Data centres and AI: legal challenges between sustainability and energy efficiency in an Italian and EU context
Andea Marega
Ughi e Nunziante Studio Legale, Rome
Cristina Maria Celotto
Ughi e Nunziante Studio Legale, Rome
Introduction
The global digitalisation and widespread adoption of Artificial Intelligence (AI) have amplified the role of data centres as fundamental pillars of modern technological infrastructure. However, the energy consumption of such facilities raises several issues in light of the climate neutrality objectives set by the European Union in its Green Deal.
The role of data centres in the digital ecosystem
Data centres are physical infrastructures designed to host servers, storage systems, and networking equipment. They represent the beating heart of global technological infrastructure. Every online activity, from sending emails to streaming videos, depends on the ability of data centres to process, store, and distribute vast amounts of data, ensuring efficiency, security, and operational continuity.
Among the key benefits, the following can be highlighted:
- support for digital transformation – they enable technologies such as AI, the Internet of Things (IoT), and 5G;
- improvement of connectivity – they facilitate the rapid exchange of data between users and applications; and
- guarantee of security and compliance – they manage sensitive data in compliance with international regulations on privacy and data protection.
The growth of energy consumption
According to data from the European Data Centre Association (EUDCA), energy consumption of data centres accounts for approximately two-to-three per cent of global electricity demand, with growth estimates of up to 13 per cent by 2030.
AI as a tool for optimising energy consumption
In the described context, AI, with its computationally intensive workloads, significantly contributes to increased energy consumption. However, AI itself can be employed to improve the energy efficiency of data centres through various approaches, such as:
- Intelligent cooling management – advanced algorithms analyse data from thermal sensors and dynamically regulate cooling to maintain optimal temperature with minimal energy expenditure.
- Dynamic resource allocation – AI can efficiently balance workloads, distributing operations to underutilised or less energy-intensive servers, reducing waste.
- Predictive maintenance – through machine learning, failures or malfunctions in equipment can be anticipated, enabling timely interventions that prevent energy waste and improve reliability.
- Power supply optimisation – AI can monitor and predict energy demand peaks, allowing more efficient resource management and the potential use of renewable energy.
- Automation of server management – algorithms can shut down or reduce the energy consumption of inactive or underutilised servers, optimising the balance between performance and energy consumption.
Italian and EU regulations on data centre sustainability
The regulation of data centre energy consumption and environmental sustainability operates on two levels in Italy: EU-wide and national.
EU regulations
At the EU level, the main regulatory instruments adopted to address energy challenges include Directive 2012/27/EU, Regulation (EU) 2019/424, the European Green Deal, and the Fit for 55 plan.
Directive 2012/27/EU establishes a common framework of measures to promote energy efficiency within the EU, setting energy efficiency requirements for large IT infrastructures.
Regulation (EU) 2019/424 establishes eco-design requirements for servers and data storage systems to improve energy efficiency and reduce the environmental impact of these devices. It applies to servers used in data centres or corporate environments and to data storage systems employed for preserving and managing digital information.
Since 1 March 2020, servers and storage systems must comply with several requirements, including:
- Energy efficiency – devices must meet minimum energy efficiency standards based on their type and configuration and must include power management features (eg, power-saving modes and automatic shutdown).
- Materials and repairability – devices must be designed to facilitate repair and upgrades, reducing waste, and ensuring firmware updates are available for at least eight years after the last model is sold.
- Recoverability and recyclability – devices must allow disassembly without specialised tools for material separation and must provide clear information on materials used and disposal methods.
- Data transparency – manufacturers must provide detailed information on energy consumption, power unit efficiency, heat emissions, and compliance with environmental standards.
The European (EU) Green Deal and the Fit for 55 plan are fundamental EU strategies for ecological transition and combating climate change. The Green Deal, launched by the European Commission in December 2019, aims to make the EU ‘the first climate-neutral continent’ by 2050. This plan foresees a radical transformation of the EU economy, focusing on clean energy, circular economy, sustainable mobility, biodiversity, and sustainable agriculture.
The Fit for 55 plan is a legislative package launched by the European Commission in July 2021 to translate into concrete regulations the goal of reducing emissions by at least 55 per cent by 2030. Among its key objectives is energy efficiency, aiming to reduce energy consumption by nine per cent by 2030 compared to initial projections.
Italian regulations
In Italy, regulation focuses on implementing EU directives and promoting sustainable solutions. Accordingly, the National Integrated Energy and Climate Plan (PNIEC) and the ENEA Guidelines (National Agency for New Technologies, Energy, and Sustainable Economic Development) have been adopted.
The PNIEC includes specific objectives for energy efficiency in digital infrastructures and foresees significant investments in the enhancement of such infrastructures in the energy sector, with particular attention to cybersecurity, data management, and interconnection between networks and facilities.
ENEA has developed specific guidelines for designing high-energy-efficiency data centres, providing recommendations for the construction of computing centres which optimise energy consumption while reducing environmental impact and operational costs.
Key legal challenges
The evolution of data centres raises a series of significant legal issues, among which several complex aspects emerge.
Regarding the environmental profile, closely linked to the high energy consumption of such infrastructures, possible regulatory interventions include the imposition of binding obligations for emissions reduction and the definition of energy efficiency criteria requiring the adoption of advanced cooling technologies and heat recovery systems.
Another critical area, partially related to the previous one, concerns energy procurement. It is necessary to introduce specific regulatory obligations to promote the use of renewable energy sources and regulate in detail long-term energy purchase agreements, commonly known as Power Purchase Agreements (PPA).
A third relevant issue concerns cybersecurity and privacy protection. The exponential increase in data processing, particularly regarding AI, heightens the risk of cyberattacks and personal data breaches. As a result, advanced security protocols and technological tools must be adopted to ensure the protection of information managed within data centres.
The integration of AI and data centres raises complicated legal questions but also offers opportunities for sustainable development. Among the possible legal solutions to address the aforementioned challenges are the introduction of specific environmental certifications for data centres, the implementation of incentive mechanisms for infrastructures adopting sustainability models, and, conversely, the application of sanctioning measures against those failing to comply with established environmental standards.