India's data center boom is scaling fast. It's benefits... may not
Scale is visible. Value is harder to hold.
In Navi Mumbai, the structures look out of place.
They rise behind controlled entry points and silent perimeters, surrounded by refineries and logistics yards. Little moves in or out. No crowds. No visible output. Yet these facilities sit at the center of one of India’s most aggressive infrastructure pushes.
They are data centres.
What appears to be inert real estate is, in fact, the physical layer of India’s digital ambitions. And increasingly, of global artificial intelligence.
Over the past few years, India has shifted from being a peripheral node in global data infrastructure to an active construction zone. Installed capacity has reached roughly 1.3 gigawatts, nearly triple its 2020 level, with significantly more planned.
The pace matters less than the composition of who is building.
Adani Group has committed up to $100 billion in data centre investment over the next decade. Microsoft is expanding its cloud and AI footprint with multibillion-dollar commitments. Alphabet is building out large-scale infrastructure to support AI workloads. NTT DATA remains one of the largest existing operators.
The surface narrative is straightforward. India is becoming a digital infrastructure hub.
The underlying system is less settled.
Data, policy, and cost
India’s rise in data centres is not accidental. It is engineered through a combination of regulatory pressure, fiscal incentives, and structural cost advantages.
The first force is data localization.
India produces an outsized share of the world’s digital exhaust, estimated at around one fifth of global data, yet historically hosted only a small fraction of the infrastructure needed to process it.
Policy is closing that gap.
The Reserve Bank of India already requires financial data to be stored domestically. Broader data protection rules are expected to extend this logic to personal and platform-level data. The effect is not just compliance. It is gravitational. Data that must stay local creates immediate demand for local infrastructure.
The second force is incentives.
India’s central government has extended tax holidays for foreign-owned data centres through 2047. At the state level, competition has intensified. Maharashtra offers discounted electricity. Karnataka links incentives to renewable energy use. Tamil Nadu and Telangana have built fast-track approval systems to ease land acquisition and permitting.
The result is a fragmented but coordinated push. Each state is effectively bidding for infrastructure that has national strategic value.
The third force is cost.
Electricity in India, while expensive relative to domestic incomes, is cheaper in dollar terms than in many developed markets. Land and construction costs are also lower. At a time when power constraints are tightening in the United States and parts of Europe, India offers something increasingly scarce.
Available capacity at scale.
For hyperscalers, this combination matters.
AI models require vast compute resources. Training and inference workloads are pushing existing infrastructure systems to their limits. India offers an alternative geography where expansion is still feasible.
This is why companies are not just entering. They are committing.
Infrastructure without friction
From a construction standpoint, the system is working.
Capital is flowing in. Land is being allocated. Power agreements are being secured. Facilities are being built at speed.
The logic is cumulative. Each new data centre increases the attractiveness of the ecosystem for the next. Cloud providers cluster. Network infrastructure follows. Service providers attach themselves to the growing base.
India begins to look less like an emerging node and more like a scaled platform.
There is also a geopolitical layer.
As tensions reshape supply chains and digital governance becomes more fragmented, countries are seeking greater control over their data infrastructure. India’s push aligns with this shift. It is not just about growth. It is about sovereignty.
Owning the physical layer of data is increasingly seen as a strategic necessity.
On paper, this creates alignment between state priorities and corporate expansion.
But alignment at the point of investment does not guarantee alignment in outcomes.
What data centres actually produce
Data centres are capital-intensive. They are also economically narrow.
Most of their employment impact occurs during construction. Once operational, they require relatively small workforces. The facilities are designed for stability, not activity. Their value lies in uptime and throughput, not human engagement.
This creates a structural tension.
India is building infrastructure at scale. But the direct economic spillovers are limited.
The comparison often invoked is the internet boom. Jensen Huang of Nvidia has suggested that data centres could have a similarly transformative effect on India’s economy.
The analogy is optimistic.
The internet created distributed value. It enabled new companies, services, and forms of participation. Data centres concentrate value. They are enablers, not endpoints.
The question is not whether they matter.
It is where the value they enable ultimately settles.
Where value moves
Much of the capital flowing into India’s data centre sector originates outside it. Hyperscalers design the architecture, control the software stack, and capture the majority of downstream value.
India hosts the infrastructure.
The economic capture happens elsewhere.
This is not unique to India. It is a feature of how digital systems are structured. The highest margins sit at the application and platform layers. Infrastructure, while essential, is often lower-margin and more commoditized over time.
There is also a second layer of leakage.
Energy.
Data centres are power-intensive. As capacity expands, they place increasing demands on local grids. In regions where energy supply is already uneven, this creates trade-offs. Power allocated to data centres is power not available elsewhere, unless supply expands in parallel.
Some states are attempting to manage this through renewable incentives. But the scale of projected growth suggests that energy will become a binding constraint.
Infrastructure that appears digital is, at its core, physical.
It consumes land. It consumes power. It locks in long-term resource allocation decisions.
Sovereignty vs control
India’s push for data localization is framed as a way to increase control over its digital ecosystem.
At a physical level, this is working. More data will reside within the country’s borders. More infrastructure will be domestically located.
At a systemic level, control is less clear.
The entities building and operating much of this infrastructure are global. The platforms that sit on top of it are global. The flows of value that pass through it are global.
Localization changes geography.
It does not automatically change ownership or influence.
This creates a subtle divergence.
India gains visibility and partial control over the movement of data. But the economic logic governing that data remains tied to global platforms.
The infrastructure becomes local.
The system it supports remains distributed.
The system under pressure
As investment accelerates, these tensions begin to surface.
States compete to attract data centres, offering incentives that reduce immediate costs but shift longer-term burdens onto public systems. Power grids face rising demand. Land use decisions become more constrained.
At the same time, hyperscalers optimize globally. Workloads can be shifted. Capacity can be reallocated. The infrastructure is fixed. The usage is fluid.
This asymmetry matters.
It means that while India anchors the physical layer, it does not fully control how that layer is utilized over time.
If costs rise or conditions change, capital can redirect. The infrastructure remains.
This is not a failure of policy. It is a reflection of how global digital systems operate.
They distribute risk unevenly.
India’s data centre boom is real. It is large. It is strategically significant.
It is also, in many ways, transitional.
The country is moving from being a source of data to a host of infrastructure. That is a meaningful shift. It changes how India sits within the global digital economy.
But hosting infrastructure is not the same as capturing value.
The facilities in Navi Mumbai will continue to expand. More will appear across the country. Capacity will rise. Investment announcements will accumulate.
From the outside, it will look like progress. And in many respects, it is.
Inside the system, something more ambiguous is taking shape.
India is securing the pipes through which its data flows.
What remains less certain is who ultimately controls what moves through them.