HCLTech Announces INR 3,500 Crore Investment to Expand AI Data Centre Infrastructure in India

July 22, 2026

HCLTech's INR 3,500 crore investment to expand AI data centre infrastructure in India reflects the rapid transformation of India's digital infrastructure landscape. As AI adoption increases, large-scale facilities require robust utility planning, high-capacity power systems, advanced cooling solutions, and reliable infrastructure design to ensure operational efficiency. In July 2026, HCLTech's board approved an investment of up to INR 3,500 crore to establish full-stack AI data centres in India with an initial capacity of 50 MW.

The announcement came alongside HCLTech's Q1 FY2027 results, the company's strongest quarter in recent history: quarterly revenue of INR 34,579 crore, up 13.9% year-on-year; net income of INR 4,624 crore, up 20.3%. HCLTech's data centre investment is both a commercial response to surging AI demand and a strategic bet on India's emerging role as a global AI infrastructure hub.

What HCLTech's Full-Stack AI Data Centre Means

HCLTech's announcement is specific in one important way: the investment targets full-stack AI data centre capability. This means the facility will not simply provide colocation space or raw compute. It will combine domestic compute capacity with GPUs, AI models, and managed services, delivering a vertically integrated platform from infrastructure through to AI application delivery.

This is architecturally different from a conventional enterprise data centre. A full-stack AI facility hosts high-density GPU clusters, NVIDIA H100 or B200-class accelerators, that operate at power densities of 30 to 100 kilowatts per rack, compared to the 5 to 10 kilowatts per rack typical of standard compute infrastructure. At those densities, Traditional air cooling alone is generally insufficient for high-density GPU deployments.

Liquid cooling systems, direct-to-chip cooling loops or rear-door heat exchangers, are required to manage thermal loads without compromising system stability. The electrical infrastructure must support not just the peak load of a fully populated GPU rack cluster but the redundancy, bypass, and switching requirements that enterprise-grade AI operations demand.

The INR 3,500 crore investment is the physical infrastructure required to capture the next tranche of that advanced AI revenue, bringing compute capacity onto Indian soil where data sovereignty requirements, low-latency AI inference needs, and government digital infrastructure mandates create sustained demand for sovereign AI facilities.

Why This Investment Arrives at a Critical Moment for AI Data Centre Expansion in India

India has 3% of the world's data centre capacity but hosts nearly 20% of the world's data. That structural gap is the commercial opportunity that HCLTech, Reliance, Adani, NTT DATA, and every major global hyperscaler is now racing to address. India's data centre capacity grew from approximately 375 MW in 2020 to approximately 1,500 MW in 2025, a fourfold increase in five years. A further fourfold increase to 6,000-8,000 MW is projected by 2030.

The government has reinforced this trajectory with direct policy support. A 20-year tax holiday for hyperscalers using Indian data centres to service global clients, one of the most commercially significant fiscal incentives any government has offered for data centre infrastructure growth in India, reduces the cost of capital for long-horizon infrastructure investments.

Data centres have received infrastructure status, unlocking access to long-tenure project financing. The National Data Centre Policy framework is being updated to create a cleaner regulatory environment for facility development and grid connection. And Minister Vaishnaw has explicitly set a USD 200 billion AI infrastructure target for India, with data centres as the primary physical layer.

HCLTech's AI Data Centre Expansion in India joins a wave of concurrent investments that define the market context. Reliance is building a 168 MW AI data centre for Meta at Jamnagar. TCS and TPG's HyperVault platform is deploying AI-focused liquid-cooled infrastructure.

Adani Group has committed USD 100 billion in renewable-powered AI data centres by 2035. Microsoft has invested USD 3 billion. AWS has committed USD 12.7 billion by 2030. AirTrunk has entered India through the acquisition of Lumina Cloudfare. Each of these players is building in the same physical environment and competing for the same grid connection capacity, cooling water availability, and engineering talent.

The Infrastructure Challenge Behind Every Megawatt of AI Capacity

The commercial excitement around India's AI data centre market frequently glosses over the engineering complexity required to deliver each megawatt of AI-grade compute. A 50 MW AI facility is not 50 MW of power in a building. It is a precisely engineered system of interconnected subsystems, grid connection and switchgear, high-voltage transformers, uninterruptible power supply systems, backup generation, power distribution units down to the rack level, precision cooling infrastructure, chilled water loops or liquid cooling circuits, building management systems, fire detection and suppression, physical security, and the network infrastructure connecting all of it to the outside world, all designed to operate continuously with no single point of failure.

Data Centre Power and Utility Infrastructure planning for an AI facility begins with the grid connection. A 50 MW facility requires a dedicated grid connection of 60-70 MW including losses and redundancy, a transmission infrastructure investment that involves DISCOM approvals, substation capacity verification, dedicated feeder design, and in many cases new substation construction.

The cooling system must be specified for the worst-case ambient temperature of the site over its 20-year operational life, not the average temperature today, and must incorporate redundancy to ensure that a cooling system failure during a peak summer heat event does not trigger a thermal shutdown of the compute layer. For liquid-cooled GPU clusters, the plumbing architecture, coolant loops, manifolds, distribution temperatures, and return conditions, must be designed from the infrastructure concept stage, not added as an afterthought once the facility is built.

These infrastructure design decisions become significantly more difficult and costly to modify once construction begins. The power density assumption embedded in the facility's transformer sizing, electrical distribution architecture, and cooling capacity determines what the facility can serve for its entire operational life. A facility designed for 10 kW per rack cannot economically be upgraded to serve 50 kW per rack GPU clusters.

As AI accelerator hardware continues to increase in power density with each new chip generation, the facilities that will retain their commercial relevance are those whose infrastructure was designed with headroom for the next generation, not just the one being deployed at commissioning. This is where AI Infrastructure Development in India requires the most careful upfront engineering discipline.

HCLTech committed INR 3,500 crore for 50 MW of AI compute in India. The megawatts are easy to announce. The power systems, cooling infrastructure, and grid connections that make each megawatt operational, that is where the engineering work is.

IMARC Engineering's Perspective

HCLTech's INR 3,500 crore commitment makes explicit what every major player in India's data centre market understands: AI workloads require a fundamentally different infrastructure design from conventional cloud computing. At IMARC Engineering, we support data centre developers, operators, and hyperscalers with the utility planning, civil and structural engineering, electrical and mechanical design, and EPCM project execution that AI-optimised facilities require.

The data centre power and utility infrastructure demands of a 50 MW AI facility, high-voltage grid connection, dedicated transformer capacity, uninterruptible power supply architecture, backup generation, liquid cooling distribution loops, chilled water systems, and fire suppression matched to high-density GPU environments, must all be designed simultaneously and integrated from the concept stage.

A facility designed for AI workloads from the ground up achieves better PUE, lower operational costs, and higher tenant confidence than a general-purpose data centre retrofitted for GPU density.

As India’s data centre market expansion accelerates toward USD 13.1 billion by 2034 and AI infrastructure development in India drives each new generation of facility above the power density of the last, the engineering quality of the infrastructure layer will determine which projects earn the enterprise and sovereign AI contracts that India's ambitions require.