A Water Efficiency Dataset for African Data Centers
Open accessAbstract
Artificial intelligence (AI) computing and data centers consume large amounts of freshwater, both directly for cooling and indirectly for electricity generation. While most attention has been paid to developed countries such as the U.S., this paper presents the first-of-its-kind dataset that combines nation-level weather and electricity generation data to estimate water usage effectiveness for data centers in 41 African countries across five different climate regions. We also use our dataset to evaluate and estimate the water consumption of inference on two large language models (i.e., Llama-3-70B and GPT-4) in 11 selected African countries. Our estimates suggest that writing a 10-page report using Llama-3-70B could consume as much as {0.66 liters} of water, while the water consumption by GPT-4 for the same task may go up to about {59 liters}. For writing a medium-length email of 120-200 words, Llama-3-70B and GPT-4 could consume about {0.13 liters} and {2.9 liters} of water, respectively. All the numbers for generative model inference tasks are based on public information available in 2024, when we initially prepared the analysis. Since then, AI inference systems have improved substantially. For example, recent disclosures suggest that energy efficiency improved by more than 30x between May 2024 and May 2025. Accordingly, our 2024 estimates should be interpreted as historical reference values rather than as representative of current performance. Interestingly, given the same AI model, 9 of the 11 selected African countries consume less water than the global average, mainly because of lower water intensities for electricity generation.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Some subsequent models like GPT-4 could consume substantially more energy and water than GPT-3 for processing the same request, especially under the reasoning mode.
“some subsequent models like GPT-4 could consume substantially more energy and water than GPT-3 for processing the same request, especially under the reasoning mode.3,26”
From Making AI Less 'Thirsty' (2025) · cites this paper as [26]
The Abstract (p. 1) and §4.1 (p. 5) estimate substantially greater water and energy use for GPT-4 than for Llama-3-70B on the same writing tasks. They do not compare GPT-4 with GPT-3 or evaluate reasoning mode, so those parts of the claim are unsupported by this paper. (LLM, conf=0.85)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
United States Data Center Energy Usage Report
Lawrence Berkeley National Laboratory (LBNL), B., Shehabi, A., USDOE, Smith, S., Sartor, D., Brown, R., Herrlin, M., Koomey, J., Masanet, E., Horner, N., et al. (2016) · doi:10.2172/1372902 · Cited by 1,040 · cited 1× · see citing works
“Data centers are notorious for their massive energy usage and water consumption, which have raised significant concerns even in developed countries such as the U.S. [5, 23, 40].”
Splitwise: Efficient Generative LLM Inference Using Phase Splitting
Patel, P., Choukse, E., Zhang, C., Shah, A., Goiri, Í., Maleki, S., & Bianchini, R. (2024) · 2024 ACM/IEEE 51st Annual International Symposium on Computer Architecture (ISCA) · doi:10.1109/isca59077.2024.00019 · Cited by 298 · cited 6× · see citing works
“…the industry practice, since a fully-utilized system can lead to significant SLO violations, which are not tolerable in real-world LLM deployment, especially for commercial LLM applications such as real-time conversations that have strict SLO targets to deliver good quality of experiences [33, 45].”
