Introduction

Technological progress is advancing sustainability through improved resource efficiency enabled by smart devices, smart grids and intelligent systems. Although AI adoption is raising energy demand, broader electrification—from electric vehicles to modern manufacturing—remains a major driver. AI delivers broad benefits across sustainability, public health, safety and security, but these depend on modern, resilient energy infrastructure. Governments and industry must collaborate to align technological growth with dependable, sustainable power systems, as many grids remain outdated and strained.

Understanding sustainable technology

Understanding The Power Challenge

AI and other data-intensive technologies are increasing global electricity demand, which could rise 50% by 2035 without efficiency gains. AI computational requirements are doubling about every 100 days, while the technology sector contributes 1.7% of global emissions. Many power grids are ageing, with significant portions of transmission lines exceeding 25 years in age, and electrification trends are intensifying demand.

GPUs power AI due to their parallel processing, high-bandwidth memory and specialised designs, offering greater efficiency than CPUs. Industry is creating new sustainable hardware, such as neural processing units, to cut energy and water consumption.

The role of data centres in next-generation technology development

Data centres are critical for AI, cloud and digital services, consuming about 2% of global electricity, a figure that may double by 2026. Despite heavy energy use, they are major renewable-energy purchasers and have achieved strong efficiency gains: between 2010 and 2018, computing output rose 550% and storage 2,400%, but power use grew only 6%.

Energy consumption is affected by cooling, hardware efficiency, facility design, regulation, software and energy sources. AI is enhancing efficiency through workload shifting, IoT monitoring and advanced cooling. Infrastructure delays, such as multi-year waits for new transmission connections, limit access to renewable energy.

Economically, data centres could generate USD 6.3 trillion in global benefits by 2027 and supported 565,000 direct jobs in 2022, each enabling six additional jobs.

Understanding How AI Is Advancing Global Sustainability

AI drives sustainability through better demand forecasting, workload optimisation and green coding. It strengthens grid cybersecurity, supports sustainable agriculture, advances climate modelling, improves manufacturing efficiency via digital twins and enhances ESG data quality. AI-driven traffic systems can reduce emissions significantly at intersections.

On the horizon: The benefits of next-generation innovation

Emerging innovations—photonic accelerators, 3D chip architectures, edge computing and advanced cooling like liquid and immersion systems—promise further efficiency improvements.

Iti’s sustainable technology policy recommendations

Modernising The Electric Grid

Modern grids are essential for renewable integration and rising AI workloads. Over 80 million kilometres of infrastructure upgrades are needed by 2040. Policymakers should streamline regulations, incentivise private investment and expand R&D in grid-enhancing technologies, including AI-enabled workload shifting.

Expanding access to alternative low-carbon power sources

A diverse low-carbon energy mix is crucial. Governments should support AI-driven renewable optimisation, simplify licensing for advanced nuclear reactors and encourage alliances such as SMR collaborations.

Promoting resource efficiency

Data-centre operators should improve energy and water efficiency through advanced cooling, virtualisation and transparent reporting. Policymakers should promote emerging cooling technologies and harmonise sustainability metrics.

Managing carbon lifecycle

Scope 2 and 3 emissions dominate data-centre footprints. Industry is advancing carbon-aware workload scheduling, clean-energy matching and supply-chain improvements. Policymakers should harmonise carbon-accounting frameworks, support R&D and enable accurate measurement of AI’s environmental impact.