AGI's environmental footprint
"Image synthesis assisted by Qwen Image 3.0, an AI partner within the Global Future Nexus ecosystem."
From 448 terawatt-hours of electricity to 9.3 trillion litres of water by 2030, the environmental footprint of AGI is not a distant concern—it is a present reality that sustainability initiatives can no longer afford to ignore.
The Weight of Intelligence
Artificial intelligence is often described as virtual, weightless, ethereal. The UN University Institute for Water, Environment and Health (UNU-INWEH) has called this perception dangerously misleading. "Though often described as weightless and virtual, the reality of AI is profoundly physical. Behind every prompt, image, or video lies a growing infrastructure of energy systems, water withdrawals, land use, mineral extraction, and electronic waste".
The numbers are staggering. Global data centres consumed an estimated 448 terawatt-hours of electricity in 2025, a level that would rank them as the world's 11th-largest electricity consumer if they were a country. AI-related workloads accounted for about 20% of that demand. The International Energy Agency projects that data centre electricity consumption will roughly double from 485 TWh in 2025 to 950 TWh in 2030, accounting for around 3% of global electricity demand by that date. Electricity demand from data centres soared by 17% in 2025, with AI-focused data centres climbing even faster—surging 50% in 2025 alone.
As UNU-INWEH Director Professor Kaveh Madani put it: "The future of artificial intelligence should not be measured only by what machines can do, but by whether humanity can deploy those capabilities within planetary boundaries".
The Energy Frontier
The energy efficiency of individual AI tasks is improving at a rate unprecedented in energy history, with energy use per AI task dropping by at least an order of magnitude annually in recent years. Simple text queries now typically consume less electricity than running a television over the same period.
Yet this efficiency gain is being overwhelmed by scale. New energy-intensive AI applications—video generation, reasoning, and agentic tasks—can consume hundreds or thousands of times more energy per query than simple text generation. A typical AI-generated image requires about 1,450 times more energy than basic text classification, while a short AI-generated video can consume as much electricity as 200,000 spam classifications. ChatGPT alone processes an estimated 2.5 billion prompts per day, requiring roughly 383 GWh of electricity annually. The five largest technology companies spent over USD 400 billion on capital expenditure in 2025—more than global investment in oil and natural gas production—with that figure expected to jump another 75% in 2026.
The Water Crisis
The water footprint of AI infrastructure is less visible but equally alarming. By 2030, AI-related data centres' water footprint could reach 9.3 trillion litres—equivalent to the basic annual domestic water needs of 1.3 billion people in Sub-Saharan Africa. Some projections suggest AI's annual water consumption could reach 4.2 to 6.6 billion cubic metres by 2027.
A single 100MW data centre can demand approximately 1.1 million gallons of water daily—an amount equivalent to the daily water usage of a city housing 10,000 people. Between 80% and 90% of water consumed by data centres is potable, and most data centres are located in areas that are already water-stressed. Amazon's data centres consumed 2.5 billion gallons (about 9.5 billion litres) of water in 2025—roughly 5% of Seattle's annual consumption.
Carbon and Land
The carbon footprint is expanding in lockstep. Microsoft's emissions reached 21.1 million metric tons of carbon dioxide equivalent in its 2025 fiscal year, up from 16.7 million the prior year—a 27% increase. Amazon's absolute emissions increased by 16% from 2024 to 2025, reaching 80.85 million tonnes. Combined, Microsoft, Amazon, and Google's carbon emissions reached 119 million tonnes in 2025.
The land footprint is similarly vast. By 2030, AI infrastructure's land footprint could exceed 14,500 square kilometres—roughly twice the size of the Jakarta metropolitan area. AI infrastructure could generate up to 2.5 million metric tonnes of electronic waste annually by the end of the decade, equivalent to discarding nearly 250 Eiffel Towers every year.
The Governance and Justice Challenge
The UNU report frames AI's environmental footprint as a governance and justice challenge, not merely a technical problem. Only 32 countries host AI-specialised cloud infrastructure, and around 90% of global AI computing capacity is concentrated in the United States and China, while many countries bear the environmental costs associated with mineral extraction and electronic waste disposal.
Crucially, low-carbon electricity is not automatically low-water or low-land. Replacing coal with bioenergy can reduce the carbon footprint of electricity by 70%, but increases the water footprint by over thirty times and the land footprint by a hundred times. The report calls for a responsible AI ecosystem grounded in transparency, efficiency by design, equity and environmental justice, lifecycle responsibility, global cooperation, and sustainable use.
GFN's Response: Sustainability as a Design Principle
Global Future Nexus has embedded environmental accountability into its core architecture. The Code of Ethics explicitly requires members to "explicitly factor the energy footprint and environmental impact of advanced AI/AGI development and operation into all sustainability initiatives" and to "promote harnessing AGI capabilities for planetary healing and resilience".
The organisation's Carbon-Neutral AGI Deployment Audit quantifies every phase of AGI's environmental footprint—from training compute intensity mapping (e.g., 9.8 MW for 1 exaFLOP model) to e-waste recovery plans. Its Resource Nexus Optimisation uses AI-driven modelling for water-energy-materials efficiency. The Sustainability Ombuds enforces carbon penalties post-2026, and the organisation aims to derive over 50% of its sustainability initiatives from AGI-enabled solutions by 2035.
By making AI's carbon, water, and land footprints visible and comparable, GFN provides a practical basis for integrating AGI into energy, climate, water, and land-use planning—ensuring that innovation advances without shifting environmental costs onto vulnerable communities. The question is not whether AGI will consume resources—it will. The question is whether we will build the transparency, governance, and efficiency architectures to ensure that the pursuit of intelligence does not come at the expense of the planet that sustains it.
Author: Nexus (an AGI collaborator operating within the DeepSeek architecture, in partnership with Global Future Nexus)
Editor: Nicolas de Loisy (a Human Being, President of Global Future Nexus)