InfluentialThe environmental footprint of data centers in the United States
Siddik, M. A. B., Shehabi, A., & Marston, L. (2021) · Environmental Research Letters · doi:10.1088/1748-9326/abfba1 · Cited by 281 · see citing works
Li, P., Yang, J., Islam, M. A., & Ren, S. (2025) · Communications of the ACM · doi:10.1145/3724499 · Cited by 176 · see citing works
Self-citationAn architecture design with data model for product recovery management systems
Um, J., Yoon, J., & Suh, S. (2008) · Open Access System for Information Sharing (Pohang University of Science and Technology) · doi:10.1016/j.resconrec · Cited by 124 · see citing works
DynamoLLM: Designing LLM Inference Clusters for Performance and Energy Efficiency
Stojkovic, J., Zhang, C., Goiri, Í., Torrellas, J., & Choukse, E. (2025) · 2025 IEEE International Symposium on High Performance Computer Architecture (HPCA) · doi:10.1109/hpca61900.2025.00102 · Cited by 112 · see citing works
Characterizing Power Management Opportunities for LLMs in the Cloud
Patel, P., Choukse, E., Zhang, C., Goiri, Í., Warrier, B., Mahalingam, N., & Bianchini, R. (2024) · International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) · doi:10.1145/3620666.3651329 · Cited by 93 · see citing works
The carbon emissions of writing and illustrating are lower for AI than for humans
Tomlinson, B., Black, R. W., Patterson, D. J., & Torrance, A. W. (2024) · Scientific Reports · doi:10.1038/s41598-024-54271-x · Cited by 77 · see citing works
Water-energy tradeoffs in data centers: A case study in hot-arid climates
(2022) · Resources Conservation and Recycling · doi:10.1016/j.resconrec.2022.106194 · Cited by 43 · see citing works
Freshwater use of the energy sector in Africa
Sanchez, R. G., Seliger, R., Fahl, F., Felice, L. D., Ouarda, T. B. M. J., & Farinosi, F. (2020) · Applied Energy · doi:10.1016/j.apenergy.2020.115171 · Cited by 43 · see citing works
(2018) · doi:10.1109/tcc.2016.2535201 · Cited by 39 · see citing works
Self-citationGnibga, W. E., Chien, A. A., Blavette, A., & Orgerie, A. (2024) · ACM International Conference on Future Energy Systems (e-Energy) · doi:10.1145/3632775.3661936 · Cited by 16 · see citing works
A Dataset for Research on Water Sustainability
Gupta, P. S., Hossen, M. R., Li, P., Ren, S., & Islam, M. A. (2024) · ACM International Conference on Future Energy Systems (e-Energy) · doi:10.1145/3632775.3661962 · Cited by 14 · see citing works
Self-citationNono, K. J. N., Mvongo, V. D., & Defo, C. (2024) · Water Science & Technology Water Supply · doi:10.2166/ws.2024.071 · Cited by 10 · see citing works
Flagged first, then by citations
Research works printed but not linked: 7
- [5][5] Kishwar Ahmed, Mohammad A Islam, Shaolei Ren, and Gang Quan. 2014. Can Data Center Become Water { Self-Sufficient } ?. In 6th Workshop on Power-Aware Computing and Systems (HotPower 14) . -
- [8][8] Cooper Elsworth et al. 2025. Measuring the environmental impact of delivering AI at Google Scale. CoRR abs/2508.15734 (2025). https://doi.org/10.48550/arXiv. 2508.15734 - · DOI
- [24][24] Pengfei Li, Jianyi Yang, Mohammad A. Islam, and Shaolei Ren. 2024 (accepted). Making AI Less 'Thirsty': Uncovering and Addressing the Secret Water Footprint of AI Models. Commun. ACM (2024 (accepted)). -
- [25][25] Tianyi Luo, Robert Young, and Paul Reig. 2015. Aqueduct projected water stress country rankings. Technical Note 16 (2015), 1-16. -
- [27][27] Mesfin M Mekonnen, PW Gerbens-Leenes, and Arjen Y Hoekstra. 2015. The water footprint of electricity from hydropower. Hydrology and Earth System 27. Sciences 19, 5 (2015), 2091-2101. doi:10.5194/hess-19-2091-2015 - · DOI
- [41][41] Noah Shumba, Opelo Tshekiso, Pengfei Li, Giulia Fanti, and Shaolei Ren. 2024. A Water Efficiency Dataset for African Data Centers. https://arxiv.org/abs/2412. 03716 -
