As artificial intelligence expands across Asia’s digital economy, carriers face a new strategic constraint: energy. What was once a background operational cost now shapes network architecture, capital investment priorities, sustainability roadmaps, and even competitive positioning.
From hyperscale AI data centers to dense 5G wireless layers and distributed edge platforms, communications infrastructure is entering a stage where energy efficiency is no longer an option but a foundation for growth.
Across the Asia-Pacific region, carriers are modernizing networks, deploying automation tools, and integrating renewable energy strategies to ensure that increased traffic demand does not translate directly into increased power consumption.
Growth in AI infrastructure is reshaping communications energy requirements
AI workloads are significantly increasing the energy intensity of digital infrastructure across Asia. Unlike previous traffic growth cycles driven primarily by video streaming, AI requires continuous compute availability across cloud and edge environments, forcing operators to rethink where and how they deploy processing power.
The proliferation of AI infrastructure is fundamentally changing communications energy demands. Distributed AI systems require unprecedented amounts of connection bandwidth between distributed data centers, driving massive new fiber additions and power-hungry network equipment. Bandwidth purchased for data center connectivity surged nearly 330% from 2020 to 2024 due to hyperscale expansion and AI developments. Meanwhile, metro dark fiber purchases increased by 268% from 2023 to 2024, and the total number of metro fibers increased by more than 600%. This includes a 2,300% increase in the largest single metro contract by fiber count.
This expansion is especially important because the energy demands of AI data centers drive geographic dispersion across multiple energy grids and require enhanced data center interconnect (DCI) infrastructure that did not previously exist. Additionally, AI cabinets now consume 48 to 120 kilowatts of power, compared to 6 to 10 kilowatts of power for traditional server racks, requiring a re-architecture of cooling and network infrastructure. Additionally, 800 gigabits per second optical transceivers come with significant energy costs that require sophisticated cooling solutions.
According to GSMA estimates, energy consumption represents approximately 20-40% of communications network operating expenditure. As mobile traffic continues to grow across the digital economies of Southeast and South Asia, improving efficiency is becoming essential not only for sustainability but also for margin protection.
Communications networks are increasingly evolving into distributed digital infrastructure platforms that support inference at the edge, industrial automation, robotic control, and immersive enterprise services. As AI adoption accelerates, carriers are increasingly integrating compute, cloud, transport, and access networks to improve workload placement efficiency and reduce unnecessary transmission overhead.
AI is also becoming the most powerful energy optimization tool in the telecom industry
AI has emerged as one of the most effective mechanisms in communication to reduce energy consumption while increasing infrastructure demands.
Machine learning platforms are currently being deployed to dynamically optimize wireless transmission parameters, cooling systems, and traffic routing across large network footprints. Across the region, similar automation-driven approaches are supporting predictive maintenance and smarter infrastructure usage. These developments point towards increasingly autonomous and energy-aware networks where performance and power consumption are managed together rather than separately.
Globe Telecom in the Philippines has introduced AI and machine learning-based energy management across its radio access network, allowing it to automatically adjust power usage according to traffic demand. SK Telecom and Ericsson signed a memorandum of understanding to promote research and testing of AI-powered RAN until 2031, confirming that AI-powered RAN integrates AI and communications to improve performance while enabling energy savings.
SoftBank’s urban trials further confirmed that the AI-RAN system supports high-density traffic scenarios while handling significantly higher workloads while maintaining power consumption similar to current RAN systems. SoftBank also worked with AGC to simplify the antenna configuration of the base station, reducing the number of antenna elements to one-eighth of the conventional one, and as a result, was able to reduce power consumption by up to one-eighth. NTT DoCoMo, in collaboration with Nokia Bell Labs and SK Telecom, conducted a field trial using AI to optimize transmit and receive signals under various wireless propagation conditions.
While the expansion of AI infrastructure will create immediate power demands on telecommunications networks, AI-driven network optimization, especially through intelligent RAN technology, will allow operators to simultaneously reduce baseline energy consumption by managing resources more efficiently.
Back to infrastructure basics to power the AI era
Infrastructure transformation remains one of the fastest ways for carriers to improve energy efficiency. Replacing legacy wireless equipment with new generation systems reduces consumption at the site level, and virtualization allows computing workloads to be consolidated and dynamically allocated based on demand.
However, without cleaner power strategies, efficiency gains alone will not be enough. In many of Asia’s emerging markets, diesel generators continue to support remote base stations and local connectivity infrastructure, exposing operators to both emissions pressures and fuel price fluctuations.
As a result, renewable energy integration is becoming a central pillar of communications sustainability programs.
With the increasing deployment of solar PV towers and the use of hybrid energy systems to reduce dependence on diesel generation and improve resiliency in geographically complex and off-grid regions, operators are seeking more stable, low-carbon energy solutions.
Related: Decarbonizing ASEAN communications: 5 pathways to a greener and smarter network
As carriers deepen their partnerships with hyperscale cloud providers, connectivity infrastructure is increasingly intertwined with high-density computing platforms.
For example, Singtel’s DC Tuas facility, Nxera, is designed to support AI and high-performance computing workloads while improving energy efficiency through advanced facility design and smart optimization technology. Meanwhile, NTT Group is expanding its data centers in line with its decarbonization goals, including the introduction of renewable energy and highly efficient facility design. NTT Docomo Business has also introduced the Green Nexcenter platform to support low-carbon data center operations aligned with the infrastructure requirements of the AI era.
Taken together, these developments demonstrate how carriers are moving beyond providing connectivity to actively participate in the region’s digital energy transition.
Energy efficiency is becoming a competitive differentiator in the AI economy
As communications networks evolve into distributed computing platforms that support enterprise AI workloads, sustainability performance is increasingly shaping customer expectations. Companies pursuing digital transformation strategies are now evaluating connectivity providers not only on performance and coverage, but also on emissions transparency and renewable energy integration.
Going forward, the transition to 6G is expected to place a greater emphasis on energy-aware network intelligence built directly into the infrastructure architecture.
Telcos that invest early in integrating automation, virtualization, and renewable energy are positioned to lead the next phase of AI-enabled connectivity.
