From Gigabytes to Gigatons: China’s Data Centers and Their Carbon Footprint from 2006 to 2050

View at publisher
Authors: Yi Jin, Yao Zheng, Junjie Ni, Cuiyang Feng, Rong Yuan, Xiaoyang Zhong, Paul Behrens
Cited by: 1
References: 45

Abstract

There has been a rapid growth of data-intensive activities in cloud computing and, more recently, training AI tools (e.g., LLMs), resulting in large increases in electricity demand and carbon emissions. Moving beyond high-level national estimates, here we compile a spatially explicit inventory of Chinese data centers and integrate this within the national electricity generation and transmission network. We find that data-center electricity consumption increased from 5.4 TWh to 157.4 TWh during 2006-2023, mostly in East China. Carbon emissions increased from 4.1 MtCO2 to 96.3 MtCO2 with cumulative emissions of 812 MtCO2. We find that transmission improvements avoided 65.9 MtCO2 of potential emissions over this period. Under our future scenarios, cumulative emissions reach 3.2 GtCO2 (1.4-7.9 GtCO2) between 2024 and 2050 while planned transmission avoids 190-1335 MtCO2. This work provides high-resolution insights into where energy system and data center interventions can be targeted.

Fields of Study & Subjects

Ammonia Synthesis and Nitrogen Reduction Water-Energy-Food Nexus Studies Environmental Impact and Sustainability

Publication details

Journal / source
Environmental Science & Technology
Publisher
American Chemical Society (ACS)
Volume
60
Issue
32
Pages
22389–22396
Publication date
25 July 2026
Language
EN