The scaling of AI threatens the UK with a water crisis: data centers consume resources equivalent to entire cities
The rapid development of artificial intelligence in the UK is facing an unexpected but critical obstacle—an acute shortage of water resources. An analysis conducted by the industry organization Water UK reveals a systemic flaw in government planning: official water consumption forecasts have proven to be "fatally flawed" as they completely ignore the needs of data centers.
Modern data centers are not only giant consumers of electricity but also veritable "water monsters." Huge volumes of water are required for cooling systems to dissipate heat from densely packed server racks. Both cooling towers and other systems directly use water, and with each new project, the strain on the water supply network grows exponentially.
The situation is particularly alarming in the Slough region, which has the highest concentration of data centers in the country. Affinity Water, the company servicing this area, reported that about 125 data centers are either already under construction or in the planning stages. Some of them are requesting up to 3 million liters of water per day—a volume comparable to the peak consumption of a small town with 3,500 households. Notably, a hose pipe ban is already in effect across most of this territory, vividly illustrating the scale of the shortage.
The problem is compounded by current regulations from the Environment Agency, which require data centers to use exclusively drinking water. This currently leads to a daily consumption of about 6.6 million liters. However, given the government's plans to triple computing capacity by 2030, this figure could skyrocket to 19.8 million liters per day. Water UK emphasizes that without a radical policy overhaul, any further growth in the number of data centers in water-scarce regions "may prove impossible."
According to experts, the only solutions are either to increase water extraction from the environment, risking environmental harm, or to build new infrastructure, such as reservoirs. However, planning and constructing such facilities could take up to 15 years, making this approach extremely slow in the face of rapid AI development.
Expert opinion: The situation in the UK is just the tip of the iceberg. We are witnessing a global trend where the environmental consequences of the "green" digital revolution are being ignored in favor of growth rates. The AI industry urgently needs to invest in closed-loop water supply technologies and alternative cooling methods; otherwise, the water crisis will become as serious a constraint on development as the chip shortage.