The rapid surge in artificial intelligence (AI) and AI workloads is fundamentally reshaping the scale and nature of digital infrastructure, placing data centres at the centre of technological growth across economies. However, this growth is no longer being discussed purely in terms of capacity expansion or technological capability. Unlike traditional digital services, AI models demand significantly higher compute intensity, continuous processing and scalable storage capabilities. The expansion of data centre capacity is therefore increasingly tied to the availability of reliable, scalable and sustainable energy infrastructure. Without addressing energy, resource and environmental constraints, the scalability of AI infrastructure risks becoming unsustainable both economically and ecologically.
The sustainability discourse within the digital infrastructure domain has evolved significantly over time. Earlier, sustainability was largely framed through operational efficiency metrics such as power usage effectiveness and resource optimisation. While these metrics remain important, the industry perspective has matured towards a broader decarbonisation agenda that integrates sustainability across the entire life cycle of infrastructure development.
This transition reflects a shift from compliance-driven approaches to design-led sustainability. Instead of treating environmental performance as an operational afterthought, sustainability considerations are now embedded at the conceptualisation and design stages of data centre development. Decisions related to site selection, equipment choices, cooling systems, materials and energy sourcing are increasingly influenced by sustainability goals from the outset. Building infrastructure with sustainability embedded from the initial stages allows operators to deploy advanced technologies, optimise energy consumption and reduce long-term environmental impact.
Equally important is the growing recognition that sustainability cannot be achieved by a single stakeholder. The ecosystem approach has gained prominence, where hyperscalers, infrastructure providers, regulators, power suppliers and technology partners collectively shape sustainability outcomes. Customer expectations, particularly from hyperscalers and enterprise users, are increasingly driving sustainability requirements in contracts and operational frameworks, further reinforcing the integration of environmental priorities into infrastructure planning.
Energy transition
Energy has emerged as the most critical variable in the sustainability equation of AI infrastructure. The rapid growth of data centres is directly linked to electricity consumption, making the transition towards renewable and non-fossil energy sources a strategic necessity rather than an optional goal.
The integration of renewable energy into data centre operations is gaining traction. Industry initiatives are moving towards higher renewable energy adoption targets, with some facilities already operating on fully green power. This transition is supported by policy frameworks, open access energy mechanisms and increasing investments in renewable generation capacity.
Water resource constraints
While energy dominates sustainability discussions, water usage has emerged as an equally critical concern, particularly in regions facing resource stress. Cooling requirements for hyperscale facilities can lead to significant water consumption, creating visible trade-offs between technological expansion and local ecological balance. In water-scarce geographies, this challenge becomes even more pronounced, requiring innovative cooling strategies and resource management mechanisms.
Therefore, infrastructure planning must account for local resource availability, groundwater stress and environmental sensitivities to ensure that digital growth does not disproportionately burden host regions.
Importantly, sustainability solutions cannot adopt a one-size-fits-all model across diverse climatic and geographic conditions. Regions differ significantly in terms of water availability, land constraints, energy mix and environmental priorities. As a result, infrastructure design must be context-sensitive, balancing energy efficiency, water conservation and community needs through technically informed and evidence-based approaches rather than rigid uniform standards.
Coordinated frameworks
The evolving sustainability agenda has highlighted the need to bridge the gap between policy intent and on-ground execution. While policy support for digital infrastructure has strengthened in recent years through infrastructure recognition, incentives and state-level initiatives, faster implementation mechanisms remain essential to match the pace of AI-driven demand. Delays in approvals, transmission infrastructure and grid connectivity can slow down otherwise viable sustainable projects, even when renewable energy capacity is available. This creates a structural mismatch where green energy generation exists, but its integration into digital infrastructure is constrained by logistical and regulatory bottlenecks.
A coordinated platform for government-industry dialogue is therefore emerging as a key enabler of sustainable digital growth. Collaborative frameworks allow stakeholders to exchange best practices, align sustainability targets with practical realities and create evidence-based benchmarks that reflect diverse regional conditions. Such coordination also reduces the risk of fragmented policy approaches across jurisdictions, where varying state-level norms on energy access, renewable procurement or sustainability compliance can lead to inconsistencies in implementation.
Further, the rapid expansion of AI-ready infrastructure has led to the emergence of data centre clusters in key urban and semi-urban regions, often driven by access to connectivity, power infrastructure and market demand. However, this clustering has largely evolved organically rather than through structured planning. While such clusters generate operational efficiencies, the absence of pre-planned digital infrastructure zones can slow large-scale expansion and create uneven resource distribution. Strategically developed infrastructure zones with pre-provisioned power, connectivity and land resources could significantly accelerate the pace of deployment while improving sustainability outcomes. Planned clusters allow better grid integration, optimised transmission planning and efficient resource allocation, reducing the environmental and logistical costs associated with fragmented infrastructure development. This approach also aligns with the growing need for scalable infrastructure capable of supporting exponential AI workloads.
