Data centres and AI: challenges and answers of an (also) infrastructural revolution

The acceleration imposed by artificial intelligence on digital transformation is profoundly redefining the entire IT ecosystem. It is not just about the impact on business models or business processes: it is the physical infrastructure itself, which underpins this revolution, that is entering a new phase. The increasing adoption of large-scale language models is forcing a radical transformation: traditional data centres are no longer sufficient. From power supply to cooling to rack density, every component must evolve to keep up with the new workloads, driven by ever-increasing data processing demands.

According to IDC, server farms could account for 2.5 per cent of the world's electricity consumption by 2027, with the remaining 97.5 per cent distributed among sectors such as construction, manufacturing, transportation and energy.In the European panorama, Italy is emerging as one of the new strategic poles, asserting itself alongside the traditional markets (Frankfurt, London, Amsterdam, Paris and Dublin - the so-called FLAPD axis) also thanks to its geographical location in the Mediterranean, a fundamental junction for the submarine backbones of international connectivity.

Everything Starts with the GPU

The evolution of the scenario calls for profound reflection, technological vision and strategic choices. An "AI-ready" data centre is not simply an upgrade of traditional infrastructures: it is a new architectural category, characterised by radically higher performance requirements. At the heart of this transformation is the GPU: the graphics cards that power artificial intelligence processes work in high-density clusters, with much higher energy demands than conventional servers. To manage them, new solutions are needed, specifically designed to address unprecedented challenges on multiple levels: energy, thermal, management, environmental.

To respond to the ongoing evolution, data centres must address three main challenges: rethinking energy strategy: growing computational loads require new energy infrastructures, capable of ensuring continuity, resilience and scalability; developing high-density infrastructures: traditional systems are not designed for the very compute-intensive configurations typical of AI and consequently new approaches to powering and cooling machines are required; support the transition with efficient and sustainable technologies: it is crucial to reduce environmental impact and contain operating costs by decoupling the growth of computing power from energy consumption.

The complexities to be addressed are therefore diverse: they start with the need to equip data centres with high-performance computing capabilities and encompass equally crucial issues such as sustainability, data security and AI governance within data centres.

Infrastructure lifecycle management must therefore encompass these aspects as well, flanking the physical component with a digital governance strategy aligned to the scale of this revolution.

 

End-to-end solutions for a sustainable future

And it is to meet these challenges that Schneider Electric, recognised by Corporate Knights for the second time as the world's 1st most sustainable company, operates in the market through an integrated and modular approach that is embodied in a complete and scalable solution offering covering the entire data centre lifecycle, from the electrical grid ('grid') to the computational core ('chip'), and then extending to the thermal management ('chiller'), with the aim of optimising performance, reliability and consumption to stay in line with sustainability targets.

Schneider Electric, also thanks to the collaboration with a solid ecosystem of partners and global alliances, provides all the tools, the necessary competences and the most technologically advanced solutions to face this new phase and to accompany companies in the realisation of data centres ready for artificial intelligence.

Tackling the AI revolution today means transforming data centres from processing centres into real strategic assets for innovation and business competitiveness. To do so, solutions must be put in place that ensure maximum operational efficiency even with extreme AI loads, centralised management and smooth scalability. Schneider Electric responds to these needs by supporting both the power flow, e.g. with racks and uninterruptible power supplies dedicated to AI, and the thermal flow needs with state-of-the-art cooling solutions, enhanced by the recent acquisition of Motivair Corporation.

Sustainability is at the heart of this data centre transformation process, it is a concrete goal to be achieved and therefore an integral part of Schneider Electric's mission. Schneider Electric provides its corporate customers with a complete portfolio of energy efficiency technologies, advanced management software for monitoring and optimising consumption, and consulting and design services.

The value of an end-to-end technology partner

Artificial intelligence represents a technological challenge, but also an opportunity. To seize it, it is necessary to rethink the IT infrastructure underpinning this revolution and to rely on technology partners with a clear strategy.

Schneider Electric works alongside its customers to guide and accelerate the change with awareness and a positive impact on both business and the environment, guaranteeing constant support at every stage, from the design to the installation of data centres, from operational management to the predictive maintenance of equipment to its evolution and scalability over time. A single interlocutor, an end-to-end vision. From network to chip. From chip to chiller.