The AI Revolution Has a Cost. Who Is Going to Pay for It?

Artificial intelligence is being presented as the next great technological revolution, promising to transform everything from healthcare and education to business, entertainment and the way we live our everyday lives. Technology companies are investing extraordinary amounts of money into developing increasingly powerful systems, while governments compete to attract the infrastructure needed to support them. It is an exciting prospect, particularly when we consider what AI might eventually make possible, but there is another side to this revolution that receives considerably less attention.

In Lexington, Kentucky, a debate over data-centre development has highlighted what happens when ambitious technological projects collide with the concerns of ordinary communities. A video circulating online, featuring a performed reading of remarks attributed to local campaigner Katie Dapkus, challenges the assumption that residents should simply welcome large technology investments without questioning their consequences. The concerns include electricity consumption, environmental pressures, household costs and whether the people living around these developments have enough influence over decisions that could affect their communities for decades.

The disagreement has already produced a significant response. In October 2026, Lexington's Urban County Council extended a temporary moratorium on data-centre development until October 2027, providing additional time to examine planning rules and potential impacts. The decision does not represent a permanent rejection of the technology, but it demonstrates how the expansion of artificial intelligence is beginning to generate questions that cannot be answered by simply promising economic growth.

The cloud isn't quite as invisible as we imagine

One of the peculiar things about modern technology is how disconnected it can feel from the physical world. We describe information as being stored in the cloud, access artificial intelligence through websites and applications, and increasingly expect enormous amounts of computing power to be available almost instantly. It is easy to forget that these apparently invisible services depend upon vast facilities filled with computers, networking equipment and cooling systems, all of which require electricity and supporting infrastructure.

Data centres are not new, of course. They have supported banking, telecommunications, healthcare, online shopping and countless other services for decades. However, the rapid expansion of artificial intelligence is creating demand for increasingly powerful computing systems, particularly those used to train and operate advanced AI models. The infrastructure needed to support that growth is becoming one of the technology industry's biggest challenges.

The International Energy Agency's 2026 assessment illustrates the scale involved. It estimates that global data-centre electricity consumption could almost double from approximately 485 terawatt-hours in 2025 to 950 terawatt-hours by 2030. This includes all data centres rather than artificial intelligence alone, although AI-focused facilities are expected to account for a rapidly growing share of demand. The figures are projections rather than certainties, but they demonstrate why electricity supply has become such an important part of the conversation surrounding AI.

Electricity is not the only consideration. Data centres generate substantial heat, requiring cooling systems that can sometimes place additional pressure on local water supplies. The amount of water required varies enormously depending upon the technology used, the location and the design of the facility. Some modern centres use closed-loop cooling systems that significantly reduce their need for additional water, while others can be more resource-intensive.

In fact, the developer involved in Lexington, DartPoints, has described plans for closed-loop, low-water-use cooling at its site. That is an important detail because it demonstrates why individual proposals should be judged on their actual designs and environmental performance rather than assuming every data centre creates exactly the same problems.

Nevertheless, the underlying issue remains. Artificial intelligence might operate in a digital world, but the infrastructure required to support it has very real consequences for the physical one.

Could ordinary households end up paying more?

Perhaps the most immediate concern for British readers is what this expansion could mean for electricity bills.

Britain already faces significant challenges surrounding energy costs and the modernisation of its electricity network. Increasing demand from large data centres could require additional power generation, new grid connections and improvements to existing infrastructure. Those investments cost money, raising questions about how the expense should be divided between developers, electricity companies and ordinary consumers.

The issue is already attracting attention from Britain's energy industry. In October 2026, senior figures including representatives of Ofgem and the National Energy System Operator raised concerns that data-centre expansion could increase pressure on household electricity bills unless developers contribute appropriately towards the infrastructure they require.

That does not mean every new data centre will automatically increase electricity prices. In some locations, placing large electricity users close to underused renewable generation could actually help make the energy system more efficient. The problem depends upon where facilities are built, how much electricity is available and whether the costs of expanding the network are allocated fairly.

It is an important distinction, because there is a considerable difference between a technology company investing billions of pounds into new infrastructure and ordinary households being expected to absorb some of the associated costs through their existing energy bills.

Most people probably wouldn't object to artificial intelligence becoming more powerful simply because the computers need electricity. They might, however, have questions if supporting that expansion meant additional financial pressure at a time when household energy costs are already a significant concern.

