The United Kingdom is grappling with a looming water shortage crisis exacerbated by climate change and a burgeoning population. A significant factor contributing to this problem is the immense water usage by data centres, which are pivotal in supporting the country's expanding AI ambitions. As these facilities continue to proliferate, they are drawing increasing scrutiny over their environmental impact. Recent data reveals that a typical data centre consumes between 11 million and 19 million litres of water per day—a volume comparable to the daily consumption of a town with 30,000 to 50,000 residents.
Microsoft's global water consumption spiked by 34% during the development of its initial AI tools. This trend exemplifies the substantial water demands associated with advancing artificial intelligence technologies. In Iowa, a cluster of data centres consumed 6% of the district's water supply in just one month while training OpenAI's GPT-4. These figures highlight the growing tension between technological progress and sustainable resource management.
In response to this challenge, the UK government is backing plans for nine new reservoirs aimed at mitigating water shortages and preventing rationing and hosepipe bans during droughts. Thames Water has engaged in discussions with the government about addressing the water demand challenge posed by data centres. In a notable move, Thames Water objected to a planning application for a new data centre in Slough, near Reading, in 2021, citing concerns about water usage.
Foxglove, a group of campaigning lawyers, uncovered evidence from Thames Water's strategy documents indicating that the utility company lacked precise knowledge of how much water its existing data centre customers were using. Foxglove's CEO, Martha Dark, announced that her company is trialling a new AI tool designed to analyse operational data and identify potential water-saving measures.
Digital Realty's vice president of sustainability, Aaron Binckley, acknowledged the criticism surrounding data centres' water consumption but asserted that the industry is making "significant strides" toward sustainability. Binckley emphasized that "AI can drive sustainability as well as technological progress," highlighting the sector's efforts to balance innovation with environmental responsibility.
Dr Venkatesh Uddameri, an expert in water resources management based in Texas, confirmed that a typical data centre can use between 11 million and 19 million litres of water per day. This aligns with the findings of the Royal Academy of Engineering, which has called on the government to ensure tech companies accurately report their energy and water usage.
To address these concerns, closed-loop cooling systems that reuse water will be implemented in Microsoft's new data centres in Phoenix and Wisconsin. This approach represents a promising step toward reducing the environmental footprint of such facilities. However, the scale of the challenge remains daunting. In 2024, the Environment Agency warned that by 2050, England alone would require an additional five billion litres of water every day—equivalent to filling two million wheelie bins—just to meet the needs of its population.
A government spokesperson acknowledged the sustainability challenges faced by data centres, stating:
"We recognise that data centres face sustainability challenges such as energy demands and water use – that's why AI Growth Zones are designed to attract investment in areas where existing energy and water infrastructure is already in place."
This statement underscores the government's commitment to strategically managing resource allocation to support technological growth while minimizing environmental impact.
Meanwhile, a source within the industry stressed that:
"Meeting the increased water demand is not just for the water industry to solve."
This highlights the need for collaborative efforts among various stakeholders to address the complexities of water management as technological advancements continue.
Prof Tom Rodden cautioned about the potential long-term consequences of unchecked AI development:
"we face a real risk that our development, deployment and use of AI could do irreparable damage to the environment."
His warning serves as a stark reminder of the delicate balance between pursuing innovation and maintaining ecological sustainability.