By: Ben Selier - Vice President, Secure Power, Anglophone Africa at Schneider Electric

The best time to plant a tree, the old saying goes, was 20 years ago, and the second-best time is today. The same concept holds true to sustainability in the data centre industry. While the ideal moment to prioritise, green initiatives may have been earlier, the time is now due thanks to the proliferation of artificial intelligence. 

More than ever, double down on sustainability efforts and decouple the growth of AI from data centre energy consumption. 

AI mirroring energy use

There is no reason to think demand for high-volume processing capabilities will slow down anytime soon. The numbers tell a sobering story. Deloitte forecasts that data centres will account for  2% of global electricity consumption in 2025, about 536 terawatt-hours (TWh). With the proliferation of AI and other power-hungry applications, that figure is expected to double in just five years to 1,065 TWh. 

Investment in mission-critical infrastructure is fuelling this growth. The global data centre market size is calculated at $125.35 billion in 2024 and is projected to rise to $364.62 billion by 2034, with a CAGR of 11.39% over that period, largely driven by gen AI. 

While this growth unlocks new possibilities for industries and communities across the world, it also poses significant sustainability challenges; how data centres source, manage, and consume energy whilst also ensuring their underlying infrastructure is sustainably sourced, managed and consumed.

What’s at stake?

Even if it were possible to halt AI’s rapid ascension, it would probably be a moot point. It is reshaping industries, enhancing productivity and expected to contribute $15.7 trillion to the global economy by 2030 —  more than the economies of China and India combined. 

However, the ramifications of unrestrained energy consumption are far-reaching. From exacerbating climate change to depleting critical resources, the stakes extend beyond immediate business concerns. Inaction is not an option. 

Six practical solutions

Above all else, a commitment to sustainability must be practical. To address growing emissions in the face of AI adoption, operators must face three key challenges head-on: mitigate the negative environmental impacts of AI growth by reinvigorating sustainability strategies, focus on Scope 3 emissions, and meet reporting and regulatory requirements. Here are six ways to tackle these hurdles.

1. Assess existing infrastructure and processes for efficiency gains: Data centre operators should thoroughly evaluate current operations. Identifying inefficiencies in cooling systems, server utilisation, and energy management can pave the way for significant improvements. 

2. Adopt efficient infrastructure: The transition to energy-efficient infrastructure is crucial. Modern servers, advanced cooling systems and renewable energy sources are critical to sustainability. Additionally, designing facilities with energy efficiency in mind from the outset can substantially reduce carbon footprints. Retrofitting older facilities with advanced technologies can also yield energy savings. For example, implementing liquid cooling systems instead of traditional air cooling can reduce energy consumption.

3. Partner with suppliers who prioritise decarbonising their supply chain and mitigating Scope 3 emissions: Sustainability isn’t achievable in isolation. A significant portion of a company’s carbon footprint stems from the activities of its suppliers and partners. By collaborating with suppliers committed to decarbonising their supply chains, operators can effectively address Scope 3 emissions, the largest source of greenhouse gas emissions for businesses. 

Suppliers who release Environmental Product Declarations empower data centre operators to take greater control of their Scope 3 emissions and build a more sustainable future. 

4. Leverage advanced reporting tools: businesses must be able to monitor and manage their operations to control their energy footprints effectively. These tools collect and analyse data and automate tasks to optimise performance. Real-time monitoring systems enable operators to identify inefficiencies and implement corrective measures promptly.

5. Optimise resource allocation with AI and ML: Ironically, AI can be a part of the solution to the sustainability challenges it contributes to. ML algorithms can optimse resource allocation, improve efficiency and predict maintenance needs, reducing waste and emissions. Predictive analytics can also forecast energy demand, allowing for better integration of renewable energy sources and minimising reliance on fossil fuels.

6. Continuously reassess and improve: Sustainability is an ongoing effort. It is essential to continuously reassess operations and adjust when necessary. Operators must embrace a cycle of evaluation, innovation, and implementation to stay ahead of challenges and opportunities. Benchmarking against industry standards and adopting best practices will help to maintain progress in pursuit of sustainability.