As data centres continue to expand in both scale and strategic importance, operators are under increasing pressure to deliver resilience, efficiency and sustainability across every aspect of their operations. Alice Oakes, Service and Support Manager at Wilo, highlights how predictive asset management can help close the industry’s maintenance maturity gap.
Data centres are expected to set a leading standard for efficiency and sustainability. Supporting this is a complex ecosystem of assets, from white-space technologies such as servers, GPUs and storage systems to grey-space infrastructure, including cooling and water management. To maintain uptime rates of 99.999% or higher, widely regarded as the gold standard across the industry, all components must work together seamlessly.
Yet the sector is facing mounting pressure. Demand for capacity continues to accelerate, while new developments are often constrained by lengthy planning, construction and grid connection timescales. As a result, many existing facilities are being asked to do more for longer periods.
At the same time, operators are contending with skills shortages and growing expectations around environmental outcomes. Metrics including Power Usage Effectiveness (PUE) and Water Usage Effectiveness (WUE) are placing greater focus on how efficiently assets are maintained and managed throughout their lifecycle.
A higher standard for critical infrastructure
The strategic importance of data centres was formally recognised in September 2024, when the UK government designated them as Critical National Infrastructure (CNI).
This designation places data centres alongside sectors such as energy, defence, food, water and nuclear, reflecting their increasingly vital role in supporting economic activity and digital services.
For operators, this raises the bar considerably. Facilities must demonstrate resilience against a wide range of potential disruptions, from power failures and equipment breakdowns to cyber threats and wider operational incidents. The expectation is not simply to prevent outages, but to ensure sites can respond quickly and recover effectively when incidents occur.
While most facilities already operate established maintenance programmes, recent research suggests that many organisations have yet to align their asset management strategies with these heightened requirements.
The maintenance maturity gap
Research we conducted among 300 UK data centre managers reveals a significant gap between operational requirements and current maintenance practices.
Only 26% of respondents reported having a fully predictive maintenance strategy in place. By comparison, 28% described their model as preventative, while more than one in five (21%) still rely primarily on reactive maintenance.
Although preventative maintenance represents an improvement on reactive servicing, both methods have limitations. They rely on fixed schedules or responding after faults have occurred, rather than monitoring the actual condition of equipment in real time.
Predictive maintenance takes a different route. Often supported by condition-based monitoring (CBM), it continuously analyses data from critical assets such as pumps, motors and cooling systems to identify potential issues before they result in failure.
When integrated with automated alerts, maintenance scheduling and intelligent spares planning, predictive strategies can help reduce outages, extend equipment lifespans and improve both energy efficiency and sustainability outcomes.
The findings suggest much of the industry remains reliant on maintenance models that may struggle to support modern uptime expectations. Confidence levels reflect this challenge, with only 53% of respondents saying they were fully confident in their current strategy.
The cost of standing still
The consequences of maintenance shortcomings are already being felt.
According to the survey, data centres experienced an average of six hours of unplanned downtime over the past year. Nearly nine in ten respondents (88%) reported more than three hours of downtime, while more than half (52%) experienced in excess of six hours.
At a time when billions of pounds are being invested in the UK’s digital infrastructure, reducing avoidable disruption has become increasingly important. Predictive maintenance has implications beyond reliability alone, encompassing operational efficiency, resource optimisation and sustainability.
However, progress is not without challenges. Skills shortages and a lack of specialist expertise remain among the most frequently cited barriers to predictive maintenance adoption. These findings mirror broader industry concerns, with workforce capability and recruitment pressures continuing to affect day-to-day operations across the sector.
Taking action
Encouragingly, momentum for change is already building. Among organisations that have yet to implement a predictive maintenance strategy, 87% say they intend to do so within the next six months. This suggests the industry increasingly recognises the need to align maintenance practices with the expectations that come with CNI status.
Turning these ambitions into reality will require more than technology alone. Operators will also need the skills, processes and organisational support required to ensure predictive maintenance programmes deliver measurable value.
As data centres continue to underpin the UK’s digital economy, maintenance can no longer be viewed as a routine operational function. Instead, it should be recognised as a strategic enabler of resilience, efficiency and long-term success.
Closing the maintenance maturity gap is not simply about adopting new technologies. It is about ensuring critical assets are managed with the same level of intelligence, foresight and reliability that modern data centres are expected to deliver. For operators seeking to meet increasingly demanding standards, predictive maintenance is likely to become an increasingly important part of their approach.
