The rapid expansion of artificial intelligence (AI) has led to a significant increase in the number of data centres, which are essential for processing and storing vast amounts of information. A critical environmental concern associated with these facilities is their substantial water consumption, primarily for cooling purposes. As AI continues to grow, addressing the water usage of data centres becomes imperative to ensure environmental sustainability (The Guardian).

Water Usage in Data Centres
Data centres consume large quantities of water to maintain optimal operating temperatures for their servers. Traditional cooling methods often involve evaporative cooling systems, which can use millions of litres of water annually per facility. For instance, in 2022, Microsoft reported consuming 6.4 million cubic metres of water, while Google's usage was even higher at 19.5 million cubic metres (The Guardian).
The increasing demand for AI services exacerbates this issue. Training large AI models requires significant computational power, leading to higher energy consumption and, consequently, greater cooling needs. This escalation in resource usage raises concerns about the sustainability of current data centre operations (Financial Times).
Sustainability Challenges
The environmental impact of data centres extends beyond energy consumption. Their substantial water usage poses a threat to local water supplies, especially in regions already experiencing water scarcity. In the UK, discussions between water industry leaders and the government have highlighted potential water shortages, with projections indicating that England will require an additional five billion litres of water daily by 2050 (The Guardian).
The concentration of data centres in specific areas can strain local resources. For example, Virginia’s "data centre alley" in the United States saw a nearly 65% increase in water consumption from 2019 to 2023, primarily due to the growth of AI computing infrastructure (Financial Times).
Innovative Cooling Solutions
To mitigate the environmental impact, data centre operators are exploring alternative cooling methods that reduce water consumption. Liquid cooling and immersion cooling techniques have emerged as effective solutions. These methods involve directly cooling server components with liquids, enhancing energy efficiency and significantly decreasing water usage (Data Centre Dynamics).
Companies like Microsoft are taking proactive steps by planning to implement water-free cooling technologies in their upcoming data centres, aiming to manage high-density AI workloads more sustainably (The Guardian).
Future Strategies
Addressing the environmental challenges posed by AI-driven data centres requires a multifaceted approach. Industry experts advocate for mandatory reporting of energy and water consumption by tech companies to promote transparency and accountability. The National Engineering Policy Centre recommends that the UK government enforce such reporting to inform policies aimed at reducing the environmental footprint of data centres (The Guardian).
Additionally, adopting circular water solutions, such as using recycled water for cooling, can contribute to more sustainable operations. Implementing dry cooling systems and optimising AI workloads to be more resource-efficient are also viable strategies to lessen water usage (World Economic Forum).
As AI continues to evolve, it is crucial for data centre operators, policymakers, and technology companies to collaborate in developing and implementing sustainable practices. By prioritising innovative cooling solutions and responsible resource management, the industry can support the growth of AI while minimising its environmental impact.
Addressing Water Scarcity in the UK
To tackle water shortages, the UK government and water companies are implementing several strategies:
Demand Management: The government has set ambitious targets for water companies to reduce leakage by 50% and decrease per capita consumption to 110 litres per person per day by 2050. Current water resources management plans indicate a planned reduction in leakage by 48% and aim for a per capita consumption of 113 litres per person per day by 2050.
Infrastructure Projects: The industry is working on 18 major cross-country projects designed to deliver sufficient water for 10 million people. These initiatives include transferring water from regions with surplus supplies to those experiencing shortages, recycling and reusing water, and developing new reservoirs.
Desalination and water recycling are being explored to secure future water supplies, especially in areas where demand exceeds supply. These methods can alleviate pressure on existing resources and complement developments like reservoirs. Environment Agency Blog
Data Centre Locations in the UK
The UK hosts a significant number of data centres, primarily concentrated in key urban and industrial areas:
London and Surrounding Areas: A substantial number of data centres are located in and around London, benefiting from robust infrastructure and connectivity.
Manchester and Slough: These cities have seen considerable data centre development, offering strategic locations with necessary infrastructure.
West Coast and M4 Corridor: Regions along the west coast and the M4 corridor are emerging as attractive locations for data centres, providing access to energy resources and infrastructure.
As the UK continues to develop its data centre infrastructure, there is a concerted effort to focus on efficiency and sustainability. This includes investments in renewable energy sources and the adoption of innovative technologies to ensure that the expansion of data centres does not exacerbate existing environmental challenges, particularly concerning water usage.
Comentários