Environmental Cost of Artificial Intelligence: Carbon, Water, and Land Footprints

Miriam Aczel, Sanaz Chamanara, Mir Matin, Aria Farsi, Tshilidzi Marwala and Kaveh Madani (2026). Environmental Cost of Artificial Intelligence: Carbon, Water, and Land Footprints. United Nations University Institute for Water, Environment and Health (UNU INWEH).

Document type:
Report
Collection:
Metadata
Documents
Links
Versions
Statistics
  • Attached Files (Some files may be inaccessible until you login with your UNU Collections credentials)
    Name Description MIMEType Size Downloads
    UNU-INWEH-Report-The_Env_Cost_of_AI-2026.pdf UNU-INWEH-Report-The_Env_Cost_of_AI-2026.pdf application/pdf 26.36MB
  • Sub-type Research report
    Author Miriam Aczel
    Sanaz Chamanara
    Mir Matin
    Aria Farsi
    Tshilidzi Marwala
    Kaveh Madani
    Title Environmental Cost of Artificial Intelligence: Carbon, Water, and Land Footprints
    Publication Date 2026-06-03
    Place of Publication Richmond Hill, Ontario, Canada
    Publisher United Nations University Institute for Water, Environment and Health (UNU INWEH)
    Pages 56
    Language eng
    Abstract The rapid global expansion of artificial intelligence is creating a new and largely underexamined sustainability challenge: the carbon, water and land footprints of the electricity required to power AI systems. While AI is increasingly presented as essential for innovation, economic growth, scientific discovery and climate action, its development depends on energy-intensive data centers, advanced chips, cooling systems, electricity grids, water resources, land and critical mineral supply chains that support AI hardware. This report shows that AI’s environmental costs depend not only on how much electricity is used, but also on where that electricity is generated and which energy sources power it. In 2025, data centers consumed an estimated 448 TWh of electricity, with AI workloads accounting for around 20% of this demand; by 2030, total data center electricity use could reach 945 TWh, while AI’s share could rise to 40%. Every kilowatt-hour used to train, deploy or operate AI carries carbon, water and land implications, and low-carbon electricity is not automatically low-water or low-land. As larger models, richer media outputs and everyday AI use scale rapidly, environmental burdens risk becoming concentrated in communities already facing water stress, land pressure, energy insecurity and limited governance capacity. Without transparent measurement, efficient design, lifecycle responsibility and environmental justice safeguards, the AI transition risks reproducing the uneven patterns of extraction and burden-shifting that have shaped earlier technological and energy transitions.
    Keyword Artificial Intelligence
    Environment
    water availability
    Energy
    Land and water management
    Copyright Holder United Nations University Institute for Water, Environment and Health (UNU INWEH)
    Copyright Year 2026
    Copyright type All rights reserved
    ISBN 9789280861419
    DOI 10.53328/INR26RMA002
  • This website has been archived and its online forms are no longer available. Please contact us by email at contact@unu.edu.

  • Citation counts
    Google Scholar Search Google Scholar
    Access Statistics: 0 Abstract Views, 0 File Downloads  -  Detailed Statistics
    Created: Tue, 02 Jun 2026, 00:30:21 JST by Mir Matin on behalf of UNU INWEH