- What Exactly Is a Data Center?
- Why Could India’s Data Center Capacity Grow 5x?
- AI Is Changing the Scale of Projects
- Why Is India Attracting Data Center Investment?
- Where Does the $45 Billion Investment Actually Go?
- Power Could Be the Biggest Part of the Story
- Which Listed Companies Are Connected to the Theme?
- But 10 GW Does Not Automatically Mean 10 GW of Profitable Capacity
- Why This Data Center Boom Is Different
India’s data center industry is entering a much bigger phase of expansion.
For years, growth was driven by internet usage, digital payments, streaming, e-commerce and cloud adoption. Those drivers remain, but artificial intelligence is now adding a far more power-intensive layer of demand.
Jefferies estimates India’s colocation data center capacity could rise from about 2.1 GW in 2026 to nearly 9.6 GW by 2031, roughly a fivefold increase in five years.
The investment required is equally large. Jefferies estimates around $45 billion of data center facility capex between 2027 and 2031 and another $90 billion for servers, taking the broader investment requirement to about $135 billion.
The key question, therefore, is not just how many data centers India will build. It is why demand is growing so rapidly and where all this investment will actually be spent.
What Exactly Is a Data Center?
A data center is the physical infrastructure behind the digital economy.
Whenever a bank processes transactions, an e-commerce platform stores customer information, a streaming service delivers video or an AI model processes a prompt, computing has to happen somewhere.
Servers perform that computing, but they need secure buildings, uninterrupted electricity, cooling, network connectivity and backup systems.
One important distinction for investors is between the data center operator and the cloud provider.
A cloud company or hyperscaler may own the servers, while a colocation operator provides the building, electricity, cooling and connectivity needed to run them. Server investment is therefore often undertaken separately from the data center facility itself.
This is also why data center capacity is measured in megawatts rather than square feet. Increasingly, what matters is how much computing load the electrical infrastructure can support.
Why Could India’s Data Center Capacity Grow 5x?
India’s colocation capacity is estimated at around 2,076 MW in 2026, compared with just 404 MW in 2019. Jefferies estimates this could rise to approximately 9,576 MW by 2031.
Several forces are driving this expansion.
The first is India's growing digital economy. More people are using online banking, UPI, e-commerce, OTT platforms and cloud-based applications. Businesses are also increasingly shifting workloads from in-house IT infrastructure to cloud and colocation facilities.
The second driver is cloud computing.
Jefferies estimates hyperscalers already account for around 60% of Indian colocation data center demand, while BFSI contributes another roughly 15%. Existing facilities are estimated to be running at 95-97% utilisation, leaving limited room to absorb future demand without additional capacity.
Then comes AI.
Training and running large AI models requires far more computing power than traditional enterprise workloads. High-density GPU servers consume significantly more electricity and generate far more heat.
That means new AI-ready facilities require stronger electrical systems and more advanced cooling than conventional data centers.
AI Is Changing the Scale of Projects
Recent announcements show how quickly project sizes are increasing.
TCS has launched its HyperVault AI infrastructure business and outlined a Hyderabad AI data center campus that can eventually scale to up to 1 GW.
Bharti Airtel’s data center subsidiary Nxtra had around 300 MW of capacity and has outlined plans to scale towards 1 GW over the coming years.
Adani Enterprise’s AdaniConneX is also developing large-scale data center infrastructure across India.
The important shift is not simply that more data centers are being built. India is moving from facilities measured in tens of megawatts towards campuses capable of eventually supporting hundreds of megawatts or even gigawatts.
That changes the investment opportunity across the ecosystem.
Why Is India Attracting Data Center Investment?
India needs more than rising digital demand to become a major global data center destination. It also needs to be economically competitive.
Jefferies estimates that average power and construction costs in India can be 25-40% lower than in several competing international markets.
India also has a huge domestic user base. Keeping computing infrastructure closer to customers reduces latency and improves performance.
Connectivity matters too. Mumbai and Chennai have become important data center hubs partly because of their access to submarine cable landing stations connecting India with international networks.
Data localisation requirements provide another structural driver, while government incentives are also encouraging global cloud and computing infrastructure to be located in India.
