10th October 2026
Plans for a giant artificial intelligence campus at a former nuclear power station promise billions in investment. But the scale of the electricity demand raises a difficult question: how much should Scotland commit to one development, and what will local people actually gain?
Scotland could be about to host one of the largest data-centre developments in the world. The proposed campus at Chapelcross, near Annan in Dumfries and Galloway, carries a headline investment figure of £16.5 billion and could create hundreds of jobs.
It is the sort of announcement that governments normally welcome. A former industrial site could gain a new purpose, businesses could win contracts, young people could find apprenticeships and Scotland could play a larger role in the rapidly expanding artificial intelligence industry.
But there is a catch. The proposed campus would need an enormous amount of electricity, and its environmental impact, land use and long-term economic benefits are already being challenged.
The project raises a question that reaches well beyond Dumfries and Galloway: is Scotland preparing for the economic opportunities offered by AI, or is it taking a gamble on a development whose costs and benefits have yet to be fully established?
What exactly is being proposed?
CX Tech Limited has submitted an application for planning permission in principle for the CX Technology Campus on land north of the former Chapelcross nuclear power station.
The proposal includes eight large data-centre buildings, a training centre, industrial space, a substation and supporting infrastructure. The campus would have a proposed capacity of around one gigawatt of computing infrastructure.
The company puts the potential investment at approximately £16.5 billion. Its published projections include:
Up to 750 construction jobs over an estimated eight-year building programme.
Around 450 permanent direct jobs once the campus is operating.
A total of 320 local apprenticeship opportunities over the construction period.
Potential gross annual business rates of around £8.2 million at current rates, subject to the eventual valuation and applicable tax arrangements.
These figures make the project potentially significant for the local economy. But they remain projections attached to a proposal, not guaranteed results.
The distinction matters. The £16.5 billion is the estimated value of investment over the development, not a cheque already written or money that will necessarily be spent in Scotland. The jobs depend on the project receiving permission, securing financing, obtaining the necessary infrastructure and progressing through construction.
The company has indicated a possible construction period running from 2027 to 2035 if planning permission and the other necessary approvals are secured. There is still a considerable distance between submitting an application and delivering a project of this scale.
Why build it at Chapelcross?
Chapelcross is not an arbitrary choice of location.
The former nuclear power station began generating electricity in 1959 and ceased operation in 2004. It is now being decommissioned. Public bodies have been looking at how the wider site could contribute to a new economic future for the area.
The Nuclear Decommissioning Authority has worked with CX Power on plans for a broader energy hub. Ideas have included renewable energy, hydrogen, advanced manufacturing and digital infrastructure.
The AI campus is therefore part of a wider ambition to turn a former nuclear site into a centre for new industries.
Scotland also has considerable wind-power resources and an established energy sector. That gives the country an opportunity to attract businesses that need large amounts of electricity and reliable digital infrastructure.
There is a sound economic argument for trying to use these advantages. But the availability of renewable electricity nationally does not automatically mean that enough electricity can be delivered to one particular site, at the right times and at an acceptable cost.
That is where the controversy begins.
One gigawatt: what does that mean?
The proposed one-gigawatt capacity is difficult to grasp because we tend to think about electricity in terms of individual homes and household bills.
A gigawatt is one billion watts of power. If a facility used one gigawatt continuously for a full year, it would consume approximately 8.76 terawatt-hours of electricity.
Campaigners have cited an estimated annual demand of around seven terawatt-hours for the proposed campus. That is equivalent to an average continuous load of about 800 megawatts. The difference reflects the distinction between a facility's maximum capacity and the amount of electricity it is expected to consume in practice.
Either way, the demand would be enormous.
Seven terawatt-hours is seven billion kilowatt-hours. At that level, the campus would use electricity on a scale comparable with the annual consumption of more than two million homes, according to the Scottish Greens' comparison.
This is why the electricity question cannot be treated as a technical detail to be sorted out later. It is central to whether the project can be delivered.
