Datacentres, once a quiet backbone of the digital economy, are now one of the most visible infrastructure points of contention of the AI era. Globally, their growth increases pressure on electricity, water, land use, emissions pathways, and local communities. The IEA reports that electricity demand from datacentres rose 17% in 2025, with AI-focused facilities growing even faster. It also projects that total datacentre electricity consumption will double by 2030 while consumption from AI-focused facilities will triple, despite substantial improvements in computing efficiency.
This expansion carries impacts beyond electricity. United Nations University estimates that AI datacentres could consume 945TWh annually by 2030, their water footprint could equal the basic domestic needs of 1.3 billion people, and their energy-related land footprint could exceed 14,500 square kilometers. Most importantly, as the UN notes, the benefits of AI and cloud infrastructure are globally distributed, while the environmental and social costs are often concentrated in specific host regions.
In the UK, this is no longer a marginal infrastructure issue. Datacentres currently account for around 2.5% of the UK’s electricity consumption, with demand projected to increase fourfold by 2030, while the government has designated datacentres as critical national infrastructure and established AI Growth Zones to accelerate investment. Government analysis suggests that UK datacentre capacity, estimated at approximately 1.6 GW in 2024, could rise to between 3.3 GW and 6.3 GW by 2030, although even that may not meet projected demand.
Cumulative community impacts must be clear before approval
Evidence submitted to the UK’s All-Party Parliamentary Group for Datacentres ranked grid access and energy supply as the industry’s most important constraints, followed by energy costs and planning. However, the community impact is broader than simply power capacity. Host regions need to understand whether a proposed datacentre will increase local grid-reinforcement costs, compete with other users for water, create noise or traffic burdens, require new substations or transmission infrastructure, rely on fossil-backed generation, or deliver meaningful local benefits.
OpenAI’s reported pause of Stargate UK illustrates the consequence of infrastructure constraints. The issue is not simply whether the UK can attract AI capital, but whether it can provide competitively priced electricity, timely grid connections, clean power additions, planning certainty, and public legitimacy. Without those conditions, the UK risks creating a gap between national AI ambition and local infrastructure reality.
US datacentres: A warning about reactive policy
New York and Texas are both slowing datacentre development to protect residents and infrastructure, but by different mechanisms. In July 2026, New York Governor Kathy Hochul imposed a one-year pause on new hyperscale datacentre environmental permits and set about developing rules to protect ratepayers, water resources, and communities. In August 2026, Governor Greg Abbott directed Texas regulators and the Electric Reliability Council of Texas (ERCOT) to conduct a comprehensive audit of datacentre projects before they move forward in the grid-connection process.
Texas shows the scale of the challenge. ERCOT is evaluating more than 474GW of grid-connection requests, more than five times Texas’s record peak electricity demand, with datacentres accounting for roughly 90% of the proposed new load. Governor Abbott’s directive requires datacentre developers to disclose their expected electricity use, water use, cooling technologies, public incentives, on-site generation plans, and community-protection measures such as noise, light, traffic and emergency-response planning.
Lessons for the UK
Rapid, incentive-led growth without transparent cost allocation, consistent environmental standards, and community participation eventually produces resistance, delay and regulatory uncertainty. Once project demand exceeds grid capacity, governments must intervene, often through pauses, audits, or moratoria that create uncertainty for datacentre developers and communities.
The UK faces the same urgency but has a better opportunity to act proactively. Unlike the US, where authority is fragmented across federal agencies, states, counties, municipalities, utilities, and multiple regional electricity markets, the UK has the opportunity to design a more coherent national framework.
For a UK government datacentre community impact framework
The UK needs a national datacentre community impact framework that links approvals to regional grid capacity, additional clean energy generation, water availability, lifecycle emissions, land-use impacts, heat reuse potential, and enforceable community benefits. This would help communities see the full trade-off: what the datacentre consumes, what infrastructure it requires, what costs it shifts, and what value it returns.
In practical terms, the UK government must require datacentre developers to show that projects are not merely “low carbon” on paper, but compatible with constraints and regional economic priorities.
To stem the negative impact from datacentre development, a framework must include five core requirements:
-
Verified demand and grid readiness. Before a datacentre reserves grid capacity or receives planning prioritisation, developers must prove the project is real, deliverable, and compatible with regional grid constraints. This prevents speculative projects from occupying scarce connection capacity and delaying more credible infrastructure. This must include details about projected peak and annual electricity demand, backup power generation, plus grid upgrade requirements and who will pay for those upgrades.
-
Technical community burden and local infrastructure impact measurement. The UK needs to make visible the local burdens that often sit outside headline-making economic development claims. Communities must see proposed facility impacts noise, construction-related environmental concerns, traffic, land use, local infrastructure, emergency services, embodied carbon, and quality of life. Datacentre approvals must include a clear water-impact assessment, especially where projects use evaporative cooling or are located in areas facing water stress.
-
Socio-economic community benefits and economic value measurement. Datacentres are capital intensive but often create fewer permanent jobs than other forms of industrial development. Details must include details of expected jobs and tax incentives.
-
Ratepayer protection. Developers must fund or contribute to the grid, water, transport, and community infrastructure required by their projects, rather than shifting costs onto local households or authorities.
-
Ongoing accountability. Datacentres must report actual performance against approved forecasts, including electricity consumption, water consumption, emissions, uptime, heat reuse, and delivery of community benefits.
The cloud is not an abstract or placeless service. Its infrastructure is physical, its resource demands are concentrated, and its effects are experienced locally. Accountability must therefore extend across operators, utilities, technology customers and policymakers with metrics that are communicated to communities with transparency. Otherwise, public trust will be lost if communities cannot understand the full impact of datacentre development.
Now is the time for the UK government to move from project-by-project approval to a national strategy for responsible datacentre growth. A clear framework gives datacentre developers predictable rules, gives communities credible protections, and ensures that datacentre expansion delivers strategic national value without imposing disproportionate costs on the places that host it.
Abhijit Sunil is senior analyst with Forrester.
S 004