Britain wants to become an AI powerhouse

The UK government has made no secret of its ambition to establish Britain as a major centre for artificial intelligence. Its AI Growth Zones programme is designed to encourage investment in data centres by improving access to electricity and supporting the planning process, with the wider aim of creating employment, attracting businesses and strengthening Britain's technological capabilities.

There are genuine potential benefits to this approach. Developing more domestic computing infrastructure could help British businesses access advanced technology, support scientific research and reduce reliance on overseas facilities for particularly sensitive applications. Construction projects can provide employment, while operational facilities require specialist engineering, maintenance, security and technical expertise.

For areas seeking investment and economic opportunities, those possibilities are understandably attractive. However, the number of jobs created during construction will not necessarily reflect the number of permanent roles available once a highly automated facility becomes operational. Nor should the promise of investment automatically settle questions about environmental impact, energy infrastructure or whether local people will experience meaningful improvements in their everyday lives.

The Environment Agency has already acknowledged that the rapid growth of data centres could create additional demands upon Britain's water resources, particularly where new developments arrive before longer-term improvements to supply infrastructure are completed. Its guidance stresses that water availability should be considered early in planning, rather than addressed only after a location has been selected.

That seems particularly important when we consider the pressure already affecting parts of Britain's water network. Building substantial new infrastructure may be economically desirable, but it makes sense to understand its requirements before promising that a particular location can accommodate it.

Shouldn't local communities have a greater say?

Perhaps this is where the debate in Lexington becomes particularly interesting.

The residents raising concerns about data-centre development are not necessarily debating whether artificial intelligence has a future. Instead, they are questioning how the decisions surrounding that future are being made, and whether the people who may experience the consequences have sufficient opportunity to influence them.

It is a question that extends well beyond America. Whenever a major industrial or infrastructure development is proposed, there is often a tension between the economic opportunities being promised and the potential costs borne by the surrounding community. Some residents welcome the investment and employment possibilities, while others worry about noise, environmental pressures, changes to their neighbourhoods or demands upon existing public services.

Both perspectives deserve serious consideration. Dismissing every concern as opposition to progress is no more helpful than assuming every new technology project must be environmentally damaging. The more useful approach is to establish what a development will actually require, which benefits can reasonably be expected and what safeguards are necessary to address potential problems.

For data centres, that could mean clearer information about projected electricity and water use, the infrastructure required to support them, the number and type of jobs likely to be created and the arrangements for meeting additional costs. Importantly, communities also need confidence that commitments made during planning will remain meaningful once construction has finished and the facility begins operating.

After all, technological progress should not require communities to accept promises without evidence.

Progress should benefit more than the companies creating it

There is something genuinely extraordinary about the speed at which artificial intelligence is developing. Tasks that once appeared firmly within the territory of science fiction are increasingly becoming practical applications, while the possibilities for medicine, research, accessibility and countless other industries are enormously exciting.

But perhaps our enthusiasm for what AI can achieve has moved ahead of our willingness to discuss what is required to make it happen.

We are becoming accustomed to hearing about billions being invested, faster processors, more powerful models and the extraordinary opportunities supposedly waiting just around the corner. What receives less attention is the electricity that needs generating, the water that may be required, the infrastructure that must be constructed and the communities living alongside it all.

None of those considerations automatically makes artificial intelligence a bad thing. Every major industrial development throughout history has required resources, infrastructure and difficult decisions about balancing economic opportunity with wider consequences. AI is simply the latest technology forcing us to confront those questions, albeit on an increasingly international scale.

The discussion taking place in Lexington offers a useful reminder that the future is not something that exists entirely within the boardrooms of technology companies or the announcements of politicians. Eventually, technological ambition has to become something physical, built somewhere, connected to existing infrastructure and incorporated into the lives of ordinary people.

Artificial intelligence may well deliver some of the most significant advances of our lifetime, and there are plenty of reasons to welcome that possibility. However, if this really is the beginning of another great industrial revolution, surely one of the lessons we should have learned from previous generations is that economic progress needs to consider the people and places affected along the way.

The question should not simply be how quickly we can build the infrastructure required for AI, but how we can ensure that its benefits are realised without overlooking its costs.

Because when the technology companies promise us a better future, it is reasonable to ask who will be paying to build it.