REVIEW 12 cited by
Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models
read the original abstract
The growing carbon footprint of artificial intelligence (AI) has been undergoing public scrutiny. Nonetheless, the equally important water (withdrawal and consumption) footprint of AI has largely remained under the radar. For example, training the GPT-3 language model in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 liters of clean freshwater, but such information has been kept a secret. More critically, the global AI demand is projected to account for 4.2-6.6 billion cubic meters of water withdrawal in 2027, which is more than the total annual water withdrawal of 4-6 Denmark or half of the United Kingdom. This is concerning, as freshwater scarcity has become one of the most pressing challenges. To respond to the global water challenges, AI can, and also must, take social responsibility and lead by example by addressing its own water footprint. In this paper, we provide a principled methodology to estimate the water footprint of AI, and also discuss the unique spatial-temporal diversities of AI's runtime water efficiency. Finally, we highlight the necessity of holistically addressing water footprint along with carbon footprint to enable truly sustainable AI.
Forward citations
Cited by 12 Pith papers
-
From Accounting to Coordination: A Virtual Water-Aware Electricity-Computation-Water Nexus Framework for Data Center Dispatch
The paper develops an ECW nexus framework using differentiable optimization layers and fixed-point coordination to internalize virtual water in data center power dispatch, showing 3-5% water withdrawal reductions on I...
-
The Hidden Cost of Thinking: Energy Use and Environmental Impact of LMs Beyond Pretraining
Full development of 7B and 32B Olmo 3 models used 12.3 GWh datacenter energy and emitted 4,251 tCO2eq, with development overheads accounting for 82% of compute and reasoning models costing 17x more to post-train than ...
-
Life Cycle Assessment of Pre-training the Lucie 7B Open-Source Large Language Model on the Jean Zay Supercomputer
Lucie 7B pre-training cost 21 tCO2eq (36.7 gCO2eq per H100 GPU-hour) including amortised manufacturing, plus ~76 m3 on-site water, on Jean Zay.
-
The Hidden Water Geography of U.S. Hyperscale Data Centers in the AI Era
Facility-resolved mapping of 472 U.S. hyperscale data centers shows ~300 GL/yr operational water, three-quarters electricity-related, with cooling and grid hotspots in different places.
-
Scalable Joint Resource Allocation for SLO-Constrained LLM Inference in Heterogeneous GPU Clouds
Two constraint-aware greedy heuristics (GH and AGH) solve mixed-scale LLM allocation on heterogeneous GPUs under SLO constraints in under one second with over 260x speedup and near-optimal cost compared to exact MILP.
-
The Rising Unsustainability of AI Graphics Cards Production
Analysis of a compiled dataset shows steadily rising production-related environmental impacts for NVIDIA workstation graphics cards from 2013 to 2025.
-
GreenZ: A Sustainable UX Framework for Complex Digital Systems
GreenZ is a conceptual three-layer sustainable UX framework built on ten principles, five operational systems, and practical tools, centered on an eight-type Digital Waste Taxonomy and a model questioning AI necessity...
-
Scalable Joint Resource Allocation for SLO-Constrained LLM Inference in Heterogeneous GPU Clouds
Constraint-aware greedy heuristics jointly allocate models, heterogeneous GPUs, TP/PP, and routing for SLO-constrained LLM inference in seconds, with better out-of-sample robustness than cost-minimal MILP plans.
-
MADP: A Multi-Agent Pipeline for Sustainable Document Processing with Human-in-the-Loop
MADP multi-agent pipeline with human-in-the-loop achieves 97% full automation on 955 real documents, 98.5% accuracy on ablation set, and 69-70% reductions in FTE, energy, and emissions versus manual processing.
-
From Cradle to Cloud: A Life Cycle Review of AI's Environmental Footprint
A review of AI sustainability studies finds inconsistent life cycle definitions and predominant reliance on coarse CO2e proxies, with limited coverage of water, materials, and multi-impact assessments.
-
Modest, artistic, and radical solutions to the environmental impact of image-generating machine learning
Surveys energy footprints of image ML and proposes modest technical solutions including tiny models, low-precision hardware, and true-cost accounting driven by critiques of shareholder efficiency metrics.
-
LLM Harms: A Taxonomy and Discussion
This paper proposes a taxonomy of LLM harms in five categories and suggests mitigation strategies plus a dynamic auditing system for responsible development.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.