There are wide-ranging estimates of data centre energy use in the literature, causing confusion for policymakers and decision-makers. Estimates and
projections for 2020 range from less than 200 TWh to 1 200 TWh, while projections for 2030 range from just over 200 TWh to nearly 8 000 TWh – a factor of almost 40.
The objective of this study is to conduct a comprehensive and critical review of existing models and assessments of the energy use of data centres. Based on this analysis, it aims to answer the following research questions:
• How have previous studies estimated data centre energy use, and what factors have contributed to wide-ranging estimates?
• What is the most plausible range for data centre energy use today?
• What best practices should be followed when conducting assessments, and which pitfalls should be avoided?
This study identified and reviewed over 100 articles, reports, and statistics published since 2014, covering thousands of energy estimates and projections from 2010 to 2040 across over 200 scenarios and cases. Nearly half of these publications had been published since January 2024, indicating the recent and rapid growth in
interest. The studies cover global, regional, and country-level estimates from academia, industry, governments, and intergovernmental agencies. We also collated published energy data from 60 of the largest data centre operators since 2018. Finally, we reviewed emerging literature on AI energy use in data centres. This is the most
comprehensive published review of data centre energy estimates to date.