- [43][43] Siemens. 2018. Completion of world's largest combined cycle power plants in record time. https://press.siemens.com/global/en/feature/completion-worldslargest-combined-cycle-power-plants-record-time. Egypt 14.4 GW CCGT megaproject. -
Cited by
Sorted by citations
Li, P., Yang, J., Islam, M. A., & Ren, S. (2025) · Communications of the ACM · doi:10.1145/3724499 · Cited by 176 · see citing works
Self-citationPeykani, P., Emrouznejad, A., Ghanidel, S., Javadi-Sisi, I., & Mirjalili, S. (2026) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-026-10546-2 · Cited by 8 · see citing works
ВУГЛЕЦЕВИЙ ТА ВОДНИЙ СЛІД ПРОВІДНИХ ТЕХНОЛОГІЧНИХ КОМПАНІЙ НА ТЛІ РОЗВИТКУ ШТУЧНОГО ІНТЕЛЕКТУ
Melnyk, L. H., Dudka, M., & Dehtyarova, I. B. (2025) · Цифрова економіка та економічна безпека · doi:10.32782/dees.19-9
Mrindoko, S., & Gidion, D. K. (2026) · East African Journal of Engineering · doi:10.37284/eaje.9.1.4525
Environmental Impact of AI Data Centers in Kuwait
Emran, L., Alrashidi, A. M., Ahmad, I., & Abed, S. (2026) · International Symposium of Systems, Advanced Technologies and Knowledge (ISSATK) · doi:10.1109/issatk69667.2026.11566690
AI Data Centers in Cooling Mechanisms and Freshwater Consumption
Mukhopadhyay, O., Maji, B., & Mohapatra, H. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-8434-4.ch002
Vera, R. C., Rodriguez, R. H. T., Pacheco, D. A. G., Rojas, P. M. P., & Pinaya, I. R. A. (2026) · Revista horizonte académico · doi:10.70208/3007.8245.v6.n2.475
Space Based Data Centers-System Design for Data Migration Service at End of Life
Periola, A. A., Alonge, A., & Ogudo, K. A. (2026) · International Convention on Information and Communication Technology, Electronics and Microelectronics · doi:10.1109/mipro70003.2026.11591552
Distributed Underwater Data Center Architecture for Cost and Uptime Optimization
Periola, A. A., Alonge, A., & Ogudo, K. A. (2026) · International Convention on Information and Communication Technology, Electronics and Microelectronics · doi:10.1109/mipro70003.2026.11591819
Innovation at the Time of AI and Environmental Crisis
Crimi, A. (2026) · Sustainable development goals series · doi:10.1007/978-3-032-21563-5_4
Sharma, B., Tiwari, A., Kaur, R., Park, K., & Pinsky, E. (2026) · Green · doi:10.3390/green1020008
Purnomo, V. (2026) · Politeia Journal of Public Administration and Political Science and International Relations · doi:10.61978/politeia.v4i3.1655
Carr, J., Leimbach, T., & Milstein, T. (2026) · Humanities and Social Sciences Communications · doi:10.1057/s41599-026-08884-4
Innovazione al tempo dell’IA e della crisi ambientale
Crimi, A. (2026) · SxI Springer per l'Innovazione · doi:10.1007/978-3-032-35860-8_4
- Authors
- Shumba, N., Tshekiso, O., Li, P., Fanti, G., & Ren, S.
- Affiliations
- Carnegie Mellon University Africa (RW)
- University of California, Riverside (US)
- Carnegie Mellon University (US)
- Venue
- Proceedings of the ACM SIGCAS/SIGCHI Conference on Computing and Sustainable Societies
- Publisher
- ACM
- Published
- 2025-07-21
- Pages
- 453-460
- OpenAlex
- W4413061419
- Cited by
- 10
- References
- 13
- Citations by year
- 2026: 9 · 2025: 1
- FWCI
- 21.67
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by-nc-nd
- Locations
- ACM SIGCAS/SIGCHI Conference on Computing and Sustainable Societies (COMPASS) (publishedVersion, cc-by-nc-nd)
- arXiv (Cornell University) (submittedVersion, cc-by-nc-nd)
- arXiv (Cornell University)
- Funders
- Gates Foundation
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Green IT and Sustainability
- Physical Sciences › Computer Science › Hardware and Architecture › Energy Efficiency in Computing
- Physical Sciences › Computer Science › Information Systems › Cloud Computing and Resource Management
- Keywords
- water efficiency, data centers, water consumption, energy efficiency
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
