Advanced Search
Article Contents
ABOUT THIS ARTICLE
Energy Use > 2026 Vol. 2 > No. 3 > 100074
Next Article Previous Article
LETTER   Open Access     Cite

AI Data-Center Flexibility is Conditional, Not Intrinsic

  • 1.

    School of Architecture, Tsinghua University, Beijing 100084, China

  • 2.

    Key Laboratory of Eco Planning and Green Building, Ministry of Education, Beijing 100084, China

  • 3.

    State Grid Zhejiang Marketing Service Center, Hangzhou 310007, China

  • 4.

    Institute for Urban Governance and Sustainable Development, Tsinghua University, Beijing 100084, China

  • 5.

    China United Network Communications Co., Ltd. Guangdong Branch, Guangzhou 510603, China

  • 6.

    China Unicom (Guangdong) Industrial Internet Co., Ltd., Guangzhou 510320, China

    • Show all affliationsShow less
    Fund Project: This study is supported by the National Natural Science Foundation of China (Grant No.52425801, 52394223, 52130803, and 72374121), and the National Key Research and Development Program of China (Grant No. 2024YFE0212900).
More Information
  • Corresponding author: linbr@tsinghua.edu.cn
  • Received Date:  September 20, 2026
    Accepted Date:  September 22, 2026
    Published Online:  September 23, 2026
Energy Use  2(3),  https://doi.org/10.59717/ipj.energy-use.2026.100074  |  Cite this article
    • FullText(HTML)
  • 加载中
    • REFERENCES
  • [1] International Energy Agency. (2025). Energy and AI. IEA.

    View in Article Google Scholar

    [2] Colangelo P., Coskun A. K., Megrue J. et al. (2026). AI data centres as grid-interactive assets. Nat. Energy 11:254−261. DOI:10.1038/s41560-025-01927-1

    View in Article CrossRef Google Scholar

    [3] Caprara A., Yu Y., Teng F., et al. (2026). Data center workload flexibility for power system demand response: Evidence from Alibaba traces. Int. J. Elect. Power Energy Syst. 178:111940. DOI:10.1016/j.ijepes.2026.111940

    View in Article CrossRef Google Scholar

    [4] Wang Y., Guo Q. and Chen M. (2025). Providing load flexibility by reshaping power profiles of large language model workloads. Adv.Appl. Energy 19:100232. DOI:10.1016/j.adapen.2025.100232

    View in Article CrossRef Google Scholar

    [5] Hanafy W. A., Liang Q., Bashir N., et al. (2023). CarbonScaler: Leveraging cloud workload elasticity for optimizing carbon-efficiency. Proceed. ACM Measur. Anal. Comput. Syst. 7:57. DOI:10.1145/3626788

    View in Article CrossRef Google Scholar

    [6] Patel P., Choukse E., Zhang C., et al. (2024). Splitwise: Efficient generative LLM inference using phase splitting. IEEE ISCA. DOI:10.1109/ISCA59077.2024.00019.

    View in Article Google Scholar

    [7] Zhong Y., Liu S., Chen J., et al. (2024). DistServe: Disaggregating prefill and decoding for goodput-optimized large language model serving. Proceed. OSDI :193–210. DOI:10.48550/arXiv.2401.09670.

    View in Article Google Scholar

    [8] Kwon W., Li Z., Zhuang S., et al. (2023). Efficient memory management for large language model serving with PagedAttention. Proceed. SOSP. DOI:10.1145/3600006.3613165.

    View in Article Google Scholar

    [9] Sukprasert T., Souza A., Bashir N., et al. (2024). On the limitations of carbon-aware temporal and spatial workload shifting in the cloud. Proceed. EuroSys :924–941. DOI:10.1145/3627703.3650079.

    View in Article Google Scholar

    [10] Hanafy W. A., Sukprasert T., Souza A., et al. (2026). Untangling the carbon-cost tradeoffs and stampede effect challenges in cloud computing. IEEE Trans. Comput. 75:2715−2729. DOI:10.1109/TC.2026.3667610

    View in Article CrossRef Google Scholar

    • Cited by

  • Cite this article:

    Yang Y., Chen Y., Yuan M., et al. (2026). AI Data-Center Flexibility is Conditional, Not Intrinsic. Energy Use 2:100074. https://doi.org/10.59717/ipj.energy-use.2026.100074
    Yang Y., Chen Y., Yuan M., et al. (2026). AI Data-Center Flexibility is Conditional, Not Intrinsic. Energy Use 2:100074. https://doi.org/10.59717/ipj.energy-use.2026.100074

Welcome!

To request copyright permission to republish or share portions of our works, please visit Copyright Clearance Center's (CCC) Marketplace website at marketplace.copyright.com.

PDF
Figures(1)    

Share

  • Share the QR code with wechat scanning code to friends and circle of friends.

Article Metrics

Article views(324) PDF downloads(90)

Relative Articles

Cited by

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Published articles are licensed under CC BY with copyright retained by the authors.
    All contents on this site: Copyright © 2026 INNOVATION PRESS. All rights are reserved.