With the advent of AI promising to disrupt different industries and drive innovation, Europe has an opportunity to leverage this technology to optimise its energy systems to ensure a more sustainable approach to using its limited resources.

Europe’s Energy Efficiency First principle aims to ensure that the energy demand is reduced and is managed cost-effectively. However, while AI offers solutions to help advance the Energy Efficiency First principle, the technology in itself relies on energy-intensive processes.

A spokesperson for the Commission said that data centres are now covered in Article 12 of the revised Energy Efficiency Directive. They said it’s the first time there is a dedicated article for data centres and energy efficiency. [Getty Images: Qi Yang]

The EU is already making plans to ensure that the environment is explicitly protected from this technology through its AI Act.

“Providers of general-purpose AI models, which are trained on large data amounts and therefore prone to high energy consumption, are required to disclose energy consumption,” the spokesperson told Euractiv. “In case of general-purpose AI models with systemic risks, energy efficiency needs to be assessed further.”

Enhancing Code of Practice

Obligations for providers of general-purpose AI models will be specified through a Code of Practice for which the process is about to be launched.

“There are ongoing discussions on energy consumption of general-purpose AI models in the context of the Code of Practice,” the spokesperson said.

They added that the Commission is empowered to adopt delegated acts to detail measurement and calculation methodologies for disclosure obligations, if necessary, to ensure comparable and verifiable documentation beyond what will be provided for in the Code.

AI testing centre in the works

From a practical standpoint, a new Horizon Europe-funded project aims to address key regulatory challenges, including compliance with the EU AI Act, ensuring that AI systems for energy are transparent, secure, reliable, and interpretable.

The AI-EFFECT project will be a collaboration between 19 European organisations in the last three years to create a robust framework for AI integration in energy systems. The project partners also aim to establish a European AI Testing and Experimentation Facility (TEF) for the energy sector, enabling development, testing, and validation at various stages.

“By making energy systems smarter, more efficient, and more reliable, we’re directly supporting decarbonisation efforts and advancing our mission to ensure a cleaner, more resilient energy future,” EPRI Europe Managing Director Eamonn Lannoye said.

‘Make more out of less’

In the private sector, semiconductor manufacturer Infineon Technologies says that an energy-efficient and high-performing power supply is crucial for powering the future of AI.

“One of Infineon’s guiding principles is ‘Make more out of less’, which is very much in line with the Energy Efficiency First principle. The world must reduce carbon emissions and use energy much more efficiently to secure quality of life for future generations,” a spokesperson for Infineon told Euractiv.

Infineon views itself as a key enabler in the move to harness renewable energy resources and deliver energy-efficient solutions along the entire electrical energy chain through its power system solutions.

“In the field of AI, efficient power semiconductors are an important lever to address societal and technological challenges fuelled by powering AI,” the spokesperson said. They explained that reducing power losses at every stage of power conversations needed to run AI processors needs to be prioritised to reduce cooling needs and to increase energy efficiency.

“In the future, so-called wide-bandgap materials like silicon carbide and gallium nitride will play a pivotal role in energy-efficient semiconductors. The key for the next essential step towards an energy-efficient world lies in the use of these new WBG power electronics materials that allow for greater power efficiency, smaller size, lighter weight, lower overall cost – or all of these together,” Infineon’s spokesperson said.

According to Infineon, data centres already consume 2% of the world’s electricity demand. It expects this share to grow to 7% by 2030.

A spokesperson for the Commission said that data centres are now covered in Article 12 of the revised Energy Efficiency Directive. They said it’s the first time there is a dedicated article for data centres and energy efficiency.

The article obliges data centres to publish data on the power they consume, except data that could be commercially sensitive.

“A study to help establish a rating and labelling scheme for data centres will start soon,” the Commission spokesperson said.

According to euractiv.com