AIOTI WG Energy has released a new paper on AI Factories for the Energy sector.
The paper explores the growing interdependence between Europe’s energy system and AI infrastructure, highlighting how the increasing use of artificial intelligence in energy operations is accompanied by rising electricity demand from the computing infrastructure behind it. Data centres consumed approximately 415 TWh in 2024, accounting for 1.5% of global electricity consumption. Their demand grew by 17% in 2025, while AI-focused data centres saw a 50% increase.
The publication examines how existing European AI Factories and High-Performance Computing (HPC) infrastructure can support the digital transformation of the Energy sector. It identifies key challenges, including rising energy demand, renewable energy integration and grid volatility, cooling and efficiency requirements, ageing infrastructure, and fragmented energy data.
One of the main topics of the paper is the proposed Energy AI Factory, a usage pattern and reference architecture that builds on existing European AI Factories rather than requiring new facilities. It connects three complementary layers: computing and sensing capabilities at the grid edge, specialised intermediaries known as AI Gates that support data preparation and validation, and AI Factories providing large-scale computing resources for model training and simulation. The architecture also addresses data governance, interoperability, security, regulatory compliance, and sustainability.
The paper also proposes a checklist outlining what AI Factories need to serve the Energy sector.
Read the full document here.
