A new white paper by International Data Corporation (IDC), a global technology research and advisory firm, has revealed that the rapid growth of artificial intelligence workloads is exposing weaknesses in traditional data centre maintenance models, raising the risk of system failures.

The report, sponsored by Schneider Electric, highlighted how surging rack densities, fragmented complex multi-vendor environments, and a shortage of skilled technicians are creating conditions where system failures are harder to predict and more costly when they occur.

According to the white paper, Rack power densities in AI-focused deployments have climbed from about 15kW per rack in conventional facilities to as high as 300–600kW, significantly increasing the potential impact of equipment failure. This is further compounded by many operators expanding capacity by acquiring and upgrading existing sites, often inheriting equipment with little operational history.

“When operators acquire existing facilities rather than build from scratch, they introduce unknown equipment configurations from multiple vendors, with no operational history, requiring immediate integration with asset performance management systems,” said White Paper author Luis Fernandes, Senior Research Manager, IDC.
The report warned that in such environments, traditional maintenance schedules are no longer sufficient to ensure reliability. “In this environment, calendar-based maintenance is no longer fit for purpose,” it stressed.

Against this backdrop, IDC stressed the need to adopt condition-based maintenance (CBM) as a way to reduce risk by identifying early signs of failure.
According to the IDC White Paper, early adopters of AI-powered CBM report fewer manual interventions, OpEx reduction, less unplanned downtime, asset lifetimes extended, and overall better efficiency.

“Condition-Based Maintenance (CBM) is an optimised operating model for AI-era infrastructure that reduces manual interventions, lowers OpEx, and extends asset lifecycle,” Fernandes said.
“By scaling predictive analytics to correlate behaviour across every vendor, asset, and failure trajectory, CBM enables operators to build machine-driven, human-validated system intelligence.”
To support this innovation, Schneider Electric introduced an AI-enabled energy intelligence and expert oversight, EcoCare, to track asset and systems behaviour against operating boundaries, identifies deviations, and predicts fault trajectories in advance.

The Global Head of Services, Schneider Electric, Jerome Soltani, said: “By combining remote monitoring capabilities with AI-assisted orchestration, you can gain insights regarding the health of your assets and systems, and get an early identification of abnormal behaviour that might precipitate a failure.”

He further stressed that this will ensure that downtime is minimised, while ensuring that equipment that is working within specification is not disturbed or needlessly addressed.