Due to current trends in computing, the data stored and processed in data centers (DCs) will continue to grow substantially. These trends include cloud computing, video streaming, the Internet-of-Things (IoT), and in particular artificial intelligence (AI) and machine learning (ML). Mitigating the growing demand, efficiency gains are able to partly offset it and limit the growth of data center energy consumption.

As servers are the main devices responsible for electricity demand in data centers, their energy efficiency is particularly relevant for both the current status and future developments. For several reasons, however, it is notoriously difficult to assess, e.g.: miscellaneous types of hardware and computing workloads, lack of standardized methodologies and assumptions, consistent system boundaries, technological dependencies with other physical and software components, and the dynamic nature of the data center domain and rapid innovation cycles.

The main objective of this study is thus to provide insights into the development of server energy efficiency over the past decade.