The developer says the campus is intended to use renewable electricity and that the scheme will be designed for energy efficiency. It has also referred to direct connections with renewable-energy projects.
Those proposals deserve serious examination. Which generating projects would supply the campus? Are they already operating, or would new generation have to be built? What happens when wind output falls? What grid reinforcement and backup arrangements would be needed? And who would pay for the necessary infrastructure?
These questions are not arguments against AI. They are the questions that should be answered before Scotland commits itself to a development of this scale.
Will it mean an electricity shortage for everyone else?
Opponents argue that a development with this level of demand could place serious pressure on Scotland's electricity system.
The developer's response is that the campus would be designed around dedicated renewable supplies and new infrastructure. If those arrangements can be delivered, the effect may be different from simply taking a large amount of electricity from the existing network.
But even a direct connection does not make the wider energy questions disappear. Renewable electricity is part of a connected system. Generating capacity, transmission networks, storage and the timing of demand all matter.
There is also an opportunity cost. Electricity infrastructure and renewable generation require investment, land, equipment and time. If a large data centre becomes a priority customer, the public should be able to see how that decision fits with Scotland's wider plans for homes, existing businesses, transport and other industries.
The question is not merely whether Scotland can generate enough electricity in a year. It is whether the electricity can be delivered reliably, when required, without creating unacceptable costs or delaying other worthwhile developments.
The planning process should demand clear answers rather than rely on general assurances that the project will use green energy.
What does Scotland get in return?
The jobs figures are attractive, particularly in an area where major long-term investment can make a substantial difference.
Construction would create demand for civil engineering, electrical installation, building work, transport and other services. Apprenticeships could give young people a route into skilled employment. Once operating, the campus would require staff to maintain its equipment, manage its systems and provide security and technical support.
The developer also estimates gross annual business rates of around £8.2 million at current rates. That could be significant for the local public finances, although the final figure would depend on the property's valuation and the applicable rules.
But the distinction between construction and permanent employment is important.
The proposed 750 construction jobs would be spread across an eight-year programme. They would not necessarily all be local jobs, and the number of people employed on site would vary as construction progressed.
The permanent workforce is forecast at around 450 direct jobs. That is a substantial number, but it is much smaller than the headline construction workforce might suggest.
Nor does the total investment figure tell us how much money would remain in the local economy. Some expenditure would go on specialist equipment, technology, finance and contractors who may be based elsewhere.
The right questions are therefore practical ones. How many local people would be employed? What skills would they need? How many apprentices would complete training and move into sustained employment? What proportion of construction contracts could be won by Scottish firms? And what arrangements would ensure that local businesses had a fair chance to compete?
Those are the measures that would help people judge whether the project delivers lasting prosperity rather than simply an impressive investment announcement.
The environmental argument
The project has also attracted objections over the proposed use of land and its environmental effects.
The Scottish Greens have argued that a substantial proportion of the proposed campus would occupy greenfield land rather than make full use of previously developed ground. They have also raised concerns about the greenhouse-gas implications described in the developer's own environmental assessment.
The developer says the campus will be designed around renewable electricity, energy efficiency and low water consumption. Those are important commitments, but they need to be assessed against the detailed evidence.
Data centres use electricity to run computing equipment and more electricity to remove the heat that equipment generates. Cooling systems, backup arrangements, construction materials, transport and the wider energy supply can all affect the development's environmental footprint.
A claim that a facility will use renewable electricity does not, by itself, answer every question about its emissions or wider environmental impact.
The planning authority needs to consider the evidence on landscape, biodiversity, traffic, noise, water, drainage, emissions and the cumulative effect of the development. It also needs to examine whether the proposed location makes the best use of land that has already been developed.
The fact that the project is next to a former nuclear power station does not mean all the proposed land is brownfield or that every environmental issue has been resolved.
Why is the Scottish Government involved?
This is not an ordinary small-scale planning application.
