TCS unit HyperVault plans $7.4 billion AI data center campus in Hyderabad

HYDERABAD, INDIA — HyperVault and its partners plan to invest up to 700 billion rupees, about $7.4 billion, in a 1-gigawatt artificial intelligence data center campus designed for high-density computing.

HyperVault AI Data Center Limited, a subsidiary of Tata Consultancy Services, has secured 264 acres in Hyderabad for a large-scale AI infrastructure project that would rank among India’s biggest if fully developed, according to TCS.

The campus will have capacity of up to 1 gigawatt and will be developed in phases according to customer demand and technology requirements. It is intended to serve frontier AI companies and hyperscale cloud operators running graphics-processing-unit infrastructure for AI training, inference and other compute-intensive workloads.

HyperVault and its partners expect to invest up to 70,000 crore rupees, equivalent to 700 billion rupees. Reuters calculated the amount at about $7.41 billion using an exchange rate of 94.48 rupees to the dollar.

The scale makes the Hyderabad project significant beyond a conventional corporate data center. TCS said the facility is being designed for high rack densities and liquid cooling, including direct-to-chip cooling, as increasingly powerful AI accelerators drive greater electricity and heat-management requirements.

HyperVault expands TCS deeper into physical AI infrastructure

The project advances Tata Consultancy Services beyond its traditional IT services and consulting business and deeper into the physical infrastructure required to train and operate large AI systems.

TCS incorporated HyperVault in India on October 29, 2025, specifically to establish AI and sovereign data centers. The company initially owned the subsidiary outright.

That ownership structure subsequently changed. TPG Terabyte acquired a 49% fully diluted stake in HyperVault on March 9, 2026, after TCS and TPG announced a strategic investment agreement the previous November. TCS had said the broader HyperVault platform would be financed through TCS and TPG equity alongside debt.

The Hyderabad announcement is substantially larger than the 180 billion rupees that TCS and TPG said in November 2025 they would commit to HyperVault over the following years. The latest announcement describes the potential 700 billion-rupee expenditure as investment by HyperVault “and its partners” to build and manage the Hyderabad infrastructure; TCS has not publicly detailed in Saturday’s release how that full amount would be divided among participants or financed.

That distinction is important: the 700 billion rupees is an expected investment ceiling for a phased project, not an assertion that the entire sum has already been spent or irrevocably committed.

A 1-GW campus would add substantial AI computing capacity

TCS previously said India had roughly 1.5 GW of data center capacity and cited industry estimates projecting national capacity above 10 GW by 2030. Against that benchmark, a single campus capable of reaching 1 GW would represent infrastructure on an unusually large scale, although its eventual size will depend on the phased buildout.

HyperVault says its infrastructure is intended to support sovereign AI and data-localization requirements as well as commercial AI workloads. The company has also described plans for more than 1 GW of new capacity across key Indian locations.

TCS said the Hyderabad campus will use green energy and “water-neutral design principles.” Those are company-stated design objectives rather than independently demonstrated environmental outcomes, and Saturday’s announcement did not provide detailed power-sourcing, water-consumption or construction metrics against which those claims could yet be assessed.

The development therefore combines two of the central constraints in the global AI infrastructure expansion: access to enormous amounts of computing capacity and the electricity, cooling, networking and physical construction required to operate it.

For TCS, Hyderabad also provides a tangible site for a strategy that was largely prospective when HyperVault was created less than a year ago. The company now has land for the 1-GW campus, but its announcement makes clear that the pace and ultimate extent of construction will depend on customer demand and changing technology requirements.