How is AI Forcing Hyperscalers to Rethink Data Centre Power?

The intersection of AI and data centre infrastructure is creating a fundamental shift in how the technology industry approaches energy consumption. As AI workloads demand increasingly sophisticated computing environments, operators are being forced to reimagine both their hardware configurations and power delivery systems.

Digital Edge, a Singapore-based data centre platform, released its ESG Report 2026 in early 2026, demonstrating how AI-driven demand is reshaping the technical architecture of digital infrastructure. The report reveals that integrating sustainability into data centre operations is no longer a peripheral consideration but a core technical requirement for supporting next-generation computing workloads.

The company operates across key Asia-Pacific markets including Singapore, Japan, South Korea, Thailand, Indonesia and India, positioning it at the centre of the region’s digital transformation. This geographic footprint provides critical insights into how different regulatory environments and energy markets are responding to the infrastructure demands of AI deployment.

AI adoption drives infrastructure transformation

According to Digital Edge, the proliferation of AI and cloud computing technologies is fundamentally altering the power intensity requirements of data centre facilities. As machine learning workloads and large language models require more computational resources, operators must expand capacity while simultaneously managing greenhouse gas emissions and total energy consumption.

This technical challenge represents the sector’s most significant infrastructure transition in recent years, driven by the computational intensity of AI workloads and their associated power requirements. Digital Edge has committed to achieving 100% renewable or carbon-free electricity across its operations by 2030, though the report acknowledges that progress has been constrained by the rapid acceleration of AI-driven demand and the uneven availability of renewable energy across different regional markets.

The company’s approach structures ESG considerations around three core pillars: Respect for Resources, Respect for People and Communities, and Respect for Transparency. These frameworks aim to ensure that environmental stewardship and governance principles are integrated into both strategic technology planning and daily operations. This governance structure enables consistent decision-making across diverse markets whilst accommodating regional variations in regulatory requirements and stakeholder expectations.

Digital Edge secured a long-term renewable power purchase agreement covering up to 83 MW of solar capacity for its Mumbai campus, expected to avoid approximately 100,000 tonnes of CO₂ emissions annually. This agreement represents one of the largest renewable energy commitments in India's data centre sector and demonstrates how strategic partnerships with energy providers can deliver both operational and environmental benefits.

“Our India initiatives mark a major milestone in how we approach sustainability at scale,” says Andrew Rigoli, Chief Sustainability Officer of Digital Edge, in the report. “From pioneering large-scale recycled water development to securing long-term renewable energy, our collaboration with customers, partners and communities is proving that responsible infrastructure and performance go hand in hand.”

The Mumbai deployment also serves as a template for replicating renewable energy strategies across other markets where grid infrastructure and regulatory frameworks support long-term power purchase agreements. This approach to partnership development enables operators to secure predictable energy costs whilst advancing decarbonisation objectives.

Advanced cooling technologies emerge

Energy efficiency has become a critical technical consideration as AI workloads require higher-density computing environments. Digital Edge targets a Power Usage Effectiveness of 1.25 or better for new data centres, significantly below the global industry average. This metric measures how efficiently a facility converts incoming power into computing output versus overhead cooling and systems.

To achieve these performance benchmarks, the company has adopted a Flexible Data Center design approach, emphasising modular construction, standardised configurations and scalable infrastructure. These design principles reduce build times while improving efficiency and ensuring consistent performance across different markets.

Facilities are increasingly being engineered to support liquid cooling technologies, which are becoming essential for managing the thermal output generated by AI processing workloads. Traditional air cooling systems struggle to dissipate heat from high-density GPU clusters and AI accelerator chips, making liquid cooling a technical necessity rather than an optional enhancement.

The shift towards liquid cooling requires substantial changes to facility design, including reinforced flooring to support heavier equipment, enhanced leak detection systems and revised maintenance protocols. These infrastructure modifications represent significant capital investments but are essential for supporting the power densities associated with AI computing, which can exceed 50 kW per rack compared to traditional workloads of 5-10 kW per rack.

“A key milestone in 2025 was the rebranding of our Indonesian operation to Digital Edge Indonesia, alongside the launch of our AI-ready CGK hyperscale campus in Jakarta,” says John Freeman, CEO of Digital Edge, in the 2026 ESG Report. “Representing a US$4.5bn investment, the campus is designed to deliver 500 MW of IT capacity, scalable up to 1 GW.”

The CGK hyperscale campus serves as a major example of how data centre architecture is evolving to support AI workloads. With planned capacity scalable to 1 GW, the facility integrates energy-efficient design principles with operational resilience and advanced cooling infrastructure. The campus design incorporates redundant power systems, diverse network connectivity and modular deployment capabilities that enable rapid scaling in response to customer demand.

Technology enables resource efficiency

Beyond energy systems, the report highlights how technology is being deployed to address water consumption challenges. Data centre cooling impacts local water resources, particularly in regions experiencing high water stress. Digital Edge has implemented a large-scale recycled water initiative, deploying up to 10 million litres of treated greywater per day for cooling systems. This reduces dependence on potable water while maintaining efficient operations in water-constrained environments.

The company’s Green Finance Framework aligns capital allocation with technology infrastructure objectives, funding projects that deliver measurable environmental benefits such as renewable energy adoption and green building certifications. Through this framework, Digital Edge secured significant green financing, including large-scale green loans to support data centre developments in markets such as South Korea and Indonesia.

This financing structure demonstrates how ESG performance metrics are becoming integral to capital markets, with lenders increasingly linking interest rates and covenant terms to sustainability targets. The framework establishes clear criteria for eligible projects and requires regular reporting on environmental impact metrics, creating accountability mechanisms that extend beyond internal governance structures.

“We continue to target 100% renewable or carbon-free electricity across our operations by 2030,” says John in the report. “In 2025 and early 2026, we secured landmark green financing across Korea and Indonesia totalling almost US$1.25bn, reinforcing the alignment between capital and sustainable infrastructure.”

By combining renewable energy procurement with strategic technology partnerships, Digital Edge demonstrates how data centre operators can respond to the growing computational demands of AI while advancing technical efficiency goals. The integration of circular economy principles, including material reuse and recycling strategies, further demonstrates how technology infrastructure can scale while minimising environmental impact.

In 2025, a significant portion of operational waste was diverted from landfills, reflecting the effectiveness of these technical and operational practices in a rapidly evolving industry landscape. These waste management initiatives include partnerships with certified recycling facilities and programmes to refurbish decommissioned equipment for secondary markets, extending the useful life of technology assets whilst reducing the environmental footprint of infrastructure refresh cycles.