More than carbon: cradle-to-grave environmental impacts of GenAI training on the Nvidia A100 GPU

Sophia Falk (), David Ekchajzer (), Thibault Pirson, Etienne Lees-Perasso, Augustin Wattiez, Lisa Biber-Freudenberger, Sasha Luccioni and Aimee van Wynsberghe
Additional contact information
Sophia Falk: Universität Bonn = University of Bonn
David Ekchajzer: IMT-BS - DEFI - Département Data analytics, Économie et Finances - IMT-BS - Institut Mines-Télécom Business School - IMT - Institut Mines-Télécom [Paris], LITEM - Laboratoire en Innovation, Technologies, Economie et Management (EA 7363) - UEVE - Université d'Évry-Val-d'Essonne - Université Paris-Saclay - IMT-BS - Institut Mines-Télécom Business School - IMT - Institut Mines-Télécom [Paris]
Thibault Pirson: UCLouvain - Université Catholique de Louvain = Catholic University of Louvain
Etienne Lees-Perasso: Toledo Institute for Development and Environment (TIDE)
Augustin Wattiez: UCLouvain - Université Catholique de Louvain = Catholic University of Louvain
Lisa Biber-Freudenberger: Universität Bonn = University of Bonn
Sasha Luccioni: Hugging Face
Aimee van Wynsberghe: Universität Bonn = University of Bonn

Post-Print from HAL

Abstract: The rapid expansion of AI has intensified concerns about its environmental sustainability. Current assessments focus on operational carbon emissions using secondary data, overlooking impacts in other life cycle stages. This study presents a comprehensive, multi-criteria life cycle assessment of AI training, building on an innovative life cycle inventory methodology for electronic products that combines physical teardown and multi-element composition analysis. Results for GPT-4 training show the use phase dominates 10 categories, contributing 96% to climate change and fossil fuel depletion. Manufacturing dominates 6 categories, including human toxicity (94%) and freshwater eutrophication (81%). The GPU chip is the largest contributor in 10 categories, particularly climate change (81%) and fossil resource use (80%). While primary data produces modest changes in carbon estimates, substantial variations emerge elsewhere, e.g. minerals and metals depletion increases by 33%. This analysis expands Sustainable AI discourse beyond carbon emissions, challenging current sustainability narratives.

Keywords: Cradle-to-grave; Environmental impact; Sustainable AI; NVIDIA A100; LCA; Life cycle assessment; Artificial intelligence (search for similar items in EconPapers)
Date: 2026-09
References: Add references at CitEc
Citations:

Published in Environmental Impact Assessment Review, 2026, 121, pp.108525. ⟨10.1016/j.eiar.2026.108525⟩

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:hal:journl:hal-05667182

DOI: 10.1016/j.eiar.2026.108525

Access Statistics for this paper

More papers in Post-Print from HAL
Bibliographic data for series maintained by CCSD ().