MESH: A Cyber-Physical Energy Architecture for Sustainable and Resilient Infrastructure

(DER). At the same time, large-scale digital infrastructures—most notably data centers—are creating unprecedented concentrations of electricity demand. These facilities operate with high load factors, require continuous power delivery, and increasingly shape the operational realities of modern power systems. As artificial intelligence accelerates and computational workloads expand, the interaction between digital infrastructures and electric grids is becoming more complex, more dynamic, and far more consequential than in previous eras.

Against this backdrop, this paper introduces a new conceptual and architectural approach centered on the integration of physical electricity flows (“Watts”) and cyber-based control intelligence (“Bits”). This combined perspective, referred to as the Watts – Bits concept, emphasizes that the future stability and efficiency of energy systems will depend not only on the availability of generation and transmission resources, but also on advanced digital coordination across diverse assets, users, and regions. Traditional grid planning—relying mainly on supply-side reinforcement—is no longer sufficient in a landscape with volatile renewable generation, geographically mismatched supply and demand, and sharp increases in power-intensive digital loads.

To address these challenges, this paper proposes the MESH (Machine-learning Energy System

Holistic) architecture as a visionary conceptual framework. While its constituent technologies are evolving, MESH itself represents a target state for a decentralized, autonomous, and cooperative energy-management system. MESH leverages machine learning, real-time data acquisition, and workload-aware optimization to match energy consumption with available renewable supply across time and space. Rather than managing electricity demand as a fixed, rigid quantity, MESH enables dynamic adaptation by shifting computational workloads to regions or periods with favourable energy conditions. This transforms data centres and other large consumers from passive demand points into flexible, system-supportive participants.

The MESH approach extends beyond conventional demand response by embedding intelligence directly into digital infrastructure operations. It allows workloads to be scheduled, migrated, or throttled based on renewable generation forecasts, grid congestion metrics, marginal emission factors, and locational system constraints. In doing so, it enhances system resilience, reduces the need for costly grid reinforcements, and facilitates higher penetration of intermittent renewables without compromising energy reliability or computational performance.

Ultimately, the MESH architecture represents a fundamental shift in how energy and digital systems are conceptualized. By tightly integrating physical power systems with cyber-driven coordination mechanisms, they provide a pathway toward a more flexible, decarbonized, and resilient energy ecosystem—one capable of supporting both the rapid growth of digital demand and the accelerating transition to renewable energy.