A data center can look insignificant from a national spreadsheet and disruptive from the window of a nearby apartment. Both descriptions can be accurate at the same time. That contradiction is becoming harder to ignore as India expands its digital infrastructure and tries to understand what that expansion actually costs. The problem may not be that India lacks data about data centers. It may be that the country keeps measuring their impact at a scale that smooths out the physical consequences that matter most.

National electricity consumption, aggregate water demand and economy-wide emissions can provide useful context. They can also make concentrated infrastructure appear deceptively small. A facility that represents a fraction of national power demand can still place substantial pressure on a local electricity connection. A cooling system that consumes a manageable volume of water nationally can become a meaningful competing demand in a water-stressed district. That distinction changes the policy question. The issue is no longer simply how much electricity or water data centers consume. It is where that consumption occurs, when it occurs, how intensely it occurs and which surrounding systems absorb the pressure.

The national average can hide the physical reality

Data centers concentrate substantial electricity, cooling and infrastructure demand at individual sites, making their local resource requirements important to assess alongside national totals. Their computing loads may operate continuously, while their cooling systems can impose specific demands on electricity and water networks. Their backup systems, substations, cooling equipment and mechanical infrastructure can also create impacts that do not appear neatly inside a national consumption figure.That makes averages a blunt instrument.

India can report that data centers represent a relatively limited share of national electricity use while individual facilities can still add significant demand to local electricity infrastructure, making grid capacity and site conditions important considerations. The national number may appear relatively modest. The local engineering requirements, however, remain tied to the specific facility and the infrastructure around it. The local engineering problem does not disappear.

Water presents an even sharper version of the same problem. A cooling system does not consume water in an abstract national reservoir. It draws from a particular source, under particular hydrological conditions, alongside particular industrial, agricultural or residential users. The relevant question therefore becomes less about whether data centers consume a large percentage of India’s water and more about whether a particular facility is drawing heavily from a stressed local supply. That is a fundamentally different measurement exercise.

Location should become part of the measurement

India’s data-center expansion increasingly requires a more granular approach to site selection. The physical location of a facility determines far more than its address. It determines the available electricity infrastructure, water conditions, climate exposure, land constraints, surrounding population density and the ability of local systems to absorb additional demand. That means the impact assessment for a data center should not end with its expected megawatts. A more useful framework would examine how much water the facility can draw during peak operation, where that water comes from, what cooling system it uses and whether its demand coincides with periods of local scarcity.

Electricity assessments could examine not only annual consumption but also peak loads, connection capacity and the infrastructure required to support them. Noise deserves the same treatment. Mechanical cooling equipment, generators and other infrastructure can produce continuous or intermittent noise. A facility located in an industrial zone may present a different issue from one positioned close to residential neighborhoods, even if both facilities have identical computing capacity. The megawatt figure cannot capture that difference.

India needs measurements that follow the burden

This does not lead automatically to the conclusion that India needs fewer data centers. It suggests something more practical: India needs better ways to distinguish between facilities with very different physical profiles. Two data centers with identical IT loads can have different water requirements, different cooling architectures, different peak electricity demands and very different relationships with nearby communities. Treating them as equivalent because they consume the same number of megawatts would discard precisely the information policymakers and planners need.

Transparent reporting could help close that gap. Operators could disclose standardized information on water sources, cooling systems, peak electricity demand, backup generation, operating noise and major resource requirements. Regulators and local authorities could then assess facilities against the conditions of the places where they operate rather than relying exclusively on national averages. Such reporting would also make comparisons more useful. Efficiency would become something that can be evaluated against climate, water availability and grid conditions rather than presented as a universal virtue detached from location.

The real question is not how big the industry looks

India’s data-center debate can become trapped between two simplified positions: the industry is either too resource-intensive or too economically important to constrain. Neither captures the engineering reality. The more revealing question is whether the country knows where the physical burden of digital infrastructure actually falls. A national statistic can tell India how large the sector is. India therefore does not necessarily need a smaller data-center industry.

It needs a more precise accounting system for the infrastructure it is approving. The future debate should move away from asking how much electricity, water or carbon “AI” uses in the abstract. The harder and more useful question is where those resources are being consumed, under what conditions and with what consequences at peak operation. If India measures the impact only after averaging it across a country of more than a billion people, the most important part of the impact may already have disappeared from the calculation.