India’s artificial intelligence ambitions are bold, visible, and increasingly urgent. From government-led digital public infrastructure to private sector AI adoption, the narrative is one of acceleration. Yet beneath this momentum lies a quieter, less discussed reality: AI does not run on ambition alone. It runs on infrastructure, specifically, data centres.
These facilities, often tucked away from public view, are no longer passive enablers of the digital economy. They are becoming its very foundation. What was once treated as backend IT capacity is now a strategic asset shaping economic velocity, digital sovereignty, and technological competitiveness. The shift is subtle but decisive.
India has already crossed approximately 1.5 gigawatts of operational data centre capacity, with projections pointing toward 4–5 gigawatts by 2030 under realistic growth conditions and potentially much higher if infrastructure scales effectively . These numbers are not just technical milestones; they are signals of how seriously the country is beginning to treat its digital backbone.
But here lies the tension: growth in demand is outpacing the readiness of supply-side infrastructure.
AI’s hunger for compute is rewriting the rules
Artificial intelligence is not just another digital workload. It is exponentially more demanding. Training models, running inference, and supporting real-time AI applications require dense, high-performance computing environments that traditional infrastructure was not designed to handle.
Unlike earlier waves of digital growth, such as cloud migration or mobile internet expansion, AI workloads compress massive computing power into concentrated clusters. This shifts the conversation from “how many data centres” to “how capable those data centres are.”
India’s advantage lies in its scale of data generation. With platforms enabling real-time payments, identity verification, and digital services, the country has built one of the most active data ecosystems globally. But data without processing capability is an underutilised asset. If compute capacity does not grow in tandem with data generation, India risks becoming a producer of raw digital value rather than a processor of it.
Infrastructure is physical before it is digital
It is easy to think of AI and cloud computing as abstract, software-driven domains. But data centres remind us of a more grounded truth: digital systems are deeply physical. They depend on land, electricity, fibre connectivity, and water. Each of these inputs introduces constraints that are often overlooked in high-level policy conversations.
Electricity, in particular, is emerging as a decisive factor. Hyperscale data centres consume energy at levels comparable to small cities. As India’s data centre footprint expands, their share of national electricity consumption is expected to rise significantly, potentially reaching 2–3 per cent by the end of the decade .
This creates a dual challenge. On one hand, the sector must secure reliable and scalable power. On the other, it must align with sustainability goals, particularly as global scrutiny around energy-intensive AI systems intensifies.
The transition toward renewable energy sourcing is promising, but it introduces its own complexities intermittency, storage, and grid integration. The success of India’s AI ambitions may well depend on how effectively these energy questions are addressed.
Geography will shape the next phase of growth
India’s data centre ecosystem today is concentrated in a handful of metropolitan hubs. Cities like Mumbai, Chennai, Bengaluru, Hyderabad, and Delhi-NCR dominate due to their connectivity, infrastructure readiness, and proximity to enterprise demand.
This concentration has served the initial phase of growth well. But as AI applications become more real-time and latency-sensitive, the need for distributed infrastructure becomes more pronounced.
Edge computing, regional data hubs, and secondary city expansion are no longer optional, they are inevitable. The next wave of digital services, particularly in sectors like healthcare, logistics, and manufacturing, will require data processing closer to the point of use.
This is where India faces a strategic choice. It can either replicate the same concentration patterns in new forms or consciously design a more distributed infrastructure network that balances efficiency with accessibility.
Policy recognition of data centres as infrastructure has been a significant step forward. Incentives, state-level initiatives, and evolving regulatory frameworks have collectively helped attract investment and accelerate project pipelines.
With an estimated $60–70 billion in investments lined up over the next decade, the intent is clear . However, intent does not automatically translate into execution. The real test lies in coordination. Data centre development sits at the intersection of multiple systems: energy, urban planning, environmental regulation, and digital policy. Fragmentation across these domains can slow progress, even when capital and demand are abundant.
For instance, delays in power provisioning, land acquisition complexities, or inconsistencies in regulatory interpretation can create bottlenecks that ripple across the ecosystem. If India is to sustain its momentum, policy must evolve from enabling to orchestrating.
The global gap is both a risk and an opportunity
Despite rapid growth, India still accounts for only around 3 per cent of global data centre capacity . This gap is significant, but it is also instructive.
It highlights how much room there is for expansion, but it also underscores the competitive landscape. Other regions are not standing still. Data centre development globally is entering a phase defined by efficiency, sustainability, and advanced design.
India’s relative “late mover” position could be an advantage. Much of its infrastructure is yet to be built, offering the opportunity to integrate next-generation technologies from the outset rather than retrofitting legacy systems.
But this advantage is time-sensitive. The longer the gap persists, the harder it becomes to close, especially as AI capabilities become increasingly tied to infrastructure scale.
Sustainability is no longer optional
As AI scales, so does scrutiny. Data centres, particularly those supporting AI workloads, are under increasing pressure to demonstrate sustainable operations.
Water usage for cooling, energy efficiency, and carbon footprint are moving from peripheral concerns to central design considerations. Encouragingly, advancements in cooling technologies and recycling systems are beginning to reshape how facilities are built.
India’s challenge is not just to grow its data centre capacity, but to do so responsibly. This is not merely an environmental concern, it is a strategic one. Global enterprises, investors, and partners are increasingly factoring sustainability into their decisions. Infrastructure that fails to meet these expectations risks becoming stranded or underutilised.
A quiet bottleneck with loud consequences
The narrative around India’s AI future is often framed in terms of talent, innovation, and policy vision. All of these are important. But without sufficient data centre capacity, they risk becoming disconnected from execution.
AI models cannot scale without compute. Real-time applications cannot function without low-latency infrastructure. Data sovereignty cannot be ensured without domestic storage and processing capabilities.
In this sense, data centres are not just part of the AI ecosystem, they are its bottleneck. And unlike other bottlenecks, this one cannot be solved overnight. Data centres require long lead times, significant capital, and coordinated planning. Delays today can translate into constraints years down the line.
The infrastructure test of India’s AI decade
India stands at a pivotal moment. The foundations of a digital economy are firmly in place. AI represents the next leap. But that leap will not be defined solely by algorithms or applications,it will be defined by infrastructure.
Data centres may remain invisible to most, but their impact will be unmistakable. They will determine whether India’s AI push translates into scalable capability or remains an aspirational narrative.
The question is no longer whether India needs more data centres. It is whether it can build them fast enough, efficiently enough, and sustainably enough to keep pace with its own ambitions.
Because in the race to lead in AI, compute is not just an input. It is the ground on which the race is run.
