As AI guzzles water and energy, we are already facing a choice: datacentres or homes? | John Harris
The rapid expansion of datacentres, fueled by AI, is creating a conflict between technological growth and local communities, straining resources like water and energy.
Intelligence analysis by Gemini 2.5 Flash
The UK town of Slough exemplifies a growing global dilemma where the proliferation of datacentres, essential for AI, consumes vast amounts of water and electricity, leading to local environmental impacts and sparking significant policy debates and community resistance in places like the US.
Imagine giant computer buildings that help make AI smart, like a super-brain. These buildings get really hot and thirsty, needing tons of electricity and water to stay cool. In places like Slough, England, there are so many of them that they're making the air warmer and using up water that people and homes also need, causing a big problem for everyone.
Analysis
Slough
The town of Slough, once known for the TV comedy "The Office," has become a focal point for the rapid expansion of datacentres in the UK. With as many as 40 such facilities, and more planned, Slough hosts the densest cluster of datacentres in western Europe. This concentration is attributed to its proximity to London, its location atop critical transatlantic data cables, and its history as a partly deregulated "simplified planning zone."
Local residents in Slough are experiencing direct impacts from this industrial growth. Accounts from people like Sheryl describe a constant "whirring, buzzing sound" at night and a noticeable increase in local temperatures. Early research, including a preprint from Cambridge University, suggests that datacentres can create a "heat island" effect, raising immediate vicinity temperatures by an average of 2C, and potentially up to 9C, though corporate owners of the trading estate dispute conclusive evidence.
Equinix
One prominent example of these facilities is a datacentre owned by the California-based company Equinix, which operates with minimal corporate branding and a constant industrial whirr. The sheer scale of operations within these centres, driven by the demands of artificial intelligence, necessitates immense computing power and, consequently, significant cooling systems. These systems are major consumers of both electricity and water, linking datacentre expansion directly to climate change concerns and local resource strain.
The resource demands are particularly stark in the context of ongoing environmental challenges, such as droughts and heatwaves. A local resident, Louise, highlighted the public's growing awareness of the vast amounts of water consumed by these "AI bits," expressing frustration at the perceived waste. This public sentiment underscores a growing tension between the technological advancements powered by datacentres and the environmental costs borne by local communities.
New York
The issues observed in Slough are not isolated, with similar conflicts emerging globally, particularly in the United States. The debate surrounding datacentre expansion has entered the US political mainstream, with states and local communities beginning to push back against unchecked growth. New York, for instance, has implemented a one-year pause on new "hyperscale" datacentres, signaling a shift towards more regulated development.
This resistance is gaining bipartisan traction, with figures like Michigan's Abdul El-Sayed articulating strong public opposition and even the Republican governor of Texas calling for development bans in rural areas. Texas has also mandated regulatory review of water and energy use before new centres can connect to the power grid. These policy responses reflect a growing recognition that the economic benefits of datacentres must be balanced against their environmental footprint and impact on existing infrastructure and communities.
Key points
- Slough, UK, hosts Western Europe's densest cluster of datacentres, with over 40 facilities and more planned.
- Datacentres, essential for AI, consume vast amounts of electricity and water for cooling, linking them to climate change and local resource strain.
- Local residents report increased noise and a 'heat island' effect, with early research suggesting datacentres can raise local temperatures by 2-9C.
- The issue is sparking political resistance in the US, with New York imposing a pause on new 'hyperscale' datacentres and Texas mandating regulatory reviews.
- Many datacentre insiders argue that renewables alone cannot meet the 100% uptime requirement, leading some to plan for dedicated gas turbines.
Growing interest in 'green datacentres' suggests a potential for innovation in more sustainable cooling and energy solutions. Policy interventions, like those seen in New York and Texas, could lead to better regulation and more responsible development practices, balancing technological progress with environmental stewardship.
The unchecked proliferation of datacentres, driven by AI's demands, risks exacerbating resource scarcity, particularly water and energy, in already strained regions. This could lead to increased local temperatures, noise pollution, and significant pressure on existing power grids, potentially forcing difficult choices between industrial growth and community needs.