Together, lower costs, domestic demand, connectivity and policy support make India an increasingly attractive location for data center investment.
Where Does the $45 Billion Investment Actually Go?
This is one of the most important parts of the theme for investors. When people hear that India could see around $45 billion of data center facility investment, they may assume most of the opportunity belongs to data center operators.
It does not. A data center is an infrastructure ecosystem.
| Data center infrastructure | Share of facility capex | Estimated capex |
| Power and electrical systems | 35% | ~$16 billion |
| Land and building shell | 25% | ~$11 billion |
| Racks and fit-outs | 20% | ~$9 billion |
| Cooling | 15% | ~$7 billion |
| Network infrastructure | 5% | ~$2 billion |
| Total facility capex | 100% | ~$45 billion |
And this still excludes servers. Jefferies estimates another $90 billion of server capex, taking combined data center and IT hardware investment towards $135 billion over 2027-31.
That means the opportunity extends well beyond companies renting server space. It also involves transformers, switchgear, cables, cooling systems, backup generators, real estate, construction and network infrastructure.
Power Could Be the Biggest Part of the Story
Data centers are enormous electricity consumers.
They also need highly reliable power because customers cannot afford frequent outages. Operators therefore require grid connections, substations, transformers, switchgear, backup generation and sophisticated power management systems.
Jefferies estimates power and electrical infrastructure could account for around 35% of total facility capex, or roughly $16 billion through 2031.
This is why the data center theme reaches beyond technology companies.
Power transmission equipment manufacturers, cable companies, generator suppliers and cooling companies can all participate in the build-out.
AI makes this even more important because high-density GPU servers require more electricity and produce more heat. As server density rises, advanced cooling systems such as liquid cooling are becoming increasingly relevant.
Which Listed Companies Are Connected to the Theme?
There are several ways listed companies can participate. The first category is data center operators.
Bharti Airtel has direct exposure through Nxtra, while TCS is developing AI infrastructure through HyperVault. Adani Enterprises has exposure through AdaniConneX, while companies such as Anant Raj also participate through data center and real estate infrastructure.
Then there are the indirect beneficiaries.
Jefferies identifies companies across power transmission, electrical equipment, backup generation, cables and cooling that can benefit from rising data center construction. These include Hitachi Energy India, Siemens Energy India, GE Vernova T&D India, ABB India, Cummins India, Kirloskar Oil Engines, Polycab and KEI Industries, among others.
But investors should not treat all these companies as identical data center plays.
A transformer manufacturer has very different economics from a data center operator. Exposure to the sector, margins, competitive intensity and capital requirements can differ significantly.
But 10 GW Does Not Automatically Mean 10 GW of Profitable Capacity
This is where investors need some caution. Most 2030 and 2031 numbers refer to projected capacity or announced plans, not operating facilities today.
A 1 GW data center announcement does not mean the entire 1 GW will be commissioned immediately. Large campuses are usually built in phases depending on customer demand.
The industry is also highly capital intensive. Operators must secure land, power, connectivity and customers before capacity becomes economically productive.
Power availability could become a constraint, while AI workloads are also increasing cooling requirements.
Competition may rise too if several operators build capacity simultaneously.
For investors, a large industry capex number shows where money may be spent. It does not automatically tell us what revenue or profits an individual company will earn from that spending.
Why This Data Center Boom Is Different
India has been building data centers for years. What is changing now is the scale and type of demand.
The earlier phase was driven largely by smartphones, online banking, UPI, e-commerce, streaming and cloud adoption.
Those drivers remain, but AI computing is now adding a much more infrastructure-intensive layer.
India is increasingly trying to become not just a consumer of cloud and AI services but also a location where a meaningful share of the underlying computing infrastructure is physically housed.
Jefferies estimates India's colocation data center rental market could increase from roughly $2 billion in 2026 to more than $9 billion by 2031.
But the opportunity surrounding those facilities could be considerably larger because every additional megawatt requires electrical equipment, cooling, land, construction, cables, networks and servers.
That is what makes the projected fivefold increase in capacity significant.
The real story is not simply that India may have more buildings filled with servers by 2031. It is that cloud computing and AI could create a new infrastructure investment cycle across power, engineering, construction, real estate, cooling and network infrastructure.