Why Hyderabad is becoming India’s data centre corridor.

From the city that wrote the software to the city that runs it

For two decades, Hyderabad’s reputation was built in office towers. HITEC City and the Financial District filled with engineers writing software for the rest of the world. The servers that software ran on mostly lived somewhere else.

That is changing quickly. Hyderabad’s live data centre capacity grew from 60.9 MW in 2022 to 151.4 MW at the end of 2025, and about 1.9 GW more is under construction, committed or planned. Only Mumbai has a larger pipeline in India, according to Knight Frank’s India Data Centre Market Update 2025. In 2026 alone, Microsoft opened its largest Indian cloud region in the city, AWS broke ground on a new facility, and TCS announced a 1 GW AI campus.

For anyone who builds mission-critical infrastructure, this matters. A corridor of this size will need hundreds of buildings engineered to tolerances most of the Indian construction industry has never been asked to meet. This essay looks at why Hyderabad is winning these projects, how it compares with India’s other hubs, and what the corridor will demand of the people who build it.

We write as an engineering practice headquartered in Jubilee Hills, a few kilometres from where much of this is happening.

Who is building in Hyderabad

The corridor is being driven by global hyperscalers, a large Indian IT company moving into infrastructure, and colocation operators such as NTT and AdaniConneX, which Knight Frank notes are scaling up in the city.

These investment figures cover different periods and geographies, so they should not be added together as a Telangana total.

The state’s own ambition is larger still. At Microsoft’s region inauguration in September 2026, IT Minister D. Sridhar Babu said Telangana has more than 300 MW operational and 2 GW under development, and is targeting 5 GW of hyperscale capacity and about US$30 billion of investment by 2029, as reported by Telangana Today. The state’s figure is statewide and counted differently from Knight Frank’s Hyderabad IT-load figure, which is why the two numbers differ.

Why Hyderabad: five reasons operators give

1. Stable ground, away from the coast

Hyderabad sits on the Deccan Plateau in seismic Zone II, the lowest-risk zone under IS 1893 (2016). That code remains in force after the Bureau of Indian Standards withdrew its 2025 revision in March 2026. Being inland also keeps it away from coastal cyclone and flood risk. Knight Frank describes this “disaster-safe geography” as one of the city’s main advantages over more expensive coastal hubs. For a building expected to run without interruption for twenty years, the ground it stands on is the first redundancy.

2. Room to build at campus scale

Modern AI data centres are not single buildings. They are campuses of several halls, substations and cooling plant, often hundreds of acres. Hyderabad’s outskirts still offer that land. The TCS campus alone takes 264 acres, and the state’s planned data centre city takes 1,500. Mumbai’s operators compete for far scarcer and costlier land.

3. A state that wants the business

S&P Global’s Data Center Policy Attractiveness India index, published in October 2026, ranks Telangana’s policy package the most attractive in the country, citing its electricity and regulatory incentives. The state’s data centre policy offers dual-grid power, renewable energy access and electricity at generation cost for eligible projects, along with incentives aimed at high-density GPU halls and liquid cooling.

4. Hyperscalers already here

AWS and Microsoft both run multi-zone cloud regions in the city. Once one hyperscaler builds, suppliers, network carriers, colocation operators and enterprise customers follow, and each new project becomes easier to justify than the last.

5. People who understand the workload

Hyderabad’s IT workforce knows the AI workloads these buildings will host, and the pipeline of people who can operate the facilities is being built deliberately. AWS’s data centre skilling programme in Telangana has trained 250 students, 110 of whom joined its operations team.

How Hyderabad compares with India’s other hubs

Mumbai remains India’s data centre capital by a wide margin. Hyderabad’s case is not about today’s capacity. It is about where the next wave is going.

Hyderabad’s live capacity is fifth among Indian cities, but its pipeline is second only to Mumbai’s. That gap between today and tomorrow is why so much construction is about to happen there.

Hyderabad’s real weakness is distance from the sea. International traffic still lands in Mumbai and Chennai, so the most latency-sensitive workloads tend to stay there. Hyderabad suits cloud regions, disaster recovery and AI training, where power and space matter more than milliseconds. HDFC Bank, for example, is one of the first customers using Microsoft’s Hyderabad region for disaster recovery.

What the corridor will demand of its buildings

The advantages that brought operators to Hyderabad are real. The constraints are just as real, and most of them land on the building.

Power is the first constraint

Telangana hit an all-time peak power demand of 19,543 MW in September 2026, up 23% on the same day a year earlier, and peak demand within Greater Hyderabad is expected to pass 5,000 MW next summer. The Union power ministry told Parliament in July 2026 that AI data centres could add 26.3 GW of national demand by 2031–32, as set out in this account of the power and water test. Every gigawatt campus therefore arrives with its own substations, dual grid feeds, backup generation and redundant distribution. A building that cannot keep its power path alive during a grid fault has failed at its main job.

Water decides the cooling strategy

Cooling uses an estimated 40% of a data centre’s electricity, and conventional evaporative cooling uses large volumes of water. Hyderabad faces a projected domestic and industrial water shortfall of 240 million gallons a day by 2027–28. That makes water-light cooling a design requirement, not a preference: closed-loop and direct liquid cooling, dry coolers, treated wastewater instead of fresh water, and metered water use. Microsoft says its Hyderabad region uses zero water for cooling, and the TCS campus is designed around liquid cooling.

AI density changes the structure

GPU halls concentrate far more power, heat and weight per rack than the server rooms of a decade ago. Liquid-cooled racks, coolant distribution units and pipework add floor loads and need routes coordinated with power and structure from the first drawing. Zone II ground lowers seismic risk, but it does not remove the need to anchor heavy equipment and design to code.

Speed must not outrun discipline

Campuses of this size are built in phases, often while earlier halls are already live. That calls for modular and pre-engineered construction, careful sequencing around live operations, and full integrated systems testing before every handover. The pressure to deliver capacity quickly is the same pressure that, in any mission-critical building, tempts contractors to skip a test or substitute a component. As we argued in our essay on honest construction, a building whose failure falls on others has to be verified, not trusted.

“If the corridor handles power and water well, Hyderabad’s advantages will compound. If it does not, they will become its limits.”

Where Prana stands

We are a Hyderabad engineering practice, and the corridor is being built on our doorstep. To be clear about our position: we do not build hyperscale data centres today. We are preparing deliberately to be one of the Indian teams trusted to build the facilities around them, and in time the halls themselves.

Our engineering studio has developed a reference design for a 5 MW modular mission-critical hall. It covers high-density thermal envelopes, N+1 electrical topologies and modular pre-cast structure, and was validated in simulation before any material was cut. On our industrial projects we already coordinate structure, MEP and airflow in a single model, mount precision equipment on vibration-isolated foundations, and have every project independently inspected.

The corridor will also need substations, utility buildings, staff facilities, warehouses and supplier plants, all built to the same standard of reliability. That is where disciplined regional builders can contribute now.