As artificial intelligence continues to expand, tech giants are racing to build the infrastructure that powers AI applications. AI data centres, which host powerful servers running complex machine-learning models, consume massive amounts of energy and produce substantial heat. Recent research shows that these centres not only strain electricity and water resources but also raise local land temperatures, creating a phenomenon dubbed the “data heat island effect.”
Energy-Hungry Infrastructure
Every AI request, whether to ChatGPT, Gemini, or Claude, relies on a data centre filled with specialized servers that operate around the clock. Unlike traditional servers, AI data centres use high-performance chips to perform thousands of calculations in parallel. This results in energy consumption far exceeding that of typical web servers.
According to the International Energy Agency, data centres consumed approximately 415 terawatt-hours of electricity in 2024, around 1.5% of global supply, and usage is projected to nearly double to 945 TWh by 2030. Hyperscale AI data centres, the largest facilities of their kind, often require 100 to 300 megawatts of power continuously, enough to power hundreds of thousands of homes.
Water Usage and Cooling Requirements
Managing the enormous heat generated by these centres requires advanced cooling systems that consume vast quantities of water. A single 100-megawatt hyperscale data centre can use up to 2.5 billion litres of water per year, equivalent to the annual needs of 80,000 people. This demonstrates the significant environmental footprint associated with AI infrastructure.
Global Distribution of AI Data Centres
The construction of AI data centres is accelerating worldwide. As of June 2026, there are over 11,600 active data centres globally, with the United States hosting the largest number—more than 4,300. Europe follows, led by the United Kingdom, Germany, and France, while Asia’s largest clusters are in China and India. Southeast Asia is emerging as a rapidly growing market for data-centre capacity.
Hyperscale data centres have nearly doubled since 2021, from 700 to 1,297 facilities. Major upcoming projects include Meta’s $27 billion Hyperion campus in Louisiana, Microsoft’s $20 billion data centre expansion in Wisconsin, Amazon’s $25 billion investment in Mississippi, Google’s $15 billion Project Spade in Missouri, and Oracle’s Project Stargate in Texas.
Data Heat Island Effect
Cambridge-led research shows that AI data centres raise land surface temperatures around them by an average of 2°C, with some areas recording increases of up to 9°C. Using NASA satellite data, researchers mapped over 6,700 centres outside densely populated areas, finding that the warming effect can extend up to 10 kilometres from the facility.
This localized warming mirrors the urban heat island effect, where concentrated human activity causes cities to become warmer than surrounding rural areas. More than 340 million people live within 10 kilometres of AI data centres and may experience higher temperatures that affect health, energy demand, and overall wellbeing.
Sustainability Challenges
The rapid expansion of AI data centres presents environmental challenges. Policymakers, developers, and communities need to consider energy consumption, water usage, and the heat impact when planning new facilities. Sustainable cooling systems, renewable energy sources, and careful site selection will be essential to mitigate the data heat island effect.
AI data centres are critical to powering machine learning and cloud computing, but their environmental footprint is growing. Understanding and addressing these impacts is key to building a sustainable AI ecosystem that balances technological advancement with climate and community considerations.