Grid integration remains a particularly critical area of focus. While renewable generation capacity is expanding, last-mile interconnection and transmission readiness determine how effectively sustainable power can be delivered to infrastructure hubs. Strengthening transmission networks, enabling faster grid connectivity and improving energy market flexibility can help unlock the full potential of renewable energy adoption within the digital infrastructure ecosystem.
Sustainability in digital infrastructure involves complex trade-offs between energy efficiency, water usage, land constraints and community impact. Decisions that optimise one sustainability metric may create pressures on another, particularly in resource-constrained regions. For instance, cooling systems that maximise energy efficiency may require higher water consumption, while water-efficient systems may increase power usage. Managing these trade-offs requires context-specific design strategies that prioritise long-term ecological balance alongside operational performance.
Geographic diversity further complicates sustainability planning. Different regions present varying constraints related to climate, water availability, land use and grid stability. Infrastructure solutions that are optimal in one region may be unsuitable in another, reinforcing the need for flexible and evidence-driven sustainability frameworks. Local sourcing of materials, adaptive cooling strategies and region-specific infrastructure designs are increasingly being adopted to address these variations.
Technological innovation
Technological innovation is playing a central role in reconciling AI expansion with sustainability objectives. Contrary to concerns about rising energy intensity, advances in chip design, infrastructure architecture and operational optimisation are steadily improving efficiency levels across digital infrastructure systems. Over time, the industry has demonstrated that technological progress can significantly enhance compute output per unit of energy, indicating strong potential for long-term efficiency gains even as AI workloads grow.
The sustainability journey is also closely linked to the broader evolution of cloud infrastructure, where earlier transitions from inefficient on-premise server environments to large-scale cloud systems already delivered measurable efficiency dividends. AI infrastructure appears to be following a similar trajectory, with current inefficiencies likely to reduce as innovation accelerates across hardware, cooling systems and operational models. This suggests that AI, while resource-intensive today, has substantial scope to become more energy-efficient as technologies mature and scale.
In sum
Research, development and ecosystem-wide innovation are therefore essential to ensuring that infrastructure growth remains sustainable. Improvements in semiconductor efficiency, cooling technologies, renewable integration and intelligent infrastructure management collectively contribute to reducing the environmental footprint of large-scale digital systems. Importantly, sustainability is increasingly recognised not as a regulatory burden but as a strategic and economic advantage, aligning environmental responsibility with operational efficiency and long-term cost optimisation.
Infrastructure design is also evolving to reflect these priorities. Sustainability considerations now influence material sourcing, supply chain decisions, cooling technologies and life cycle management practices. Closed-loop systems, advanced chillers and hybrid cooling solutions are being adopted to minimise resource consumption while maintaining high performance standards. These design innovations demonstrate that sustainability and scalability can coexist when embedded at the architectural level rather than treated as retroactive adjustments.
The way forward
India is uniquely positioned within the Asia-Pacific region to lead the development of sustainable digital infrastructure. The country’s increasing non-fossil energy capacity and large-scale renewable deployment provide a favourable foundation for aligning infrastructure growth with sustainability goals.
Simultaneously, the scale of digital adoption and projected data consumption underscores the urgency of building infrastructure that is both resilient and environmentally sustainable. As AI missions, cloud expansion and digital transformation initiatives accelerate, demand for compute infrastructure is expected to grow significantly, further reinforcing the need for sustainable planning and execution.
The path forward lies in execution efficiency, ecosystem collaboration and long-term strategic alignment between infrastructure growth and sustainability objectives. With strong policy intent, expanding renewable capacity and increasing industry commitment to decarbonisation, the digital infrastructure sector is moving towards a model where AI growth and sustainability are not competing priorities but mutually reinforcing goals. If supported by coordinated policy frameworks, technological innovation and context-sensitive design, sustainable digital infrastructure can emerge as the foundation for scalable, responsible and future-ready AI ecosystems.
Based on a discussion among Aruna Sundararajan, Chairperson, BIF; Bimal Khandelwal, CEO, STT GDC (India); Alexander Smith, Principal, Global Infrastructure and Energy, Google; Sukrit Anand, VP, Investments and Strategy, Digital Connexion; and David Skelton, Head, APDCA Secretariat, at the India AI Impact Summit 2026
