India’s data center market has moved beyond being a future growth story. It is expanding in real time at a pace and scale that draw global capital, reorder the competitive landscape, and force infrastructure operators to rethink where the next phase of AI compute will be built. The market reached an estimated value of approximately $10.48 billion in 2025 and is projected to grow to $27.2 billion by 2032, reflecting a compound annual growth rate of around 14.6%. Capacity will rise from 5,450 MW in 2026 to more than 15,000 MW by 2031. Mumbai’s under-construction capacity now ranks sixth globally, surpassing London and Dublin. For an infrastructure market that barely featured in global rankings five years ago, these figures indicate a structural shift rather than a cyclical uptick.
The momentum is being driven by a convergence of forces that are reinforcing each other simultaneously. AI workloads are pushing rack densities from conventional 8 kilowatts to 25 to 40 kilowatts, which is beyond air cooling limits and requiring liquid cooling investment that adds up to $800,000 per megawatt in capital expenditure. Additionally, India’s Digital Personal Data Protection Act is creating compliance-driven demand for domestic data infrastructure across banking, telecom, healthcare, and government sectors. Together, these forces are compressing the timeline between investment decision and construction start in ways that are producing one of the fastest capacity buildout cycles in the Asia-Pacific region.
Where the Investment Is Coming From
Foreign direct investment has become the primary engine behind India’s data center buildout, but the conviction behind that capital is what distinguishes this cycle from earlier rounds of India tech investment. Between 2020 and April 2025, India attracted approximately $14.7 billion in data center investments, with foreign institutional investors contributing nearly 86% of the total. Microsoft has committed $17.5 billion to India. Google finalized a $15 billion partnership with Adani for cloud regions. Amazon Web Services expanded its Mumbai region by 40% in January 2025 and continues to add availability zones. Moreover, the Indian government now permits 100% foreign direct investment in data centers, and the Union Budget for 2026 to 2027 introduced a 20-year tax holiday for foreign cloud providers using India-based infrastructure.
Domestic players are scaling in parallel and creating a dual-engine growth model that global capital alone cannot create. AdaniConneX has set a target of 1 gigawatt of capacity backed by a $10 billion investment roadmap. Reliance Industries has announced a 3 gigawatt data center campus in Gujarat that would triple current national capacity, powered by renewable energy and Nvidia AI semiconductors. CtrlS has committed $2 billion for AI-ready green campuses. Yotta Data Services has ordered more than 16,000 Nvidia H100 GPUs to power its Shakti Cloud with plans to scale to over 32,000 units. As explored in our analysis of India’s vertically integrated AI infrastructure model, the combination of domestic operator ambition and foreign hyperscaler capital is creating an infrastructure stack that no single player could build alone.
The GPU and Cloud Demand Picture
GPU deployment in India is accelerating faster than most global observers anticipated. The IndiaAI Mission has allocated 34,000 GPUs for domestic AI development. Cloud providers are interoperating with this government-backed compute pool while simultaneously expanding their own GPU cluster capacity within India-based regions. Reliance Jio, in partnership with Nvidia, is deploying GH200 Grace Hopper Superchips and DGX Cloud services to establish India’s first GPU-as-a-service cloud infrastructure. Consequently, India’s GPU-capable cloud capacity is expanding from a near-standing start to a significant regional presence within a two to three year window.
Cloud-based deployment models are expected to capture roughly 50% of India’s data center market in 2026 as enterprises rapidly shift workloads to scalable environments. Large enterprises drive 60% of demand and require extensive infrastructure for AI applications across banking, telecom, and e-commerce. However, as we have shown in our analysis of India’s data center transmission problem, power transmission and grid infrastructure constrain this demand rather than capital or intent.
State-level policy, renewable energy procurement, and behind-the-meter generation are now addressing that constraint and making previously limited markets viable for hyperscale development. Additionally, our coverage of why India’s data center boom is an execution problem shows that delivery timelines, not announcement volumes, will determine how much of this capacity actually materialises.
The Cities Shaping the Next Phase
Mumbai remains the dominant market with approximately 41% of total installed capacity among India’s top seven cities. However, the next phase of growth is redistributing across a broader set of markets driven by land cost, power availability, and renewable energy access. Bengaluru shows the highest forecast growth rate at a compound annual rate of nearly 24%, supported by low renewable tariffs and multiple hyperscale projects. Chennai anchors five active submarine cable systems and is emerging as the connectivity hub for latency-sensitive workloads. Hyderabad is attracting AI-focused campus development including CtrlS’s immersion-cooled AI Mega Campus and Oracle’s dual-region infrastructure plan.
The emergence of Tier 2 markets represents a further evolution of India’s data center geography. Gujarat, powered by Reliance’s 3 gigawatt campus entering phased commissioning from 2026, is becoming a significant incremental contributor. Andhra Pradesh’s Data Centre Policy explicitly targets 200 megawatts of additional capacity with land allotment and fiscal support. These markets offer the power availability, land scale, and renewable energy access that primary metros are increasingly unable to provide at commercially viable cost.
Furthermore, construction costs in India average around $6.64 per watt in Mumbai, which is 30 to 40% below comparable markets in the US and China. That structural cost advantage is durable and is one of the primary reasons global operators are treating India not just as a market to serve but as a strategic infrastructure base to build for the long term. As our analysis of how sovereign wealth funds are reshaping AI infrastructure investment showed, the patient capital backing these India bets is making multi-decade infrastructure commitments that short-cycle investors cannot match.
