India's Data Centre Pipeline Surges to Multi-Gigawatt Scale as AI Reshapes South Asia's Digital Infrastructure

Subheadline: Installed capacity is projected to rise from about 1.6 GW in mid-2026 to 6 GW by 2029, drawing more than USD 50 billion in hyperscaler commitments and forcing a rethink of power, policy and regional supply chains.

Executive Summary

India's data centre market entered the second half of 2026 with roughly 1.6 GW of installed capacity and a development pipeline that, on current commitments, is projected to reach 6 GW by 2029, according to JLL's India data centre mid-year report. First-half 2026 absorption reached 101 MW, more than 20% above the three-year average, with pre-committed hyperscale capacity accounting for 82% of the total. The demand mix has shifted decisively toward high-density artificial intelligence workloads rather than conventional colocation.

The capital required to deliver that pipeline is substantial. JLL estimates cumulative investment of about USD 110 billion by 2029, spanning construction, power infrastructure, cooling systems and IT equipment. Global hyperscalers have committed more than USD 50 billion to self-build approximately 1.4 GW of capacity, equivalent to nearly 30% of new additions. A 20-year tax holiday extending to 2047 for foreign cloud providers, announced in the Union Budget 2026-27, materially lowers operating costs for internationally domiciled operators.

Policy and supply-side risks are rising in parallel. Record-low vacancy rates are squeezing availability for enterprise tenants that must compete with hyperscalers for land, power and interconnection. The Digital Personal Data Protection Act introduces mandatory breach reporting and penalties of up to INR 250 crore, with full compliance obligations taking effect from May 2027. Power procurement, cooling technology and grid access are emerging as the binding constraints on delivery timelines rather than capital or demand.

Introduction

Data centre capacity has become a proxy for a country's capacity to participate in the AI economy. For South Asia, where digital services exports, remittance-linked consumption and a young workforce underpin growth, the ability to host compute domestically determines whether value accrues locally or is captured offshore.

India's position is unusual among emerging markets. It combines scale, a large domestic user base, an established submarine cable landing ecosystem and increasingly explicit fiscal support for digital infrastructure. The mid-2026 data indicate that these factors are now converting into physical capacity at a pace that has few regional parallels. The question for operators, investors and policymakers is no longer whether demand exists, but whether power, land, water and regulatory capacity can be delivered on schedule.

Main Analysis

Demand: AI workloads reset the absorption curve

First-half absorption of 101 MW exceeded the three-year average by more than 20%, a figure that understates the shift in composition. Pre-committed hyperscale capacity represented 82% of absorption, and the workloads driving those commitments are predominantly AI training and inference, which demand materially higher power density per rack than traditional enterprise deployments. Mumbai and Chennai led deliveries, with pipeline capacity secured through power approvals rather than speculative construction.

This pattern has two consequences. First, vacancy is structurally low because capacity is largely spoken for before completion. Second, the technical specification of new builds has changed: liquid cooling, direct-to-chip and immersion cooling are moving from pilot installations to design defaults, and site selection increasingly depends on the availability of firm, scalable power rather than proximity to end users alone.

Supply: construction methods have been industrialised

India's operators have delivered roughly 60% compound annual growth in capacity since 2021, an outcome partly attributable to process innovation. Pre-fabricated steel structures, modular electrical yards and repeatable design templates have compressed construction cycles. Project sizes have expanded from around 20 MW to blocks of 50 MW and above to accommodate high-density workloads. In the first half of 2026, 85 MW was delivered, with Mumbai accounting for 48% and Chennai 50% of supply additions, reflecting some overlap as campuses are completed in phases across both markets.

Capital: hyperscalers move from leasing to self-build

Global hyperscalers are committing more than USD 50 billion to build approximately 1.4 GW of self-build capacity by 2029. Self-build changes the competitive structure of the market. It reduces the share of demand available to third-party colocation providers, but it also expands the addressable market for construction contractors, power equipment suppliers, cooling vendors and electrical engineering firms.

