Water is often treated as a side issue in digital transformation. But as AI‑driven growth picks up pace, water is becoming a strategic consideration, influencing infrastructure decisions and resilience.
Behind every AI tool sits physical, energy‑intensive infrastructure housed in data centres. Many of these centres rely heavily on water for cooling. It’s been found that data centres in Scotland are using enough tap water to fill 27 million half-litre bottles per year, with it estimated that just 10-50 responses using AI model GPT-3 could consume 500ml of water.*
Why does AI use so much water?
To operate reliably, data centres rely on a range of cooling systems - from evaporative and chilled‑water processes to more advanced liquid‑cooling methods.
AI models run continuously on high‑performance hardware, generating heat at a scale that traditional systems were never designed for. Water remains one of the most efficient ways to manage this heat safely and cost‑effectively.
Cooling, however, is only part of the picture. According to UK sustainability research**, water is used in data centres to keep systems cool and elsewhere to produce the electricity and computer chips that AI relies on. Much of the water used across these processes cannot yet be easily reused, creating further pressure.
While the sector is making progress on efficiency, demand is rising rapidly and national water resource planning has not fully accounted for the pace of AI‑driven infrastructure growth.
UK government‑backed research shows that the rapid expansion of AI is already driving a measurable increase in water demand, often in regions that are already water‑stressed***. In the UK, many proposed data centre developments are planned in areas facing long‑term supply challenges, where water systems built for a different climate and a different economy are being asked to do more, more often.
Why this matters for businesses
Every AI prompt carries a resource cost, even if the end user never sees it.
This isn’t just a consideration for technology companies. Organisations across every sector are increasingly reliant on AI‑enabled tools, cloud services and data‑driven platforms, all of which depend on upstream water and energy systems.
As water scarcity becomes a growing challenge across the UK, businesses need to understand how emerging technologies shape their wider environmental footprint, including indirect or ‘hidden’ resource dependencies.
Failing to consider water use alongside energy introduces reputational, operational and resilience risks, particularly as climate change accelerates and competition for critical resources intensifies.
How businesses can use AI responsibly
Used wisely, AI can still be part of a more sustainable business model but only if organisations understand the wider and often hidden impacts behind its use. Starting with awareness, recognising that digital tools rely on physical resources, and factoring those dependencies into technology choices, procurement decisions and long‑term planning.
Responsibility also means asking harder questions about where digital services come from, how suppliers manage water and energy demand and whether infrastructure planning considers long‑term water availability, not just short‑term productivity gains.
The opportunity lies in being deliberate in applying AI where it delivers genuine value and understanding the challenges that sit behind its use.
Even small actions matter when using AI. Educating teams on smarter, more efficient AI use, including better prompting first time, can reduce unnecessary processing, energy demand and water use.
AI doesn’t have to come at the expense of sustainability or reputation, but the outcomes depend on how thoughtfully it is used.
AI will continue to reshape how businesses operate. The challenge, and opportunity, is to ensure this transformation strengthens resilience rather than placing additional strain on already stretched resources. AI might feel ‘digital’ but it still depends on real‑world resources, and water is one that’s already under pressure.
Sources:
** AI’s thirst for water – UK Government Sustainable ICT
*** 84% of proposed UK data centres to be built in water stressed areas - Water Magazine