TCS and TPG Partner on $2B AI Data Center in India

TCS and TPG Invest $2 Billion in Indian Data Center Network

Tata Consultancy Services (TCS), a leading Indian IT company, has received $1 billion in funding from private equity firm TPG. This investment is part of a larger $2 billion initiative focused on establishing a network of large-scale data centers throughout India.

Addressing the Growing Demand for AI Compute

This project, named “HyperVault,” is a direct response to the rapidly increasing demand for computational power needed to support artificial intelligence. The infrastructure required to handle this demand is not being built quickly enough to keep pace.

India's Data Center Capacity Gap

A significant disparity exists between data generation and data center capacity in India. The nation produces nearly 20% of the world’s data, yet possesses only approximately 3% of global data center infrastructure. Consequently, major technology companies and cloud service providers are making substantial investments to expand their presence and capitalize on India’s growing adoption of AI technologies.

HyperVault: Advanced Data Center Design

TCS and TPG intend to develop data centers utilizing liquid cooling and high-density rack configurations with HyperVault. These designs are essential for supporting the demanding power and network requirements of advanced AI applications across key cloud regions.

The Rise of Liquid Cooling and its Challenges

Liquid cooling and high-density racks are becoming increasingly prevalent as GPUs, crucial for AI training and inference, consume significantly more power and generate more heat than traditional CPU servers. However, these technologies also raise concerns regarding resource utilization, particularly water consumption, in regions like India where water scarcity is a pressing issue.

Water Stress and Data Center Operations

Existing water stress in urban centers such as Mumbai, Bengaluru, and Chennai, where a large portion of India’s data center capacity is located, could potentially complicate operations. Estimates suggest a 1 MW data center can require up to 25.5 million liters of water annually for cooling, further straining already limited resources.

Power and Land Use Constraints

The rapid expansion of AI data centers is also expected to exacerbate existing pressures on India’s power supply and land availability. High-density AI clusters necessitate a reliable electricity supply and substantial industrial land areas, both of which are becoming increasingly difficult to secure in major metropolitan areas.

India as a Frontier for AI Infrastructure

Despite these challenges, global technology companies are prioritizing India as a key location for building AI infrastructure. Over the past two years, both domestic and international firms have announced investments exceeding $32 billion to expand data center infrastructure within the country.

Recent Major Investments

Recent examples include Microsoft’s commitment of $3 billion over two years for cloud and AI infrastructure, Google’s planned $15 billion investment over five years for a gigawatt-scale AI data center hub in Andhra Pradesh, and Amazon’s $12.7 billion commitment through 2030 for AWS cloud infrastructure.

TCS's Role and Initial Capacity

TCS will collaborate with hyperscalers and AI companies to design, deploy, and manage the AI infrastructure as the platform grows. The initial phase of the project aims to establish approximately 1.2 gigawatts of capacity.

Future Data Center Capacity Projections

S&P Global forecasts that over 95% of India’s new data center capacity over the next five years will be leased, with the remaining portion driven by hyperscalers constructing dedicated AI infrastructure. Local companies, such as Reliance Industries and CtrlS, are also increasing their data center capacity to meet the growing demand.

Long-Term Growth Forecast

TCS and TPG anticipate that India’s total data center capacity could surpass 10 gigawatts by 2030, a significant increase from the current level of roughly 1.5 gigawatts.