Established hubs remain Mumbai as the core market, Chennai with eight submarine cable landings, and Pune as an emerging AI compute cluster. Hyderabad is attracting gigawatt-scale commitments, Visakhapatnam offers coastal advantages for cooling and connectivity, and Delhi NCR is being positioned around sovereign cloud requirements.

Policy: fiscal incentives and a phased data regime

The Union Budget 2026-27 introduced a 20-year tax holiday until 2047 for foreign cloud providers, a measure that improves the after-tax economics of large campuses and makes India more directly comparable with competing jurisdictions in Southeast Asia and the Gulf.

The Digital Personal Data Protection Act establishes a staged compliance calendar: board constitution and administrative definitions from November 13, 2025; consent manager registration and obligations from November 13, 2026; and full compliance from May 13, 2027, covering notice requirements, security and breach reporting, child data protection, obligations for Significant Data Fiduciaries, data principal rights and cross-border transfer mechanisms. Penalties reach INR 250 crore. For operators, this creates compliance cost but also a degree of legal certainty that supports long-horizon investment.

Sustainability and power availability

Power is the principal constraint. India is targeting 500 GW of renewable energy capacity before 2030, and is expanding nuclear capacity from 8.78 GW to 22.38 GW by 2031-32, with an aspiration of 100 GW by 2047. Small Modular Reactors are being considered as off-grid options for large campuses. In the interim, grid interconnection queues, state-level tariff structures and water availability for cooling will determine which projects are delivered on schedule and which are deferred.

Regional Impact

Economic growth and employment. A build-out requiring USD 110 billion of investment by 2029 represents a significant capital formation cycle, generating demand for civil construction, electrical engineering, cooling systems and skilled technical labour. Data centre operations create comparatively few permanent jobs per megawatt, but the construction and equipment supply chain is labour-intensive and geographically dispersed.

Regional competitiveness. India's capacity trajectory positions it as the largest data centre market in South Asia by a wide margin. For neighbouring economies, this has two effects: it creates dependence on Indian-hosted compute for latency-sensitive services, and it establishes a benchmark for power pricing, land allocation and regulatory clarity against which smaller markets will be assessed.

Trade integration and supply chains. Chennai's submarine cable cluster connects India to Southeast Asia and onward routes, reinforcing digital trade corridors that complement physical ones. Equipment procurement for cooling, power distribution and networking increasingly draws on regional manufacturing, creating incremental demand for electronics and precision engineering capacity within South Asia.

Foreign investment. The combination of tax incentives, self-build commitments and low vacancy is likely to sustain inbound capital. It also raises the stakes for investment policy coherence across states, since incentives are administered at state level and vary considerably.

Technology adoption and innovation ecosystems. Domestic demand for AI compute can support domestic model development, research commercialisation and university partnerships. Without sovereign or locally accessible capacity, that work migrates to overseas platforms, taking associated talent and intellectual property with it.

Financial inclusion and infrastructure. Digital public infrastructure, payments systems and fintech services benefit from lower domestic hosting costs and improved latency, which indirectly supports the expansion of financial services into underserved areas.

Long-term resilience. Concentration of capacity in a small number of coastal clusters creates exposure to climate events, water stress and grid disruption. Geographic diversification into Hyderabad, Visakhapatnam and Delhi NCR partially mitigates this, but the concentration risk remains material.

Strategic Insights

For corporate tenants, the practical implication is that enterprise colocation capacity will remain tight while hyperscale pre-commitments dominate new supply. Multi-year contracting, early engagement on power and a willingness to consider tier-two locations are becoming standard procurement practice.

For investors, the opportunity set extends beyond real estate. Power infrastructure, cooling technology, electrical equipment, structured financing and energy contracting are all components of the value chain, and several are less contested than land acquisition. Ownership models that separate the shell from the operating platform are likely to attract infrastructure capital with longer horizons.

For policymakers, the priority list is operational rather than declarative: accelerated grid interconnection, clearer state-level tariff treatment for large loads, water-use frameworks for cooling, and predictable implementation of the data protection regime. Incentives have succeeded in attracting capital; execution capacity now determines whether that capital converts into operating capacity.

