Stop dirty datacenters': Europe-wide protests planned demanding building moratorium

The supplied article lists datacenter developments, but it does not describe Europe-wide protests or their demands. More generally, campaigns against datacenters often seek slower expansion, stronger environmental reviews, limits on water and electricity use, and clearer benefits for host communities. Their central concern is that construction may move faster than public planning.

A building moratorium is a temporary legal or administrative halt. It can stop new planning approvals, building permits, or construction for a defined period. Authorities may use that time to assess grid capacity, water availability, emissions, noise, land use, and community impact. A moratorium does not necessarily cancel existing facilities or permanently ban datacenters.

The article does mention Drax’s proposed £20B datacenter at a biomass power station. It says the project still needs planning permission, government approval, and a customer. Those requirements show why approvals matter, but they are not evidence of protests or a moratorium. The source provides no Europe-wide campaign demands, so specific claims about them need additional reporting.

The supplied article does not identify countries or cities involved in datacenter protests. It also gives no survey results or measure of opposition to new construction. Therefore, it cannot support a reliable Europe-wide map or percentage. Any precise list or figure would require separate reporting about local campaigns and public opinion.

In general, opposition tends to appear where several projects are planned close together. Residents may object to land conversion, construction traffic, noise, electricity demand, water use, or limited local benefits. Support can also exist, especially where projects promise investment, jobs, tax revenue, or better infrastructure. Public views therefore differ between regions and even between neighbouring communities.

The article’s Drax example illustrates a project still at the approval stage. It needs planning permission, government approval, and a customer before proceeding. That is different from evidence of public opposition. The article also mentions a customer gap, not a protest movement. More detailed claims about countries, cities, or scale should not be inferred from these headlines.

“Dirty” is an environmental criticism, not a technical datacenter category. Campaigners may use it when a facility relies on carbon-intensive electricity, drives new fossil-fuel generation, or causes large emissions during construction and operation. They may also object to land clearing, backup generators, noise, waste heat, and pressure on local water supplies.

The key mechanism is the facility’s total footprint. Servers consume electricity continuously. Cooling systems remove their heat and consume additional power, while some cooling designs use water. Building the site requires concrete, steel, cables, roads, and machinery. AI workloads can intensify the burden because dense accelerator servers draw substantial power and require robust cooling.

The supplied article does not discuss these impacts, so it does not establish which datacenters campaigners mean. It does mention a proposed £20B Drax facility and says approval is pending. Whether that project would be considered “dirty” would depend on its actual energy source, design, emissions, water use, and construction plan, none of which the article provides.

The supplied article gives no electricity, water, or land figures for a large datacenter. As a broad industry guide, a conventional large facility may require roughly 10–100 megawatts of continuous electrical capacity, while very large AI campuses can reach several hundred megawatts. Actual demand depends on servers, cooling, uptime, and expansion plans.

Water use varies even more. Evaporative cooling can consume hundreds of thousands to several million litres per day, while air-cooled or closed-loop designs may use much less operational water. Land needs can range from several hectares to dozens of hectares, including buildings, substations, roads, security areas, and expansion space. AI systems usually pack more high-power chips into each server room.

These are general ranges, not figures from the source. The article only says Drax wants a £20B datacenter and that on-premises HPC and AI server sales are up 20 percent this year. That signals strong computing demand, but it does not reveal the proposed facility’s power, water, or land requirements.

If datacenter construction accelerates faster than infrastructure planning, local grids may face connection queues, congestion, or costly reinforcement. Water systems can experience higher demand, especially where cooling uses evaporation. Communities may also face construction traffic, noise, land-use changes, rising pressure on housing, and disputes over whether promised jobs justify the burden.

The mechanism is continuous demand. Servers operate around the clock, and cooling adds to their electrical load. Grid operators may need new substations, transmission lines, or generation. If additional power comes from fossil fuels, emissions can rise. Water impacts depend on climate, local supplies, cooling design, and whether treated or recycled water is available.

The supplied article does not report actual outages, shortages, emissions, or community impacts. It does show that Drax’s planned £20B facility still needs planning permission, government approval, and a customer. That creates an opportunity to assess infrastructure and environmental effects before construction, rather than treating expansion as automatic.

The supplied article does not propose alternatives for expanding European computing capacity. Common approaches include improving server efficiency, using higher utilization, placing workloads where low-carbon electricity and water are available, and expanding existing sites before building new campuses. Software optimization can also reduce the computing needed for a task.

Cooling choices matter. Air cooling, closed-loop liquid cooling, recycled water, and heat reuse can reduce freshwater demand, although each design has trade-offs. Operators can schedule flexible AI workloads when electricity is cleaner or grids are less congested. Smaller regional facilities may reduce network distance, but they are not automatically greener if they run inefficiently.

The article reports that firms still want on-premises HPC and AI equipment, with related server sales up 20 percent this year. That suggests demand will not disappear. It also reports Drax’s proposed £20B facility, which still lacks approvals and a customer. Planning can therefore compare efficiency, energy sources, water use, and reuse before choosing capacity.

A datacenter receives electricity and routes it to servers, storage systems, networking equipment, lighting, and cooling. Servers use processors and memory to run software, store or retrieve data, and answer online requests. For AI, specialized accelerators perform large numbers of mathematical operations for training or inference. The facility connects those machines to networks and users.

Electricity is not converted into useful computing with perfect efficiency. Resistance and switching inside chips, memory, power supplies, and fans release energy as heat. Dense AI hardware produces especially high heat in a small space. Cooling systems move that heat away using air, liquid, chillers, pumps, or evaporative processes. Cooling equipment itself consumes additional electricity and sometimes water.

The supplied article does not explain this mechanism directly. It does report rising demand for on-premises HPC and AI servers, up 20 percent this year, and Drax’s proposed datacenter. As AI deployments grow, efficient chips, cooling, and power planning become increasingly important to cost and reliability.