As Malaysia’s AI-ready datacenter sector expands rapidly, Schneider Electric is calling on operators to rethink how critical infrastructure is maintained. In a bylined article, Ea Jia Gen, Head of Services at Schneider Electric Malaysia, argues that traditional maintenance models are no longer sufficient for the demands of modern, AI-ready datacenter facilities.
Between 2021 and mid-2025, MIDA approved RM144.4bn in datacenter and cloud computing investments in Malaysia, reflecting the country’s growing role in the regional digital economy. That scale of investment, Ea argues, brings with it a parallel responsibility: keeping those facilities running reliably as they grow larger, more complex, and more energy-intensive.
Why AI-Ready Datacenter Maintenance Needs to Change
Conventional maintenance approaches typically follow one of two models: fixed-schedule servicing or reactive response after something fails. Ea’s article points out the limitations of both. Scheduled maintenance can result in unnecessary checks, while reactive approaches allow minor problems to escalate into costly disruptions.
AI workloads intensify the pressure further. According to the International Energy Agency, global datacenter electricity demand grew by 17% in 2025, with AI-focused facilities consuming power at an even faster rate. Uptime Institute’s 2025 survey found that one in ten outages still causes serious or severe disruption, and that nearly two-thirds of operators report difficulty retaining or finding qualified staff.
The combined effect is a workforce being asked to manage increasingly complex, interconnected systems, covering power, cooling, electrical distribution, and digital monitoring, with limited time and technical resources. Traditional maintenance, built for simpler environments, was not designed for this level of interdependence.
Condition-Based Maintenance & Digital Tools For Datacenters
The alternative Ea describes is condition-based maintenance: servicing equipment based on how it is actually performing rather than according to a predetermined calendar. Connected equipment generates operational data continuously, and when that data is analyzed, teams can identify early warning signs and prioritize the most critical repairs before disruption occurs.
Ea is clear that technology is not intended to replace experienced technicians. AI can surface patterns and insights, but the judgment, context, and practical expertise needed to act on those insights still come from people. The article frames digital tools as a support layer rather than a substitute for human skill.
Schneider Electric’s approach draws on its collaborations with NVIDIA and AVEVA, encompassing validated power and cooling designs, lifecycle digital twin capabilities, and AI-supported alarm management. In practical terms, operators can use these tools to understand system performance, identify emerging risks, and determine which actions to prioritize.
The article cites one concrete outcome from Schneider Electric’s work with a global datacenter operator. By combining EcoStruxure Modular Data Centers with EcoCare services, that operator reduced intrusive on-site maintenance activities by 40% and achieved 20% operational expenditure savings compared with traditional calendar-based maintenance contracts.
Ea outlines three practical benefits of a smarter maintenance approach: reducing unexpected disruptions by catching early warning signs, helping skilled teams allocate their time to the equipment that actually needs attention, and supporting more efficient operations by keeping systems running as intended and extending equipment life.
For new datacenters, Ea recommends factoring smarter maintenance into the design stage so that data, systems, and workflows are connected from the outset. For existing facilities, a phased approach is proposed, starting with critical power and cooling assets and expanding monitoring practices across the broader facility over time.
The article concludes by framing maintenance not as a standalone service activity but as part of a wider asset lifecycle strategy. Schneider Electric argues that the strongest operational outcomes come when design, asset digitization, condition-based maintenance, modernization, and field services are integrated into a single continuous approach.
Pokdepinion: Malaysia just approved RM144 billion in datacenter investments, so naturally the next step is making sure nobody forgets to service the cooling.