In August 2026, the Scottish Government introduced a direction requiring local authorities to notify ministers of data-centre applications exceeding 50 megawatts. The aim is to give ministers strategic oversight of projects that could have major implications for energy and other infrastructure.
Chapelcross, at around one gigawatt of proposed capacity, is far above that threshold.
The Scottish Parliament has also seen calls for a pause in decisions on very large data-centre proposals until a clearer national approach has been established. Opponents argue that Scotland should not assess each development in isolation when the combined demand could have significant implications for the electricity network and the environment.
The planning application at Chapelcross has attracted thousands of objections through a campaign portal established by opponents. That demonstrates the strength of concern, but the number of objections does not by itself determine whether planning permission should be granted.
The central issue is whether Scotland has a sufficiently clear strategy for deciding which large data centres it wants, where they should be located and what conditions they should meet.
A national approach should consider electricity demand, renewable generation, grid capacity, environmental effects, water use, jobs and the economic value created for Scotland.
Without that wider picture, there is a risk that each proposal will be considered mainly on its own promises, while the cumulative effect of several developments receives too little attention.
Is AI worth this much investment?
Artificial intelligence is creating demand for computing power on a scale that would have seemed extraordinary only a few years ago. Training and operating advanced AI systems requires specialised computer chips, extensive data infrastructure and large amounts of electricity.
Countries and technology companies are investing heavily because they expect AI to change industries ranging from medicine and engineering to finance, manufacturing and public services.
Scotland has a choice. It can try to attract part of that investment, develop expertise and create skilled jobs, or risk seeing much of the infrastructure and economic activity located elsewhere.
But attracting investment should not mean accepting every proposal on whatever terms a developer offers.
A successful project should demonstrate that it can be financed, supplied with reliable electricity, operated within environmental requirements and deliver benefits that justify the resources it uses.
There is also a wider question about the kind of economy Scotland wants. Does it want to become a location for large computing facilities, or should it put greater emphasis on attracting businesses that use AI to develop products, services and jobs across the wider economy?
Those approaches are not mutually exclusive. A data centre can provide infrastructure for other industries. But the public should be able to understand how a major facility fits into the country's economic strategy.
The lesson for the Highlands
For people in Caithness and the wider Highlands, Chapelcross offers a familiar lesson about major infrastructure projects.
Large investment figures attract attention, but the local questions are always more specific. How many jobs will be available to local people? How long will they last? Will young people be able to train for them? Will local suppliers benefit? What infrastructure is required, and who pays for it?
The same questions apply to major energy projects in the Highlands, where national ambitions for electricity generation and industrial development do not always translate automatically into permanent local employment.
Chapelcross could become a significant new economic centre for Dumfries and Galloway. But its success should not be measured by the size of the investment headline alone.
The public needs evidence about the jobs, the electricity arrangements, the environmental consequences and the financial commitments behind the proposal.
A gamble that needs scrutiny
The Chapelcross proposal presents Scotland with a genuine opportunity. It could give a former nuclear site a new economic purpose, attract major investment, support skilled employment and establish a foothold in a rapidly expanding industry.
It also presents substantial questions about electricity demand, environmental impact, land use and the difference between promised investment and benefits that will actually reach local communities.
The right response is neither to dismiss AI nor to assume that a £16.5 billion headline automatically makes a project worthwhile.
Before the scheme proceeds, the developer should demonstrate how the electricity will be supplied, how environmental effects will be managed, how the development will be financed and what lasting benefits local people can reasonably expect.
The planning authorities and Scottish Government should make sure those answers are independently scrutinised and that the project is assessed against a coherent national strategy for data centres and energy.
Scotland needs investment and good jobs. It also needs a reliable electricity system, responsible land use and public confidence that major developments serve more than the interests of their owners.
The real test at Chapelcross is not whether £16.5 billion sounds impressive. It is whether the project can deliver lasting economic value without imposing unacceptable costs on Scotland's energy system, environment or communities.