Competitive risks are identifiable. Self-build by hyperscalers compresses the third-party leasing market. Power cost inflation would erode the value of tax incentives. Prolonged delay in grid connections would push delivery schedules beyond the periods on which customer commitments are based. Regional competition from Gulf and Southeast Asian markets, which offer comparable incentives with faster permitting, remains a live constraint on pricing.

Future Outlook

Over the next three to five years, India's data centre capacity is projected to reach approximately 6 GW by 2029, with the composition increasingly oriented toward AI workloads. Absorption is likely to remain concentrated in Mumbai, Chennai and Hyderabad, while Visakhapatnam and Delhi NCR develop as secondary nodes.

Power will be the decisive variable. Renewable capacity additions, nuclear expansion toward 22.38 GW by 2031-32, and the possible deployment of Small Modular Reactors will determine whether the pipeline is delivered on schedule. The May 2027 full-compliance date for the data protection framework will test operator readiness on security, breach reporting and cross-border data transfer mechanisms, and will shape how multinationals structure their Indian data residency arrangements.

Trade integration across South Asia is likely to deepen unevenly. Submarine cable investment and cross-border power trade discussions provide the connective tissue, but regulatory divergence between countries limits how far compute capacity can be shared regionally. The more probable near-term outcome is a hub-and-spoke model in which India hosts the capital-intensive AI infrastructure and regional economies participate through services, equipment supply and connectivity.

In manufacturing, the build-out creates sustained demand for cooling systems, switchgear, prefabricated structures and power electronics, some of which have viable prospects for domestic production under existing industrial policy. In financial markets, infrastructure investment trusts and green bond issuance are plausible channels for absorbing the capital requirements of the sector.

Conclusion

India's data centre expansion is no longer a speculative pipeline. Capital is committed, land and power approvals are secured for a substantial share of the projects, and demand is pre-contracted. The binding constraints have shifted to power availability, water and permitting efficiency, and to whether the compliance framework accompanying the Digital Personal Data Protection Act can be implemented without slowing deployment.

The strategic significance extends beyond the data centre sector. Compute capacity is an input to AI adoption, digital services exports and industrial modernisation. How efficiently India converts capital and policy support into operating capacity over the remainder of the decade will influence its competitive position not only in Asia but in global digital services markets.

Key Takeaways

• India's installed data centre capacity of about 1.6 GW in mid-2026 is projected to reach 6 GW by 2029, requiring an estimated USD 110 billion of investment.
• First-half 2026 absorption of 101 MW exceeded the three-year average by more than 20%, with pre-committed hyperscale capacity representing 82% of demand.
• Hyperscalers have committed more than USD 50 billion to self-build approximately 1.4 GW, equal to nearly 30% of new capacity, shifting competitive dynamics away from third-party leasing.
• A 20-year tax holiday to 2047 for foreign cloud providers, announced in the Union Budget 2026-27, improves the economics of large campuses but does not resolve power and land constraints.
• Digital Personal Data Protection Act obligations phase in through May 2027, with penalties of up to INR 250 crore; compliance readiness will influence multinational data residency decisions.
• Power procurement, cooling technology and grid interconnection are now the principal determinants of delivery timelines.
• Regional spillovers are likely to be concentrated in equipment supply, connectivity and services rather than shared compute capacity.

SEO Keywords

South Asia Economy, India data centre market, data centre investment India, Artificial Intelligence infrastructure, hyperscale data centres, Foreign Direct Investment India, digital economy South Asia, data centre capacity 2029, Digital Personal Data Protection Act, AI compute, power infrastructure India, renewable energy India, regional trade, industrial development, technology adoption, capital markets, energy transition, emerging markets, economic growth, supply chain

Suggested URL Slug

india-data-centre-pipeline-ai-investment

Sources

• JLL, India data centre 2026 mid-year report, 31 August 2026: https://www.jll.com/en-in/insights/market-dynamics/india-data-centers