Search eceee proceedings

AI, Energy-Intensive Digital Applications and Data Centers: Uncertain Energy Demand and Policy Responses

Panel: 4. Drivers to action – policies and market mechanisms

This is a non-peer reviewed extended abstract. No paper available.

Author:
Frédéric Renkens, Office fédéral de l'énergie OFEN (Suisse), Switzerland Mihaela Grigorie, Swiss Federal Office of Energy, Switzerland

Abstract

Artificial intelligence and other energy-intensive digital applications have the potential to enhance energy efficiency in various sectors and industries. Concurrently, their implementation may result in a significant increase in electricity demand, particularly through the establishment and operation of new data center infrastructures.

According to a 2020 study by the Swiss Federal Office of Energy, data centers in Switzerland consumed 2.1 TWh of electricity, accounting for 3.6% of the country's total electricity usage. This figure is expected to have increased since then (results of a new study are pending) and may continue to rise in the near future, depending on the pace of AI adoption and data center expansion.

To address this challenge, Switzerland has implemented a range of policy measures aligned with the IEA’s recommended triptych: regulations, information, and incentives. These include:

• Minimum Energy Performance Standards (MEPS) for IT equipment, in alignment EU directives.

• Surveys on current and projected energy consumption in the sector, the energy impact of AI and other digital applications, and the implementation of efficiency measures, including waste heat recovery.

• Dissemination of knowledge through various means, including best practices, labels, guides, stakeholder networking, and voluntary initiatives.

• The ProKilowatt funding program, which supports hardware energy efficiency measures.

• The new Swiss energy efficiency obligation scheme for electricity providers.

The presentation will provide an overview of Switzerland’s policy landscape, highlight ongoing initiatives, and invite discussion on their effectiveness and improvement potential.

Downloads

Download this presentation as pdf: 4-096-25_Renkens_Pres.pdf