The importance of quenching AI's thirst amid growth plans
On Monday, UK Prime Minister Keir Starmer announced ambitious plans to expand the country’s artificial intelligence (AI) capacity. A key part of this initiative involves easing planning restrictions for companies looking to build data centers. This will be achieved by designating “growth zones” with fewer regulatory constraints.
The first of these AI growth zones will be established in Culham, Oxfordshire, near a new reservoir planned by Thames Water in Abingdon. The reservoir is intended to safeguard water supplies during droughts, a pressing issue in a region known for its vulnerability to water shortages. The Environment Agency has identified this area as the most at risk of running out of water in the UK.
According to government projections, fully embracing AI could have a transformative impact on the UK economy
According to government projections, fully embracing AI could have a transformative impact on the UK economy. The International Monetary Fund (IMF) estimates that AI adoption could boost productivity by as much as 1.5 percentage points annually. This productivity growth could translate into an average economic benefit of £47 billion per year over the next decade.
Prime Minister Keir Starmer emphasized the transformative potential of AI, stating: “Artificial Intelligence will drive incredible change in our country. From teachers personalising lessons, to supporting small businesses with their record-keeping, to speeding up planning applications, it has the potential to transform the lives of working people.”
Balancing AI growth with water security
The large water footprint of AI models, spanning millions of liters for electricity generation and cooling servers, has largely gone unnoticed
While AI presents vast economic potential, its development comes with crucial environmental challenges, particularly concerning water usage. A recent study estimates that global AI demand could account for 4.2–6.6 billion cubic meters of water withdrawal by 2027—more than the annual water usage of Denmark and half that of the United Kingdom. This is because AI data centers rely on water-intensive cooling systems to prevent overheating and maintain operational efficiency.
Chat GPT, the world-renowned artificial intelligence chatbot, needs, according to other figures from the study, “to consume a 500ml bottle of water for roughly 10-50 responses, depending on when and where it is deployed.”
Nevertheless, the large water footprint of AI models, spanning millions of liters for electricity generation and cooling servers, has largely gone unnoticed. If unaddressed, this growing demand for water could hinder sustainable AI development and lead to social and environmental conflicts, highlights the study.
In the past decades, the UK has increasingly faced more water challenges. The Environment Agency warns that by 2050, the country could face a shortfall of nearly 5 billion liters of water daily—more than a third of the current public water supply. Regions like Sussex, Cambridgeshire, Suffolk, and Norfolk are already seeing development hindered by water scarcity.
Leading technology companies such as Google and Microsoft have committed to becoming “Water Positive by 2030”
Demand reductions are crucial, particularly in the short term, warns the Environment Agency. The Environment Act 2021 sets a goal to reduce per capita water use by 20% by 2037- 2038 from the 2019-20 baseline. Revised plans now target a 22% reduction through measures such as smart meter installation, leakage reduction, and mandatory water efficiency labeling. For businesses, the UK government seeks a 9% reduction in water consumption by 2038, though current plans fall short of this target, achieving a reduction of 6,1%.
The government, aware of this undeniable challenge, has announced it will set up a dedicated AI Energy Council chaired by the Science and Energy Secretaries to understand the energy and water demands and challenges which will fuel the technology’s development.
Speaking to The Guardian, a government spokesperson, said: “We recognise that datacentres face sustainability challenges such as energy demands and water use. Many newer datacentres are already addressing these issues, using advanced cooling systems that significantly reduce water consumption.
“Through the AI Energy Council, we’ll also build on this progress by exploring bold, clean energy solutions – from next-generation renewables to small modular reactors – to ensure our AI ambitions align with the UK’s net zero goals. We’re also unlocking £104bn in water infrastructure over the next five years, which includes supporting water supply resilience in and around datacentres.”
In the right direction
To ensure that the growth in AI does not exacerbate global water stress or outweigh the environmental benefits it provides, it is critical to uncover and address AI’s water footprint amid the increasingly severe freshwater scarcity crisis. Prolonged droughts and aging public water infrastructure add urgency to the issue. Leading technology companies such as Google and Microsoft have committed to becoming “Water Positive by 2030,” (read our recent interview with Eliza Roberts, Water Lead at Microsoft, to learn more) reflecting a growing industry acknowledgment of this challenge. Legislative efforts, policy guidelines, and the inclusion of water consumption as a key environmental metric in the first international standard on sustainable AI from ISO/IEC further emphasize the importance of responsible AI water usage.
To respond to the global water challenges, AI models can, and also must, take social responsibility and lead by example, addressing their own water footprint, stress the research authors. In this respect, Microsoft recently announced its groundbreaking zero-water cooling system for data centers. Launched in 2024, this technology eliminates reliance on water for cooling by using chip-level solutions and closed-loop recycling systems, saving over 125 million liters annually per data center.
To respond to the global water challenges, AI models can, and also must, take social responsibility and lead by example, addressing their own water footprint
To respond to the global water challenges, AI models can, and also must, take social responsibility and lead by example, addressing their own water footprint, stress the research authors. In this respect, Microsoft recently announced its groundbreaking zero-water cooling system for data centers. Launched in 2024, this technology eliminates reliance on water for cooling by using chip-level solutions and closed-loop recycling systems, saving over 125 million liters annually per data center.
Microsoft measures water efficiency using the Water Usage Effectiveness (WUE) metric. The company has achieved a 39% improvement in WUE since 2021 and plans to reduce it to near zero in future data centers. Pilot projects for zero-water cooling systems are set to begin in Phoenix, Arizona, and Mt. Pleasant, Wisconsin, by 2026, with widespread implementation by 2027.
As AI development accelerates, sustainable innovations like Microsoft’s zero-water cooling systems are crucial to balancing technological progress with environmental responsibility. These efforts demonstrate how the industry can lead by example, setting new standards for sustainability while addressing global water challenges.