Computing Power and Energy

Computing Power and Energy Computing-Power Synergy as a Key Component of New Infrastructure

The Energy Imperative

The Energy Imperative Rapid growth: Data centers consumed about 415 TWh in 2024—1.5% of global electricity use—and demand has grown around 12% annually since 2017. 2030 outlook: Global data-center electricity consumption is projected to more than double to approximately 945 TWh. AI impact: Training and inference require dense, continuous power; a typical AI-focused data center can consume as much electricity as 100,000 households. Local pressure: Data centers are geographically concentrated, amplifying grid congestion and infrastructure requirements. New constraint: Chips alone cannot deliver capacity; generation, transmission, substations, cooling, and backup power are equally critical.

Synergy in Practice

Synergy in Practice Computing-power and energy synergy coordinates workloads, data centers, grids, and renewable resources as one integrated system. Energy follows computing: Expand reliable generation, storage, and grid capacity around major computing clusters. Computing follows energy: Locate facilities near abundant, affordable renewable energy. Workloads follow conditions: Shift flexible tasks across regions and time based on price, grid load, renewable output, and carbon intensity. Integrated scheduling: Connect cloud platforms, computing networks, power markets, and grid operators. Service differentiation: Keep latency-sensitive inference near users while moving training, simulation, rendering, and batch processing when feasible.

Integrated Architecture

Integrated Architecture Layer Core Components Primary Role Energy Renewable generation, storage, grids, microgrids Supply reliable, lower-carbon electricity Facilities Data centers, AI clusters, HPC, liquid cooling Convert electricity into computing capacity Networks High-speed computing and data networks Enable workload mobility across regions Orchestration Energy and workload scheduling platforms Optimize location, timing, cost, and utilization Governance Carbon accounting and green-power procurement Measure performance and support compliance Scheduling inputs: Capacity, urgency, latency, electricity prices, renewable forecasts, grid congestion, efficiency, and carbon intensity. Scheduling output: Assign each workload to the most suitable location and time within reliability and policy constraints.

Strategic Value

Strategic Value Stakeholder Value Created Data-center operators Lower electricity costs and higher asset utilization Grid operators Flexible demand and improved renewable integration Cloud and AI providers Greater capacity, resilience, and location flexibility Energy producers New demand and stronger long-term offtake Governments Digital growth with lower carbon and infrastructure pressure System effect: Flexible computing can absorb surplus wind and solar generation and reduce renewable curtailment. Competitive shift: Electricity availability, price, and carbon intensity increasingly shape data-center location and workload economics. Infrastructure outcome: Computing capacity becomes a coordinated resource rather than an isolated facility.

China’s Policy Direction

China’s Policy Direction Efficiency target: China aims to reduce average data-center power usage effectiveness below 1.5 by 2025. Renewable target: The plan calls for data-center renewable-energy utilization to increase by 10% annually. Hub concentration: More than 60% of newly added computing capacity is targeted for national hub regions by the end of 2025. Scale: China had more than 8.1 million data-center racks in use by the end of 2023. 2030 direction: Energy and carbon efficiency per unit of computing power should reach internationally advanced levels, with higher renewable utilization. Strategic alignment: Computing hubs and cross-regional power infrastructure can jointly connect energy-rich regions with digital demand.

Constraints to Resolve

Constraints to Resolve Grid access: Long interconnection timelines and transmission bottlenecks can delay computing projects. Reliability: Cost or carbon optimization must not weaken uninterrupted power supply. Workload limits: Latency, data sovereignty, security, and software dependencies restrict mobility. Measurement gaps: Workload-level electricity use and emissions remain difficult to quantify consistently. Resource impacts: Cooling introduces water consumption, heat management, and local environmental pressure. Capital exposure: Generation, grid, cooling, and computing assets require high investment under uncertain utilization. Coordination gap: Computing and power sectors lack unified technical, market, and operating standards.

Implementation Roadmap

Implementation Roadmap Assess: Map workloads by latency, flexibility, security, energy use, and carbon profile. Connect: Align computing expansion with generation, substations, transmission, storage, cooling, and backup capacity. Instrument: Measure workload-level energy, PUE, utilization, carbon intensity, and renewable consumption. Orchestrate: Pilot time- and location-aware scheduling for delay-tolerant workloads. Contract: Combine green-power procurement, demand response, storage, and long-term energy partnerships. Scale: Expand from individual facilities to regional computing-energy networks with shared operating standards.

The New Infrastructure Advantage

The New Infrastructure Advantage Core conclusion: The next generation of digital infrastructure will be defined by coordinated computing, electricity, networks, storage, cooling, and carbon performance. Strategic priority: Treat energy capacity as a first-class input to AI and data-center planning. Operating priority: Make flexible workloads responsive to renewable output, grid conditions, and electricity prices. Investment priority: Develop computing clusters and power infrastructure through one integrated roadmap. Governance priority: Establish shared metrics, standards, and accountability across technology and energy ecosystems. End state: Computing-power and energy synergy evolves from an efficiency measure into a foundational infrastructure capability.