AI's Hidden Thirst: Mapping Data Centres' Water Footprint

A 7-week investigation into the unseen environmental impact of our digital infrastructure

While the world marvels at ChatGPT's ability to write poetry and DALL-E's artistic creations, beneath this AI revolution lies a troubling secret: every chat response and generated image comes at a significant environmental cost, draining millions of gallons of water from our increasingly fragile ecosystems – yet somehow, this remains largely absent from public discourse about AI's impact. Through innovative data visualisation and careful analysis, I set out to reveal these hidden connections between US data centres and water consumption, challenging common assumptions about digital infrastructure's environmental footprint.

The Challenge

I investigated the complex relationship between data centres and water resources across the United States, with particular focus on three distinct regions facing unique challenges: Phoenix's water scarcity crisis, Seattle's hydroelectric dependence, and Ashburn's efficiency strategies.

My Approach

In the 7 week project I was able to accomplish the below:

• Developed interactive visualisations using d3.js to map water usage patterns across 58 major US cities

• Created an innovative radial calendar system to display monthly water usage effectiveness intensity

• Designed complementary circular charts showing data centre density and water scarcity footprints

• Built a comprehensive prototype in Figma to demonstrate these relationships

Key Findings

My analysis revealed surprising patterns across three major data center hubs:

• Phoenix: Despite severe water scarcity, the region's heavy investment in solar energy actually results in lower overall water consumption than expected, as energy production typically demands more water than cooling systems

• Seattle: While celebrated for its "green" hydroelectric power, the region's data centers ironically show some of the highest water usage rates due to their dependence on water-intensive hydroelectric energy

• Ashburn (Data Center Alley): Emerged as a model of efficiency by strategically utilising sustainable, non-water-dependent energy sources, demonstrating how thoughtful infrastructure planning can minimise both water and energy impacts

Impact

This project goes beyond simple data visualisation – it's a crucial investigation into the environmental impact of our digital infrastructure. By revealing these hidden relationships between data centres and water resources, I aimed to inform more sustainable practices in tech industry development.

The final prototype demonstrates how thoughtful data visualisation can reveal complex environmental relationships and challenge our assumptions about digital infrastructure's impact on natural resources.

City Level - DC Water Usage

Big Tech Level - DC Water Usage