of the wider electricity system or raise costs for everyone else.
Finally, there is a debate about process. Residents in disputes like the one in Newport often say decisions felt rushed or remote, while developers and governments argue that slow planning systems hold back necessary investment.
What are the practical implications?
For people living near a proposed or existing datacentre, the practical questions are about planning. Local planning applications are usually open to public comment, and conditions attached to a permission can cover issues such as noise limits, landscaping, lighting and operating hours. Residents who understand that process are better placed to raise concerns early.
For councils, the Newport case shows the risk of treating datacentres purely as economic wins. Height, massing and closeness to homes can become the issue that defines public opinion about a project, whatever its other merits.
For the technology industry, local opposition is becoming a real constraint. Companies may face more pressure to choose sites with greater separation from housing, invest in screening and design, offer community benefits, or show clearly how many local jobs they will create.
For policymakers, the tension between national ambitions for AI and local consent is likely to grow. How governments balance faster approvals against meaningful local input will shape where future datacentres are built.
What should readers watch next?
The most immediate thing to watch is how Newport’s resistance develops, including any formal objections, council decisions or changes to how future applications are handled. The available material does not set out the timetable or the specific measures involved.
More broadly, it is worth watching whether national planning policy in the UK gives datacentres a faster route to approval, and how that is received by communities. The growth of electricity demand from datacentres, and how network operators handle requests to connect them, will also be a key issue.
Finally, similar disputes are emerging in other countries where AI-related construction is accelerating. Whether developers adapt their designs and site choices in response to local opposition, or whether governments override it, will show how the balance of power between residents and large technology firms is shifting.
Frequently asked questions
What is a hyperscale datacentre?
A hyperscale datacentre is a very large facility built to house tens of thousands of servers, usually for major technology companies or cloud providers. These sites are designed to grow in stages and to handle huge computing workloads, including training and running artificial intelligence systems. Because of their size, power needs and cooling equipment, they tend to be far more visible and contentious than the smaller server rooms and older datacentres that preceded them.
Why are datacentres being built near homes?
Datacentres need strong electricity connections, fast fibre links and suitable land, and these often come together on the edges of existing towns and cities. Developers also value being reasonably close to population centres for connectivity. As a result, sites that suit the industry are frequently next to residential areas. The growing size of AI-related facilities makes that proximity more noticeable, which is why disputes with neighbours have become more common.
Do datacentres create many jobs?
Datacentres create jobs during construction, and running them requires technicians, engineers, security staff and contractors. However, hyperscale sites are heavily automated, so the number of permanent jobs is often modest relative to the size of the building. Supporters stress the wider economic benefits, including supply chains and tax revenue, while critics argue that the employment case is sometimes overstated. The figures for any individual site depend on the project.
Why do AI systems need so much electricity?
Artificial intelligence models are trained and run on specialised chips that carry out vast numbers of calculations. Thousands of these chips working continuously draw large amounts of power, and the heat they produce must be removed by cooling systems that use further energy. This is why AI-focused datacentres need substantial grid connections and why their growth has raised questions about pressure on electricity networks and the pace of decarbonisation.
Can residents object to a datacentre planning application?
In the UK, planning applications for large developments are normally published by the local authority and open to public comment for a set period. Residents can submit objections about issues such as visual impact, noise, traffic and lighting. Councils must consider relevant planning matters when deciding. Larger or nationally significant projects may follow different procedures, and the scope for local influence can vary depending on how the application is handled.
What is happening with datacentres in Newport, South Wales?
The Guardian reports that a hyperscale datacentre about 23 metres high has been built on the western edge of Newport, close to suburban homes, and dominates the view from neighbouring gardens. Local Labour politicians have presented it as a source of jobs, while residents have voiced frustration. The report says the city is fighting back against the trend, though the specific measures involved are not detailed in the material available.
Sources and further reading
- The Guardian, technology section: reporting on the datacentre on the western edge of Newport and local reaction to it.
- UK government planning and digital infrastructure policy documents: background on how datacentres are treated in national policy.
- Electricity network operators and energy regulators: general information on grid connections and rising demand from large energy users.
- Local authority planning portals: how planning applications are published and how residents can comment.
Surfaced from the rss:guardian_tech signal “local datacentre opposition”. AI-assisted draft, editorially reviewed.
