AI-Driven Autonomous Data Center Management Framework for Enhancing Computing Efficiency in Saudi Arabia’s Digital Infrastructure

Data centers are becoming strategic infrastructure for cloud services, artificial intelligence, digital government, finance, healthcare, and industrial platforms. Their growth, however, intensifies the coupled challenges of computing utilization, cooling energy, equipment reliability, service-level compliance, and carbon-aware operation. This review synthesizes recent peer-reviewed research on artificial intelligence for autonomous data center management and develops a Saudi-oriented framework for improving computing efficiency without treating energy optimization as an isolated facilities problem. An integrative review approach is used to examine evidence on workload prediction and scheduling, virtual-machine consolidation, thermal modeling, cooling control, network energy management, anomaly detection, predictive maintenance, digital twins, and reinforcement learning. The synthesis shows that the strongest results emerge when information-technology and facility systems are optimized jointly rather than through independent controllers. Machine learning improves state estimation and forecasting, while reinforcement learning and model-predictive methods enable adaptive decisions under changing workloads and thermal conditions. Yet deployment remains constrained by simulation-heavy evaluation, limited transfer across facilities, weak explainability, fragmented telemetry, safety concerns, and the absence of common multi-objective benchmarks. For Saudi Arabia, these limitations are particularly important because rapid digital-infrastructure expansion must coexist with hot-climate cooling demands, resilience requirements, and national ambitions for efficient, sustainable digitalization. The paper proposes a layered autonomous management framework combining a trusted telemetry fabric, digital-twin state estimation, predictive intelligence, constrained optimization, closed-loop control, and governance. The framework emphasizes measurable computing efficiency, thermal safety, reliability, and accountable autonomy. Research priorities include real-world testbeds, climate-aware transfer learning, multi-agent coordination, uncertainty-aware control, carbon-intensity integration, and standardized evaluation across computing, cooling, and service outcomes.