Many people enjoy asking AI for help with work, shopping or everyday questions. Far fewer are enthusiastic about the physical infrastructure that makes those services possible appearing near their homes. Stanford researchers report that a majority of surveyed US adults would feel uncomfortable with a new data centre in their area. The contrast behind AI data centre opposition is not necessarily hypocrisy. The benefits of a service can be widely distributed while the costs of building its infrastructure fall heavily on particular communities.
Stanford’s November analysis of the AI boom reports that 60% of US adults surveyed would feel uncomfortable with a new data centre in their area, with concerns crossing political divisions. The figure refers to a particular survey population and wording. It should not be presented as the opinion of all people everywhere, or as evidence that every local development would face identical opposition.
What AI data centre opposition is really about
An AI assistant feels like software. A user types a question and receives an answer on a phone or laptop. Behind that interaction are data centres containing computing equipment, networking systems and facilities that manage electricity and heat. Expansion can require new buildings, power connections and sometimes water-intensive cooling equipment. These are physical choices with local consequences.
A national conversation about advanced AI may focus on innovation, jobs and technological leadership. A resident living near a proposed site may ask whether the development will change the local landscape, traffic, electricity network or water resources. Both perspectives can be legitimate. The public benefit of a new service does not automatically settle how a particular site should be selected or managed.
The International Energy Agency’s 2026 report on powering artificial intelligence examines risks and options for electricity systems facing rapid data-centre investment. It does not say that every facility imposes the same strain. Location, grid capacity, energy supply and the type of cooling system can change the likely effect substantially.
Why electricity is the first question
Large computing facilities require dependable power. In some areas the electricity network has enough capacity, while in others new demand may trigger costly connections or upgrades. The consequences depend on how the costs are allocated, whether generation is added and how the local system manages peaks. It is not accurate to claim that every new data centre automatically raises household bills by a fixed amount.
A developer may announce plans to use renewable energy or build supporting infrastructure. Those commitments should be examined in detail. Buying certificates or signing a long-term electricity agreement is not necessarily the same as ensuring new power is available at the hour and place a facility needs it. Local authorities and residents have reason to ask how the claims translate into physical supply.
LiveAIWire has explored the potential effect of data-centre electricity demand on households. The local planning question is related but distinct: which projects pay for upgrades and who bears the risk if projected demand or promised infrastructure changes?
Water use can vary widely
Cooling equipment is essential because computers produce heat. Some facilities use water in particular cooling processes, while others rely more heavily on different methods. The appropriate design depends on climate, available resources, energy use and operating requirements. A general statement that all data centres consume the same amount of water can be deeply misleading.
Water concerns become more urgent where droughts or competition among users already exist. Residents may want clear information about the type of water used, seasonal demand and the capacity of local systems. Developers should provide figures that match the proposed facility rather than relying on broad averages taken from another country.
Environmental questions also extend beyond operating resources. Construction affects land, materials and transport. The location of electrical infrastructure and backup systems can matter as much to a neighbouring community as the building itself. An informed consultation needs site-specific evidence instead of promotional language.
What communities might gain
New facilities can bring investment, construction work and some long-term technical employment. They may also contribute to local business activity or infrastructure upgrades. These potential benefits deserve consideration, but they should be presented with realistic estimates and clear distinctions between temporary and permanent jobs. A large building project does not necessarily create a large permanent workforce.
Benefits also depend on local policy and agreements. A community may reasonably ask whether a proposed development supports public infrastructure, whether it pays appropriate taxes and whether any promised training or employment opportunities are accessible to residents. A company announcement is not the same as a binding local commitment.
The economic value of the AI services delivered from a data centre may be difficult to connect directly to the neighbourhood hosting it. Users in other regions can enjoy the product while the local community deals with roads, land use or utilities. That distribution helps explain why public support for technology does not automatically imply support for every infrastructure project.
Why the survey should not decide planning policy alone
The Stanford figure highlights concern, but surveys do not replace environmental assessments or local planning procedures. Some people who express discomfort might support a particular proposal after reviewing its design and benefits. Others might oppose it even if a company offers compensation. The process matters because people need a credible opportunity to examine evidence and influence decisions.
Public consultation is also more useful when it begins early. Presenting residents with a nearly final plan and a short list of predicted benefits is unlikely to build confidence. Developers should disclose important assumptions and explain who will verify them. Independent scrutiny may reveal that an alternative site or infrastructure arrangement is more suitable.
LiveAIWire’s earlier article on the growing electricity needs of AI infrastructure provides broader context. The relationship between global demand and local pressure needs careful explanation: rapid industry growth can coincide with very different outcomes for individual communities.
An argument about trust as well as technology
A resident may be less concerned about the idea of a data centre than about whether a developer has been candid about water, power and traffic. Trust improves when forecasts are accompanied by measurable commitments, when uncertainty is acknowledged and when there is a process for handling problems after construction.
The politics of AI infrastructure can also differ from attitudes towards AI products. People who use a service every day may still prefer stricter oversight of the industry that provides it. That is an ordinary position in many sectors. Consumers benefit from transport, food and energy systems while supporting planning limits or safety rules for particular facilities.
The meaningful question is not whether residents should support or oppose AI in general. It is whether a particular development has been assessed fairly, its costs and benefits explained accurately and its risks managed. Community opposition becomes more understandable when those distinctions are ignored.
Stanford’s finding captures the growing visibility of AI’s physical footprint. As more people use the technology, its infrastructure choices will become more local and more contested. The responsible response is transparent evidence and practical planning, not assuming that enthusiasm for an app implies consent to every building needed to operate it.
A proposal should include evidence residents can question
An effective planning consultation gives people something more specific than a prediction of investment. It should explain expected electricity use under different operating conditions, the proposed water system, construction traffic and the arrangements for dealing with unexpected demand. Residents should also be told which facts have been independently assessed and which remain estimates provided by the developer. The quality of the evidence matters because a large facility may operate for many years after its original owners or customers change.
Technology companies can also reduce conflict by comparing locations rather than treating the first available site as inevitable. A design that works in a region with abundant grid capacity may be unsuitable in an area where new electricity demand competes with household and industrial needs. LiveAIWire’s earlier account of the AI energy trap and rising infrastructure demand provides a related perspective. Improving the efficiency of computing does not guarantee that overall demand will fall if the industry expands rapidly.
A local decision should therefore consider alternatives, not just the benefits of one proposal. Public authorities have a role in weighing impacts that no individual consumer or technology provider can resolve alone. Transparent comparison can help communities distinguish a credible development from one that transfers too many costs to its neighbours.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity and the social impact of emerging technology. LiveAIWire is an independent, human-led technology publication using AI-assisted research, editorial production and original AI-assisted editorial illustrations under his direction.
