Grid-interactive data centers for sustainable and resilient energy systems: Evolution, trends, and future directions

Data centers (DCs) are becoming some of the most energy-intensive and grid-influential infrastructures in modern power systems, driven by the explosive growth of artificial intelligence (AI) workloads, cloud computing, and data-intensive digital services. Conventional facilities were designed primarily for reliable computing service delivery, but high-density AI workloads and fast load transients now demand more flexible and controllable energy-computing integration. Grid-interactive data centers (GDCs) provide a pathway for this transition by coordinating computational workloads, energy storage, renewable generation, cooling infrastructure, and grid-facing control functions. However, existing studies continue to address energy management, workload flexibility, demand response, power quality, dynamic load modeling, and ancillary-service participation as largely disjoint research streams. To address this fragmentation, this review provides a systems-level synthesis of next-generation GDCs for sustainable and resilient energy systems. It traces the evolution from conventional and energy-aware data centers to AI-optimized DCs and next-generation GDCs. The review identifies key enabling technologies, including grid-forming interfaces, flexible workload scheduling, distributed energy storage systems, AI-based control, digital twins, and cybersecurity, and introduces a six-layer GDC architecture that integrates physical infrastructure, energy systems, communication and control, intelligence and co-optimization, grid interfaces, and stakeholder-market interaction. The review further examines operational management strategies, grid integration challenges, sustainability metrics, and deployment barriers, and highlights digital-twin-based operation, coordinated smart inverters, and fast multi-timescale optimization frameworks as critical research directions for future GDC deployment.