By Canninah Dladla, Cluster President for Sub-Saharan Africa, Schneider Electric
Artificial intelligence (AI), cloud computing and advanced analytics are accelerating demand for data centre capacity at an eye-watering pace. But it’s a double-edged sword – the one side is creating is tremendous opportunities for innovation and economic growth, the other brings scrutiny around environmental footprint, particularly the consumption of energy and water.
According to the World Economic Forum’s Nature Positive: Role of the Technology Sector report, the technology sector has a critical role to play in reducing pressure on natural systems while enabling digital transformation.
The opportunity here is not only to build more infrastructure – the now‑famous quote, ‘if you build it, they will come’ undoubtedly holds true – but also to focus on how we build it, creating data centres that contribute to both a resilient and sustainable future.
The true measure of tomorrow’s data centre lies not in how much compute it delivers, but in how responsibly it delivers it.
The changing data centre
Liquid cooling has become a major talking point in this epoch of AI data centres. Often viewed as a new technology designed for this era, its foundations actually stretch back decades. Early IBM mainframes in the 1960s and Cray supercomputers used liquid cooling with great success to manage increasing computational demands.
The difference today is scale; generative AI has fundamentally changed the way data centres are designed and operated. Modern accelerated compute servers combine multiple graphics processing units (GPUs), central processing units (CPUs) and data processing units (DPUs) to process increasingly complex workloads.
These systems deliver exceptional performance but also generate significantly more heat than traditional cloud servers.
Staggeringly, AI-optimised servers, with massive GPU clusters, consume 3,000 to 5,000 watts (W) per server, with specialised training clusters exceeding 10,000 W. In contrast, conventional CPU-based cloud servers generally consume 150W to 500W.
Therefore, as rack densities continue to rise, traditional air cooling is reaching its physical limits.
Why liquid cooling is becoming essential
As mentioned, air cooling has supported the data centre industry for decades, but AI and high-performance computing workloads are creating a new thermal challenge.
It is here where liquid cooling has emerged as a truly viable alternative due to its superior heat dissipation capabilities and lower energy consumption compared to air-based methods. It offers numerous benefits:
- Closer to the heat source – liquid cooling tackles thermal load at its origin, circulating coolant through cold plates attached directly to high‑performance chips.
- Density unlocked – it enables higher compute density while reducing reliance on energy‑hungry fans, ensuring consistent performance for demanding AI workloads.
- Single‑phase reliability – liquid cooling’s direct‑to‑chip cooling with liquid‑only circulation delivers predictable thermal performance, simpler maintenance, and improved reliability for mission‑critical environments.
Turning sustainability ambition into measurable impact
This exciting transition from air cooling to advanced liquid cooling demonstrates how innovation can address both performance and sustainability challenges.
A recent life-cycle assessment, conducted by Schneider Electric and Microsoft, found that moving from air cooling to advanced liquid cooling technologies such as direct-to-chip and immersion cooling can significantly reduce environmental impact.
Significantly, the study showed reductions in greenhouse gas (GHG) emissions of up to 21%, energy demand reductions of up to 20%, and blue water consumption reductions of up to 82% when transitioning to liquid cooling solutions, depending on the energy scenario.
The above is undoubtedly noteworthy and encouraging, particularly in light of increased scrutiny to develop, build and operate data centres that conserve energy and water.
Importantly, in regions such as Sub-Saharan Africa, where energy resilience and resource efficiency remain key priorities, technologies that enable greater output from existing infrastructure will be increasingly important.
Building the nature-positive data centre
With innovation like liquid cooling stepping into the fore, the next generation of data centres will be designed as intelligent, adaptive systems that work more harmoniously with the environments around it,
This will require a holistic approach, thus combining efficient electrical infrastructure, renewable energy integration, intelligent energy management, advanced cooling architectures and digital monitoring.
Through our EcoStruxure vendor‑neutral, open architecture, Schneider Electric helps data centre operators and hyperscalers design, build, and operate infrastructure that balances performance with sustainability.
We offer AI-enabled energy management, grid-interactive UPS systems, storage solutions and advanced cooling technologies provide operators with the visibility and control needed to optimise resources.
Ultimately, the future data centre must support the digital ambitions of society while respecting the natural systems that sustain it.
