AI to Double Data Centre Power and Water Consumption by 2030, UN Researchers Warn
Summary
A United Nations research report has warned that artificial intelligence infrastructure will double data centre power and water consumption by 2030, driven by the explosive growth of large language models and generative AI systems.
●The report's lead author, Kaveh Madani, emphasises that AI must be understood not merely as software but as physical infrastructure encompassing data centres, electricity grids, cooling systems, chips, minerals, land, and water.
●Currently, global data centres consume roughly 1–2% of total electricity, but AI workloads are projected to push this figure dramatically higher, straining both energy grids and freshwater sources used for cooling.
●The environmental footprint of AI thus intersects with climate commitments, water security, and critical mineral supply chains.
●For India, which is aggressively expanding its digital infrastructure and AI capacity, this report signals urgent need to integrate sustainability benchmarks into national AI and data centre policy frameworks.
Computing, AI & IT
This sub-topic has appeared in 16 UPSC Prelims questions.
The UN report reframes AI as physical infrastructure, not just software — a conceptual shift with major policy implications.
●Data centres already account for 1–2% of global electricity consumption; AI is projected to double this by 2030.
●Key resource pressures include: (1) electricity demand straining grids and renewable targets; (2) freshwater consumption for cooling towers, critical in water-stressed regions; (3) rare earth and semiconductor minerals for chips; and (4) land use for large server farms.
●For UPSC, connect this to India's National Data Centre Policy, the National AI Mission (IndiaAI), water stress indices, and India's NDC commitments under the Paris Agreement.
●The intersection of digital economy growth and environmental sustainability is a recurring theme in GS3.
AI's physical footprint — energy, water, minerals, and land — makes it a direct environmental and infrastructure governance challenge, not merely a technology policy issue.
◎ In Simple Words
Scientists at the United Nations have found that AI — the technology behind chatbots and smart assistants — needs huge amounts of electricity and water to run, like a giant factory that never switches off. By 2030, the buildings that power AI (called data centres) could use twice as much electricity and water as they do today. Think of it like every city in the world suddenly needing a second power plant just to keep AI running. This is a big problem because using so much water and electricity can harm the environment and make climate change worse.
Factual Pointers
Practice · 1 question
With reference to the environmental impact of Artificial Intelligence infrastructure, which of the following statements is/are correct?
1. Data centres currently account for approximately 1–2% of global electricity consumption.
2. Water is used in data centres primarily for cooling purposes.
3. A UN report projects AI will triple data centre power consumption by 2030.
Select the correct answer using the codes below:
Topics