Meta’s Georgia well-water dispute shows why India’s AI data-centre boom needs water safeguards

When construction began near their home in Georgia, US, Beverly and Jeff Morris say their well water deteriorated. Sediment clogged pipes, household appliances failed and water pressure fell. Their experience near Meta’s Stanton Springs data centre has become a warning in India, where AI and cloud-computing facilities are expanding in cities already facing competing demands for water and electricity. But there is a crucial distinction: the residents’ problems are documented through reporting and a lawsuit; whether Meta’s construction caused them is contested.

What happened near Meta’s Georgia data centre?

The Morrises live roughly 1,000 feet (about 305 metres) from Meta’s Stanton Springs campus in Newton County, Georgia. After construction started in 2018, the couple reported sediment in their private well, repeated appliance damage and failing taps. A July 2025 New York Times investigation recorded their account and wider concerns about water availability and rising local water charges. The Instagram reel circulating in India presents this as a data centre drying up the entire area. The evidence supports a serious household complaint, not that sweeping conclusion.

In late August 2026, four families near the campus filed a lawsuit against Meta and a regional development authority. They allege that tree clearing, blasting and excavation disturbed the underground aquifer, sending silt into their wells. Those are allegations before a court, not judicial findings of contamination or liability. Meta has disputed a causal link and pointed to a commissioned groundwater study; in April 2026 it said its operating facility buys water from the local water authority and does not use groundwater directly. That statement addresses operational water supply, but does not by itself settle allegations about the effects of construction on neighbouring private wells.

Why the distinction between construction and cooling matters

Two different water questions are often mixed together. Construction may affect soil, drainage or groundwater and needs site-specific investigation. After a facility begins operating, its cooling system can use water, while the electricity supplying it may have a separate water footprint. A data centre drawing municipal water does not automatically mean it has drained neighbouring private wells; equally, compliance with a water-supply agreement does not resolve every possible construction complaint.

For the Georgia families, a fair investigation would compare water-quality and well-performance records before and after construction, hydrogeological conditions, excavation and blasting records, independent testing and alternative explanations. Residents should not have to bear unreasonable costs or lose access to safe water while a credible complaint is investigated. Equally, responsibility must be assigned on evidence rather than proximity alone.

India faces a measurable water-planning challenge

India’s concern is not imaginary, even though the reel’s claim that the same Georgia disaster has already arrived in Mumbai has not been established. The Council on Energy, Environment and Water (CEEW), in a February 2026 assessment, estimated that Indian data centres use around 150 billion litres of water a year and projected that demand could more than double by 2030. It also reported that a typical 100-megawatt hyperscale facility can consume around 20 lakh litres per day for on-site cooling, depending on design, climate and technology. That is an illustrative facility, not a fixed amount used by every data centre.

A May 2026 WRI India assessment found more than half of India’s data centres located in water-stressed regions. The problem is especially important where several facilities are concentrated close to residential, agricultural and industrial users. The impact of any particular project depends on which water source it uses, the season, local demand, recycling and cooling technology. National totals cannot determine what happens to an individual neighbourhood or well.

Electricity also deserves examination. Data centres operate continuously and growing demand can require grid upgrades and new generation. Whether households pay more or suffer power shortages depends on project-specific investment, tariffs and supply planning; neither outcome follows automatically from the existence of an AI data centre.

What India’s policy currently provides—and where it can improve

In a March 2026 parliamentary reply, the Union government said the electronics ministry had received no concerns about data-centre water consumption and pointed to more efficient cooling technologies and existing groundwater-extraction regulation. An August 2026 government reply clarified that AI data centres do not automatically require a separate environmental clearance simply because they are data centres; projects meeting specified building or township thresholds may require prior clearance, during which water availability, reuse and safeguards are considered. These are important distinctions, but they do not establish that every project’s local water effects have been independently tested or publicly explained.

India can protect both digital infrastructure and basic water access by requiring credible local water-balance studies before approval, including dry-season conditions and cumulative demand from neighbouring facilities. Projects should publicly disclose expected and actual withdrawals, water sources, recycling and cooling methods; regulators should require baseline testing near vulnerable wells, independent monitoring, enforceable limits and a prompt remedy for proven damage. The same discipline should apply to grid capacity and who pays for necessary upgrades. Recycled water and lower-water cooling can reduce pressure where technically suitable, but their benefits must be verified rather than assumed.

The lesson from Georgia

The Georgia case does not prove that Meta drained an entire county or that Indian data centres are already poisoning drinking water. It does show the human cost when households report serious water problems near large industrial construction and causation becomes a long, unequal dispute. India’s independently documented growth in data-centre water demand and concentration in water-stressed regions makes preventive scrutiny necessary now. Approving projects without credible local water evidence, public safeguards and effective remedies would be unjust to communities asked to bear risks for infrastructure serving a much wider market. Expansion is defensible only when safe household water supplies are protected in practice, not merely promised.