Everyone wants AI. But who wants a data center next door?

Artificial intelligence is sold as something weightless. You open an app, type a question, and receive an answer. But every answer depends on physical buildings, electricity, equipment and cooling. Those buildings have neighbors.

A Reuters investigation published October 9 describes resistance to an Amazon data center near Mumbai. Residents in Thane are concerned about the loss of green space, noise and pressure on local utilities. Amazon disputes the suggestion that it intends to draw on household water and power; it says the facility will rely on water-free cooling and its own electricity arrangements.

The land under the cloud

For a community living beside a large industrial facility, 'the cloud' is not a metaphor. It can mean earthworks, construction vehicles, a changed view, future noise and uncertainty about what happens to local services. Reuters spoke with residents who say they were not adequately consulted as land near their homes was cleared.

The company sees a different picture: investment in infrastructure that can support a growing digital economy. The project is tied to Amazon's much larger announced India AI infrastructure spending plans. Both perspectives deserve examination, but only residents experience the immediate neighborhood impact.

Water is a local question, even when computing is global

S&P Global has estimated that between 60% and 80% of operating data centers in India could face high water stress this decade. That is an exposure estimate across existing facilities, not proof that the specific Thane project will consume drinking water. Location, cooling technology, resource contracts and climate vary.

Amazon told Reuters that this facility will use water-free cooling, a distinction that matters. Yet one mitigation does not settle every concern. Residents can reasonably ask about electricity supply, construction, ecological changes and how commitments will be monitored over time.

The opportunity and the cost appear on different balance sheets

National governments calculate AI investment, jobs and future industrial capability. Technology firms calculate the cost of getting new computing capacity online. Households measure quality of life and the reliability of nearby services.

S&P Global's outlook suggests Indian data-center electricity demand could rise substantially by 2030. That projection does not mean all demand will burden public utilities equally, but it highlights the need to discuss grid planning and resource contracts in public. A data center can be economically worthwhile while requiring stronger consultation and accountability.

India's government now acknowledges the electricity question

On October 14, Reuters reported that India's junior IT minister Jitin Prasada said the government was preparing a policy for data centers. He linked that announcement to a concern familiar far beyond India: electricity demand from large computing facilities should not leave ordinary consumers paying more or struggling for supply.

That announcement is not a settled set of rules. The minister did not specify the policy's provisions or timetable, so it would be premature to claim that residents have won new protections, that projects will be relocated, or that national electricity allocations have changed. The announcement does show that local questions about power, costs and accountability have moved into a national policy discussion.

The overlooked question is consent

A productive debate does not have to pit digital growth against residents. It should ask what is disclosed before permits are granted, which independent measurements are published, and whether complaints can be resolved after a facility opens.

TAP's question is not whether India should abandon AI infrastructure. It is whether the people living beside its infrastructure get a meaningful voice in decisions that reshape their neighborhoods. The benefits of invisible technology should not be an excuse to make its physical costs invisible too.