PARIS ­| February 20th, 2025 – BG IRIS, Bryan Garnier’s research platform, is pleased to release “Energy-Aware AI: Confronting the power-hungry reality of scaling AI”, an in-depth analysis exploring the impact of AI on data center infrastructure and energy consumption.

As generative AI fuels an unprecedented surge in compute demand, data centers designs are evolving to handle soaring power densities and thermal needs of AI servers. With data center electricity consumption set to double by 2030, we explore the strategies and innovations that can enable the industry to scale sustainably.

  • The AI era is triggering a seismic shift in global capex priorities, with data centres poised to attract nearly USD500bn in 2025 alone.
  • AI is turning data centers into power-hungry giants, with their electricity consumption set to exceed 1000TWh by 2030, putting pressure on energy systems.

The AI-driven investment cycle is reshaping dynamics for electrical and thermal system suppliers, as the demands of accelerated computing drive convergence between IT equipment and infrastructure (power and cooling).

In response, the industry is seeing rapid advancements in advanced cooling, high-efficiency power systems, and energy management solutions. Policymakers and enterprises alike are seeking pathways to balance innovation with sustainability, shaping the future of AI infrastructure.

To dive deep into this topic and discover more about the emerging technologies and strategies shaping sustainable AI infrastructure, download the industry brief.