Chris Wellfair, Projects Director at Secure I.T. Environments, explains how stepping towards AI-driven systems can supercharge efficiency and reliability – without letting the technology run wild.
We’re all used to hearing about AI in the news, and 2025 has of course continued that trend. Just look at the ‘shock’ news of China’s launch of DeepSeek in January, which has shown that these innovations are not limited to the US tech giants – much to their, and their investors, disappointment.
Aside from the headline grabbing news about how AI will change lives or take over the world, it is having an impact on the data centre industry. AI in itself is driving the construction of more powerful data centres, but in a way that won’t grab the headlines, it is proving very beneficial to improving the efficiency of data centre operations.
AI in data centre operations
AI technologies are being integrated into data centre management to enhance operational efficiency, reduce costs, and support sustainability goals critical to both private and public sector initiatives. These AI-driven data centres can autonomously monitor and adjust power consumption, streamline cooling systems, and predict hardware failures before they occur. However, meeting AI’s computational demands requires data centres to rethink infrastructure, embracing scalable and flexible designs that can handle high-performance computing (HPC) environments.
There are a few key areas where AI can improve data centre operations both in terms of efficiency and reliability:
Enhanced energy efficiency
AI-driven data centres continuously monitor energy consumption patterns and automatically adjust systems to maintain optimal energy use. This ensures that cooling and power supply systems operate with maximum efficiency, cutting down on excessive energy expenditure. As the AI era progresses, data centres will need to adopt more efficient energy management strategies, from integrating alternative energy sources to implementing dynamic power allocation systems that optimise energy use based on workload demands. This approach not only aligns with sustainability goals but also enhances cost competitiveness.
Predictive maintenance
AI tools analyse data from sensors and equipment to predict potential failures and schedule maintenance before a problem disrupts operations. A supplier can use the data from across all its installations to build an AI model that helps data centre managers minimise downtime and extend the lifetime of the infrastructure hardware in their data centre.
Optimised cooling systems
Cooling systems in data centres typically consume a significant portion of total energy. Accelerated compute systems supporting AI workloads generate enormous amounts of heat, making advanced cooling solutions, like liquid cooling, essential for maintaining operational efficiency. Liquid cooling, particularly beneficial for high-density racks and edge computing locations, is gaining traction as traditional air cooling proves insufficient for these demanding environments.
Sensors in a data centre can be used to understand how the humidity and temperature of the data centre change with workloads and external environmental factors. Over time, this information can allow an intelligent cooling system to self-optimise how it cools the environment in different scenarios.
Improved security measures
With AI’s machine learning algorithms, data centres can detect unusual activity patterns, enhancing their cybersecurity posture. Real-time analysis enables rapid responses to potential threats, which is crucial given the rising number of cyber-attacks targeting data infrastructure. AI can also be used as part of the DC security infrastructure, enabling facial and voice recognition for physical access controls, whether to applications, cabinets, or secure areas of a data centre.
One AI step at a time
AI-driven data centres are starting to transform the landscape of data centres in the UK, offering enhanced efficiency, sustainability, and security. Of course, one of the biggest worries for any data centre team is letting go – handing over control. This is an understandable response, as ultimately it is the DC team that retains operational responsibility, and nobody wants the AI going rogue! Implementation of AI systems covering any of the areas above, can be incremental, and should always be setup in ways, where outside certain parameters human approval should be sought. For example, if the AI decides it is going to turn off the cooling infrastructure for a prolonged period of time, or automatically book a preventative maintenance visit.
Every data centre team will have a view on what is the right mix of AI automation and human control, and the AI will need to earn trust. But that does not mean it should be ruled out. AI’s ability to trawl data and identify opportunities for efficiencies far surpasses anything that the operations team can achieve, and frankly they’ve usually got other things to worry about. Use AI to its strengths, to make recommendations, and in time automate. Your DC will thank you for it, and you will achieve unexpected operational efficiencies that benefit your team and the wider business.
