A Mission-Critical engineering specialist, Raja Kumara Swamy Donthula has explained that the rapid growth of Artificial Intelligence (AI) is placing new demands on data centres, with electricity, cooling, control systems and infrastructure reliability becoming increasingly important to facilities supporting intensive computing workloads.

Donthula said that as computing requirements increase, engineers are required to consider how the different systems supporting a data centre work together throughout its lifecycle, from design and construction to commissioning and operations.

He noted that modern data centres depend on electrical distribution, standby generation, uninterruptible power supply (UPS) systems, cooling infrastructure, controls, life-safety systems and communications infrastructure.

“While these systems may be designed and delivered through separate engineering workstreams, their interaction becomes particularly important when a facility is tested and prepared for operation.

“The relationship between the systems becomes most apparent during commissioning, when equipment and infrastructure that may have been designed and installed separately must demonstrate that they can operate together under defined conditions,” he said.

According to Donthula, commissioning activities can include multiple testing levels, factory and site acceptance activities, quality assurance and quality control walkdowns, as well as operational-readiness documentation.

He said the increasing density of computing infrastructure is making this coordination more important, as higher computing requirements can affect electrical capacity and thermal management at the same time, while control systems and redundancy arrangements must respond appropriately as operating conditions change.

His words: “For data-centre operators, commissioning provides an opportunity to establish whether design requirements have been maintained through construction and whether interconnected systems are ready for operational use.

“The growth of hyperscale infrastructure has further increased the importance of these considerations. As facilities expand to accommodate cloud services and AI workloads, engineering requirements extend beyond server capacity to the physical systems needed to maintain continuous operation.”

Donthula said AI workloads ultimately depend on facilities capable of supplying large amounts of power, managing heat and maintaining operations when normal conditions are interrupted.
He added that as computing densities increase, power and cooling are therefore becoming a more prominent part of the wider AI infrastructure conversation.

He also identified predictive maintenance, digital twins, multi-agent systems and AI-assisted decision support as emerging technologies that could be applied to data-centre infrastructure.
“Data centres already generate substantial operational information through electrical systems, cooling equipment, control platforms, sensors and maintenance processes. The emerging question, he said, is how this information can be used more effectively to support engineering and operational decisions.

“Predictive systems, for example, could help engineers identify changes in equipment behaviour before they develop into more serious operational issues, while digital twins could provide virtual representations of physical infrastructure for analysis, planning and scenario evaluation.
“Multi-agent systems could allow specialised AI systems to focus on areas such as electrical conditions, thermal behaviour or maintenance information, with their outputs brought together to support human decision-making,” he said.

However, Donthula stressed that the application of these technologies in mission-critical infrastructure would still require clearly defined operating boundaries, traceable outputs and appropriate human oversight.

Greater automation does not automatically translate into greater reliability, particularly where safety, engineering standards and accountability remain central to operations.
The physical infrastructure supporting AI is also becoming an important consideration as cloud services and AI applications expand across Africa.

According to Donthula, the availability of dependable power, cooling and connectivity will remain important to the development of digital infrastructure.

As AI continues to drive demand for computing capacity, Donthula said the infrastructure supporting that growth will require greater coordination between physical systems, operational data and digital technologies noting that the challenge for data-centre operators and engineers will be to adopt new capabilities while maintaining the reliability, safety and accountability expected of mission-critical facilities.