How Smart Maintenance is Shaping Data Centres of the Future

AI continues to drive unprecedented power demands across the data centre industry, leaving operators with a continued dilemma: Should they retrofit legacy facilities to be more sustainable and compliant, or take the risk of building entirely new locations?

Scale is therefore a significant consideration, with data centres needing to scale their operations, while simultaneously striving to meet their sustainability targets.

To tackle this issue, Data Centre Magazine speaks with President of ABB’s Electrification Service Division, Stuart Thompson, who shares his experience on how predictive maintenance and strategic retrofitting are enabling operators to turn aging infrastructure into competitive advantages while also dramatically reducing environmental impact.

Please introduce yourself and your role at ABB.

My name is Stuart Thompson and I’m the president of ABB's Electrification Service Division. ABB initially started out with the provision of electrical equipment, but now I’m responsible for leading a US$1.2bn business. Over time, we saw a rising demand for servicing solutions among our customers which led to the creation of a dedicated Electrification Service business to better serve them. 

As a qualified electrical engineer, I’m passionate about helping customers maximise energy efficiency and get the most out of their operational investments by improving performance and extending the lifetime of their electrical infrastructure. By enabling customers to adopt a circular mindset, my team achieves this through strategic upgrades, retrofit solutions and digital services — including cutting-edge applications of Augmented Reality and AI – all supported by specialised advisory services that deliver deeper insights for greater optimisation, availability, reliability and safety of electrical systems.

What would you say are some of the largest energy challenges currently facing the data centre industry?

Data centres today face several critical energy challenges. First is the dramatic increase in power demand driven by AI, cloud computing and digital transformation, with hyperscalers now regularly building facilities requiring over 100MW of power. As our reliance on digital technologies only continues to grow, this isn’t set to change anytime soon. 

Secondly, operators must balance this growing demand with evolving sustainability requirements which change from market to market based on the regulatory environment they’re working within. Many are committed to net-zero targets while simultaneously scaling operations, creating a significant technical challenge.

Third is the reliability imperative. Our research shows unplanned outages can cost around US$125,000 per hour, and for data centres, even minutes of downtime can mean millions in losses. Lastly, data centres increasingly face grid constraints in high-demand regions like Southeast Asia, forcing operators to look for alternative solutions to overcome the challenge of limited grid capacity while maintaining reliability.

These challenges create what we call an ‘energy trilemma’ – balancing security, affordability and sustainability of energy supply. This is where energy storage, microgrids and intelligent energy management systems become critical, enabling data centres to enhance resilience while reducing their carbon footprint.

How is the industry’s approach to energy management changing?

The data centre industry is absolutely transforming its approach to energy management. We’re seeing a shift from capital expenditure to operational expenditure models, especially through ‘as-a-Service’ offerings that maximise existing assets with minimal upfront investment.

Digital-first strategies are replacing manual processes, with real-time monitoring systems leveraging AI to continuously optimise energy usage. The use of Battery Energy Storage Systems (BESS) to provide back-up renewable power is also gaining significant traction, with $5 billion invested in 2022 alone, which is triple from the previous year.

Most importantly, we see energy evolving from a cost centre to a value generator, with facilities participating in grid services and energy markets to create new revenue streams while supporting grid stability.

How does predictive maintenance fit into this? How can this technology ensure data centre operations are more sustainable?

Predictive maintenance represents a fundamental shift from reactive approaches to data-driven anticipation of issues before they occur. I’ve seen firsthand how this shift is economically crucial, as our research has shown that waiting for equipment to fail typically costs organisations up to 10 times more than implementing proactive maintenance strategies. When we extend asset lifecycles through predictive maintenance, we’re dramatically reducing the environmental footprint associated with manufacturing and disposing of equipment. This moves us away from the wasteful ‘take-make-dispose’ mindset toward a circular economy approach.

Modernisation plays a big part in this transformation. Degraded equipment consumes significantly more power and updating these systems through targeted retrofits rather than complete replacements delivers dual benefits. By keeping modernised systems at peak efficiency, predictive maintenance directly reduces energy consumption and carbon emissions, while minimising the higher carbon footprint of emergency replacements and expedited shipping. This shows how operational excellence and environmental responsibility can go hand in hand with data-driven maintenance strategies that extend the life of existing infrastructure.

How is ABB approaching the integration of AI into maintenance systems and what impact might this have on data centre sustainability?

We’re taking a practical approach to AI in maintenance systems, focusing on specific operational challenges where it delivers measurable improvements. Our AI algorithms can now identify subtle equipment patterns that even experienced technicians might miss – detecting partial discharges in switchgear or analysing thermal patterns in transformers weeks before issues become visible through conventional methods.

What’s transformative for sustainability is how these AI systems optimise maintenance timing based not just on equipment condition, but on operational impacts and even energy market conditions. This means work can be scheduled during periods of low renewable energy production, preserving clean energy for critical operations.

I’m particularly excited about our progress with digital twins and augmented reality solutions that allow remote experts to guide on-site personnel, which I’ve seen reduce service-related emissions by up to 60% by eliminating unnecessary travel. Beyond energy savings, the extended equipment lifecycles and reduction in unnecessary parts replacements align perfectly with data centres’ increasing focus on both operational efficiency and environmental responsibility.

As AI continues to increase demand for electricity, how is ABB positioning itself to support customers with this surge, while prioritising sustainability? 

The AI-driven surge in electricity demand is staggering – it’s expected that by decade’s end, data centres in the US alone are projected to consume nearly 12% of the country’s total electricity, up from just 3.7% today. I believe this challenge requires a multifaceted approach.

We’re focusing on modernisation of existing infrastructure through strategic upgrades rather than new builds. I’ve seen our retrofitting solutions enhance energy capacity by up to 20% while extending equipment life by decades.

Digitalisation is equally critical as implementing advanced energy management systems can help to optimise operations in real-time to balance both cost and carbon impact. And finally, we’re helping customers integrate renewable energy through battery storage systems that address intermittency issues.

What I find most valuable is the real synergy between these approaches. For example, digital systems can predict renewable energy production peaks and adjust cooling systems to pre-cool facilities during these periods, reducing load when carbon-intensive power is the primary source. This integration of strategies helps meet surging electricity demands while simultaneously reducing environmental impact.

How can the data centre industry better address the carbon footprint of older data centres without requiring complete rebuilds?

The industry needs to fundamentally shift from a replacement mindset to strategic reinvention when addressing older data centres. Rather than viewing aging facilities as liabilities, we should recognise them as opportunities for innovation and transformation.

Digital retrofitting represents the cornerstone of this approach – implementing smart monitoring capabilities into existing infrastructure can reveal efficiency opportunities that have been hidden for years. When combined with targeted equipment upgrades in high-impact systems like cooling and power distribution, these facilities can achieve performance metrics that approach those of new builds.

I believe the most forward-thinking operators are now implementing comprehensive energy management ecosystems that connect every system within the facility. This holistic approach enables dynamic optimisation across all operations, aligning energy use with both business priorities and sustainability goals.

The industry must also embrace circular economy principles more broadly. By designing for component-level maintenance rather than wholesale replacement, we can dramatically extend infrastructure lifespans while reducing resource consumption and waste.

The bottom line is that decarbonising existing data centre infrastructure is fundamentally a business imperative – driving operational efficiency, reducing long-term costs and creating market differentiation, while simultaneously addressing environmental responsibilities in an increasingly resource-constrained landscape.

To read the full article in the magazine, click HERE.