Energy Efficiency Challenges for Data Centres in the AI Era

Data centres are the unseen powerhouses of our digital world, supporting everything from streaming services to online transactions. However, the growing demands of artificial intelligence (AI) are significantly increasing the energy these facilities consume, posing fresh sustainability challenges.

As AI technologies advance, the need for more computing power and storage is rising rapidly, with generative AI alone potentially consuming up to 33 times more energy than traditional software. This surge in demand is putting immense pressure on data centres to operate more efficiently while minimising their environmental impact.

“While data centres have historically focused on energy efficiency through advancements in cooling, server technology and renewable energy, the increasing demand for water to cool these facilities poses a new and critical challenge,” said Simon Yeoman, CEO of Fasthosts.

All Fasthosts data centres are now powered by 100% renewable energy, but Simon highlights that water usage is becoming an equally important issue.

Balancing energy and water consumption

Data centres have long been known for their substantial energy use. Despite improvements in efficiency, the relentless growth of AI is pushing energy consumption to unprecedented levels. Beyond energy, water has emerged as a crucial resource, particularly for cooling purposes, adding a new layer of complexity to the sustainability challenge.

“The escalating demands of data processing and AI require robust cooling systems, which in turn lead to higher water usage,” Simon said. “This dual challenge of energy and water consumption presents a substantial sustainability hurdle for the industry to navigate.”

Innovative solutions, such as liquid cooling technologies, are being explored to address this issue. These systems use closed-loop mechanisms that drastically reduce water usage while enabling heat reuse.

By circulating a coolant liquid through server components, these systems absorb and dissipate heat efficiently, conserving water and enhancing cooling effectiveness.

In regions with suitable climates, air-cooled systems offer a more sustainable alternative. These systems utilise ambient air to cool data centres, reducing reliance on water-based cooling methods.

 “Combining air cooling with advanced air filtration techniques can further enhance sustainability by cutting down water use and keeping temperatures within optimal ranges,” Simon added.

Existing water-based cooling systems can also be optimised through water recycling and reducing evaporation. 

“This way, a data centre in a drought-prone area can reuse water within its cooling cycle and minimise water loss,” he continued. “This approach conserves water and helps the data centre remain efficient and operational even during water shortages.”

Embracing AI for smarter data centres

As the adoption of AI grows, there is an urgent need for a paradigm shift in how data centres operate. The focus must move beyond just reducing kilowatt-hours or litres of water used; instead, resources must be employed as efficiently and effectively as possible to maximise both performance and sustainability.

To tackle these challenges, modern data centres are increasingly relying on advanced tools for real-time monitoring and optimisation. Simon highlights the concept of data enmeshing, which integrates data from various sources — such as power consumption, cooling efficiency, and environmental conditions — providing a comprehensive view of operations. 

“This approach helps spot inefficiencies, such as overworked cooling systems, and can make targeted improvements,” he said.

Capacity planning is another critical aspect of efficient data centre management. By predicting future needs and ensuring the infrastructure can handle increased loads without over-provisioning, data centres can avoid wasting energy on underutilised servers. Server virtualisation is also key, allowing multiple virtual servers to run on a single physical server, which significantly reduces the number of physical servers needed and, consequently, cuts energy consumption and cooling requirements.

Moving towards a sustainable future

Beyond the immediate steps to improve efficiency, several other strategies can help prioritise sustainability in data centres. Transitioning to renewable energy sources, such as partnering with companies that generate renewable electricity or investing in on-site solar panels, is crucial in reducing reliance on fossil fuels and minimising environmental impact.

Sustainable design practices, including modular construction and the use of eco-friendly materials, also contribute to long-term sustainability.

“Modular construction allows for phased expansion, making each new addition as energy efficient as possible,” Simon said. “It goes without saying that using eco-friendly materials reduces the overall carbon footprint of the data centre, provided they are sourced in a responsible way.”

The relationship between AI and data centre sustainability is both challenging and full of potential. While AI drives up energy and water consumption, it also offers the tools to transform resource management within data centres. 

“AI is not just a contributor to the data centre sustainability challenge; it can also be a powerful ally,” Simon concluded. 

By embracing AI and maximising its potential, the data centre industry can continue to innovate sustainably.

**************

Make sure you check out the latest edition of Energy Digital Magazine and also sign up to our global conference series - Sustainability LIVE 2024

**************

Energy Digital is a BizClik brand.