India’s TCS gets TPG to fund half of $2B AI data center project

Indian IT giant Tata Consultancy Services (TCS) has raised $1 billion from private equity firm TPG as part of a multi-year, $2 billion project to build a network of gigawatt-scale data centers in the country.

The project, called ‘HyperVault’, comes as demand for AI computing is rising faster than companies can build the energy-hungry infrastructure needed to support it.

The gap between supply and demand for AI computing in India is particularly large: the country generates almost 20% of the world’s data, but is responsible for only about 3% of global data center capacity. Big tech companies and cloud providers have been too invest billions of dollars to expand local capacity and leverage the growing adoption of AI products in the country.

With HyperVault, TCS and TPG plan to develop high-density liquid-cooled data centers with the power and networking capacity needed to support advanced AI workloads across large cloud regions, the companies said.

Liquid cooling and high-density rack designs are becoming more common as the GPUs needed to power AI inference and training consume significantly more power and generate more heat than conventional CPU servers. But such designs also raise questions about resource use in countries like India, where water scarcity is already a problem.

In urban hubs like Mumbai, Bengaluru and Chennai, where much of India’s data center capacity is concentrated, existing water scarcity can complicate operations. S&P Global, citing estimates from the Uptime Institute: noted that a data center load of 1 MW can require up to 25.5 million liters of water per year for cooling, putting pressure on already stressed infrastructure.

The rapid construction of AI data centers will further strain India’s power and land use, two other bottlenecks identified by industry analysts. High-density AI clusters require reliable electricity supplies and large plots of industrial land, two requirements that are increasingly difficult to achieve in large urban regions.

Nevertheless, global tech companies consider India as their frontier for building AI infrastructure. According to S&P Global, local and global technology companies have announced investments of more than $32 billion over the past two years to expand the country’s data center infrastructure.

In January, Microsoft said it would invest $3 billion over two years in India’s cloud and AI infrastructure, and in October Google said it would spend $15 billion over five years to build a gigawatt-scale AI data center hub in the southern state of Andhra Pradesh. And in 2023, Amazon committed $12.7 billion to build AWS cloud infrastructure in India through 2030.

TCS said it would partner with hyperscalers and AI companies to design, deploy and operate the AI ​​infrastructure as the platform expands. The company plans to build approximately 1.2 gigawatts of capacity in the initial phase.

More than 95% of new data center capacity in India over the next five years will come from leased facilities, with the rest driven by hyperscalers building dedicated AI infrastructure, S&P Global estimates. Local players such as Reliance Industries and CtrlS are also expanding their data center capacity to meet the rising demand.

TCS and TPG predict that India’s total data center capacity could exceed 10 gigawatts by 2030, up from around 1.5 gigawatts today.