India's race to build data centers is reshaping the country's power system, raising difficult questions about energy, water, and public resources. The country's installed data center capacity was just 2.2 gigawatts (GW) in 2025, but projections are climbing fast. Firms like Wood MacKenzie now expect India to cross 12 GW by 2030, while India's Ministry of Power estimates data center power demand could reach as high as 26.3 GW by 2031-32.

The scale of this expansion is staggering. Google alone is setting up three data centers near Visakhapatnam on India's east coast, with a combined capacity of 2.5 GW. Reliance is building a 1 GW facility nearby. RMZ-Colt Data Centres is adding 1.25 GW, and Sify is contributing 50 MW expandable to 500 MW. In total, the Visakhapatnam district alone could soon host as much as 6 GW of data center capacity, with similar buildouts happening in Mumbai, Chennai, Hyderabad, Bangalore, and other cities.

Why Are Global Tech Giants Partnering With Local Companies Instead of Building Alone?

The most striking trend in India's data center market is how foreign firms are structuring their expansion. Google has partnered with Adani, Meta with Reliance, and Colt with RMZ. Rather than building their own facilities, hyperscalers are tying up with relative newcomers who can handle the complex logistics of securing land, water, and power at massive scale.

This shift reflects a fundamental challenge facing the industry worldwide. As data center campuses grow to hundreds of megawatts or even gigawatt scale, finding enough land, water, and power becomes increasingly difficult. Local partners with political connections and infrastructure expertise can navigate these constraints far more effectively than foreign companies operating alone.

"Infrastructure-backed platforms are better positioned to develop large integrated campuses with stronger control over land acquisition, utility provisioning and long-term power sourcing. The complexity and cost of building captive facilities at meaningful scale, securing land, power, approvals and cooling simultaneously, is making leasing from specialised operators the more rational choice for most enterprises and even some hyperscalers," explained Anshuman Magazine, a commercial real estate analyst at CBRE.

Anshuman Magazine, Commercial Real Estate Analyst at CBRE

This partnership model is reshaping India's data center market in ways that extend far beyond real estate deals. It is creating a two-tier industry structure that could define India's role in the global AI infrastructure boom for years to come.

How Is India's Data Center Market Splitting Into Two Competing Tiers?

India's data center industry is cleaving into two distinct camps. On one side are powerful newcomers with deep pockets and government connections, partnering with hyperscalers to build gigawatt-scale facilities. On the other side are traditional data center operators, often smaller and more technically specialized, who are being pushed toward smaller deployments.

According to a head of a leading Bangalore-based data center firm, the market split looks like this: "Pureplays will have about 30% of the market. There will be about a hundred of us. And then, there will be about ten large companies. They will tie up with hyperscalers like Amazon and Google." This consolidation is already visible in how facilities are sized. While the first camp builds GW-scale facilities, traditional operators are sticking with smaller units, no more than 100 MW in size.

This division matters because it determines not just who profits from India's data center boom, but also how the country's power grid, water systems, and regulatory frameworks will need to adapt. The two tiers face fundamentally different technical and logistical challenges.

What Are the Key Differences Between Small and Gigawatt-Scale Data Centers?

• Power Sourcing: Smaller 100 MW facilities can draw power primarily from state electricity grids supported by long-term utility agreements, diesel backup systems, and renewable energy via power purchase agreements. Gigawatt-scale facilities require a completely different approach, needing to mix and match power from multiple generators and potentially revamping the grid itself.
• Grid Requirements: A 1 GW data center needs 1 gigawatt-hour of power running continuously all day long, which means the surrounding grid must have multiple gigawatts of capacity available. This is far beyond what traditional grid infrastructure in most Indian cities can currently support.
• Renewable Energy Challenges: If a gigawatt-scale data center were to run entirely on solar power, it would need to install 5 GW of solar capacity to offset solar's low efficiency and intermittency. This creates a massive renewable energy footprint that most regions cannot accommodate.

"India can easily host multiple 100 MW data centres. The grid can support these," said the head of the Bangalore-based data center firm. "1 GW is hard. You will need to revamp the grid. Also, you will not get 1 GW from a single generator."

Head of Leading Bangalore-Based Data Center Firm

These technical constraints are forcing India to make strategic choices about where gigawatt-scale facilities can be built and how the country's power infrastructure will need to evolve.

Is India's Power Grid Actually Ready for This Expansion?

India's installed power generation capacity stands at 520 GW, with plans to expand to 777 GW by 2030. On paper, this suggests the country has room to accommodate data center growth. However, the real question is not whether India has enough total capacity, but whether the grid can be reconfigured to deliver power reliably to specific regions where data centers are being built.

The Visakhapatnam district example illustrates the challenge. With 6 GW of data center capacity potentially coming to one region, local grid infrastructure would need significant upgrades. This is not simply a matter of having enough power generation somewhere in the country; it requires transmission infrastructure, distribution networks, and grid management systems capable of handling massive, continuous loads in specific locations.

Water availability presents a parallel challenge. Data centers require enormous quantities of water for cooling systems. However, as the sources note, the quantum of water used by data centers is dwarfed by what India uses on agriculture, suggesting that water scarcity, while real, may be manageable at the national level if distributed thoughtfully.

Why Is India Attracting This Data Center Investment Now?

Several factors are converging to make India an attractive destination for data center investment. The push for data localization, where companies must store data within India's borders, is one driver. Another is the rising use of artificial intelligence (AI) by both private companies and India's union and state governments, which require local computing infrastructure.

Geopolitical events have also played a role. The war in the Persian Gulf, which saw Iran bomb data centers in Bahrain and elsewhere, has prompted companies to diversify their infrastructure away from politically volatile regions. Financial incentives rolled out by India's central and state governments have further sweetened the deal.

The financial returns are compelling. "Data centres are projected to generate anywhere between 12%-23% as return on investment," according to an executive at a Delhi-based AI firm. This has attracted both international investors like KKR and Blackstone, as well as local firms like Adani and Hiranandani.

What Does This Mean for India's Role in the Global AI Economy?

India's data center boom reveals something uncomfortable about the country's position in the global AI infrastructure race. While India is building the physical facilities that will power AI services worldwide, the country remains largely a raw input supplier. Foreign firms are making the strategic decisions about where to build, what technology to deploy, and how to structure operations. Local partners are securing land, water, and power, but the technology and the profits flow elsewhere.

This dynamic could shift over time as Indian data center operators gain experience and scale. But for now, India's data center buildout is less about India becoming a technology leader and more about India becoming the infrastructure backbone for global AI companies seeking cheaper power and favorable regulatory environments.

The next few years will reveal whether India can manage this expansion without the power shortages, water conflicts, and local opposition that have plagued data center development in other countries. The answers will shape not just India's energy future, but also the global geography of AI infrastructure for decades to come.