Generative AI in data centre design: challenges & opportunities

27-08-2025

Generative AI is accelerating demand for digital infrastructure at an unprecedented pace. For data centre owners, operators, and investors, the challenge is no longer whether AI will reshape facilities, but how to ensure your data centre withstands its intensity while remaining efficient and sustainable.

Long corridor of server racks inside a data centre, symbolising AI-ready digital infrastructure

Generative AI is no longer a niche technology; it drives demand for computing power and digital infrastructure worldwide. Unlike traditional IT workloads, it demands far greater processing capacity, higher energy use, and advanced cooling systems. This reality forces the data centre industry to rethink resilience, efficiency, and long-term planning.

For data centre owners, operators, and investors, this shift raises critical questions: How should facilities be designed, expanded, or retrofitted to meet this surge in demand? And how can your data centre stay efficient and sustainable in a rapidly evolving technological landscape?

Designing for AI means looking beyond immediate demand. It’s about future-proofing data centres with flexibility, efficiency, and responsibility at the core.

Stijn de KruijfData centre facility and sustainability expert

Why generative AI reshapes demand

A single AI query consumes many times more computing resources than a standard search. As adoption accelerates across industries, demand for high-performance capacity continues to soar.

McKinsey projects that global AI-ready capacity will more than triple by 2030, growing at an average of 33 per cent each year. In Europe, pressure is already mounting. Major hubs such as Dublin, Frankfurt, Amsterdam, and Paris face increasing scarcity of power, land, and permits.

This surge is not just about scale but also about intensity. Generative AI workloads drive higher rack densities, demanding far more electricity per square metre than most existing facilities were designed to handle.

Generative AI pushes infrastructure to its limits — but also inspires smarter, greener solutions. This is our opportunity to build data centres fit for both growth and sustainability.

Stijn de KruijfData centre facility and sustainability expert

Design and operational challenges

Generative AI forces a fundamental rethink of how data centres are designed and operated. Power infrastructure comes under immediate strain, with AI servers pushing electrical systems beyond conventional limits. Meeting this demand requires upgraded grid connections and smarter integration of on-site generation and storage.

Cooling is equally critical. Air-based systems often cannot cope with high-density loads. Technologies such as liquid-to-chip cooling and full immersion systems are gaining traction, enabling thermal stability and greater operational efficiency.

Space and scalability require new approaches. Retrofitting existing sites can involve extensive redesign, while new builds must incorporate flexibility to accommodate evolving hardware requirements.

Sustainability intensifies the challenge. The International Energy Agency (IEA) projects that data centre electricity consumption will more than double by 2026. Owners and operators must balance AI-driven growth with decarbonisation goals.

Existing vs. new facilities

For existing facilities, the challenge lies in upgrading within physical and regulatory limits. Operators often face constraints on grid connections, floor loading, or cooling capacity. Selective retrofits, such as liquid cooling integration or modular power expansions, can extend facility lifespans, though they also introduce complexity.

For new builds, the opportunity is to design “AI-ready” facilities from the ground up. Yet uncertainty remains. Hardware architectures evolve quickly, and tailoring a site too narrowly to today’s specifications may limit adaptability tomorrow. Experienced engineering consultancy helps strike the right balance between immediate readiness and long-term flexibility.

We help you keep data centres up to date to meet evolving challenges

Creating a data centre today means focusing on the demands of tomorrow. Our team of 200 technical experts have a proven track record in creating technologically advanced mission critical facility designs that are fully future-proof.

A brief note on AI in optimisation

While this blog focuses on the impact of generative AI on infrastructure demand, it is worth noting that AI also plays a role inside data centres. Predictive maintenance, energy optimisation, and automation offer new opportunities to improve operational efficiency. However, these applications remain secondary to the central challenge: ensuring facilities can physically host and cool the workloads that generative AI requires.

Looking ahead

Generative AI represents both a challenge and an opportunity for the data centre industry. Facilities that embrace new power and cooling technologies, rethink design principles, and plan for scalability will be best placed to capture demand in this fast-growing market.

At the same time, collaboration across regulators, utilities, investors, and technology providers is essential to ensure growth remains sustainable. In Europe, where power and space constraints are acute, innovation and shared responsibility will determine which facilities thrive in the generative AI era.

Become a partner

We work with clients worldwide to design and operate data centres that are resilient, efficient, and ready for the demands of generative AI.
Connect with us today to explore how your facility can adapt to the demands of generative AI.

MartienArts

Director Mission Critical Facilities