The UK’s ambitious AI growth plans are hitting a wall—and it’s not just about computing power or energy. It’s about water. Yes, water. Personally, I think this is one of the most overlooked yet critical challenges of our tech-driven era. While the government touts AI as the future, the water industry is sounding the alarm: there simply isn’t enough water to sustain the datacentres that power this vision. What makes this particularly fascinating is how it exposes the disconnect between technological ambition and environmental reality.
Datacentres, the backbone of AI and cloud computing, are water guzzlers. Cooling those rows of servers requires millions of litres of water daily—often drinking water. From my perspective, this raises a deeper question: why are we prioritizing chatbots and data storage over basic human needs in water-stressed regions? The UK’s south and east, already under hosepipe bans, are home to over three-quarters of the country’s datacentres. One thing that immediately stands out is the absurdity of this mismatch.
Water UK’s critique of the government’s forecasts is scathing. They call them ‘fatally flawed’ for ignoring datacentre water demands. What many people don’t realize is that these forecasts are built around household consumption, not the industrial-scale needs of tech giants. Affinity Water, for instance, revealed that some datacentres are requesting up to 3 million litres of water a day—equivalent to the peak demand of 3,500 homes. If you take a step back and think about it, this is a staggering imbalance.
The government’s response? A vague promise to ‘protect our water supply’ while building nine new reservoirs. But here’s the kicker: reservoirs take 15 years to plan and build. By then, the UK’s water shortfall could be catastrophic. A detail that I find especially interesting is the lack of a drought hierarchy—a clear plan for who gets water when supplies run low. Without it, we’re left with a societal trade-off: households or datacentres?
What this really suggests is that we’re sleepwalking into a crisis. The Environment Agency’s permitting rules force datacentres to use drinking water, while Greenpeace highlights that water companies leak five times more water than a hosepipe ban could save. It’s a double failure: inefficient infrastructure and misguided policy. In my opinion, this isn’t just a water issue—it’s a governance issue.
The designation of datacentres as ‘critical national infrastructure’ adds another layer of complexity. Oliver Hayes from Global Action Plan raises a valid concern: are we prioritizing AI-driven frivolities over essential services? I think it is highly likely that we’re heading toward a future where datacentres running chatbots are deemed more important than households during droughts.
This isn’t just a UK problem. Ireland is grappling with similar issues, and globally, the water-tech nexus is becoming a flashpoint. What makes this trend particularly alarming is its invisibility. Most people don’t connect their smart devices to water scarcity, but every search query, every AI interaction, has a hidden hydrological cost.
If we’re serious about sustainable growth, we need a radical rethink. Why not mandate datacentres to use recycled water or locate them in water-abundant regions? Why not invest in cooling technologies that minimize water use? The government’s £75m campaign to reduce household water use feels like a Band-Aid on a bullet wound.
In the end, this isn’t just about water—it’s about priorities. Are we building a future that serves people or algorithms? Personally, I think the UK’s datacentre plans are a wake-up call. We can’t innovate our way out of environmental limits. It’s time to stop treating water as an infinite resource and start treating it as the lifeline it is.
