Data science as a tool of neo-imperialism often operates through algorithmic bias that ignores marginalized voices. Furthermore, data justice does not prioritize environmental protection, and corporate-dominated governmental monitoring bypasses toxic blind spots. Data science is not a standalone panacea for environmental racism. Despite technological advancements, statistical models tend to reinforce the very inequalities they aim to solve, encouraging slow violence. A significant milestone of the Environmental Data Justice Movement is its focus on data ownership. To achieve environmental justice, the focus of data science must shift from extractive methods—where data is taken from communities without their input—to collaborative modeling, a shift that is not currently taking place.
In a recent letter dated 6 June 2026, addressed to the Minister of the Union Ministry of Environment, Forest and Climate Change (MEFCC), Government of India, Mr. E.A.S. Sama, former secretary to the Government of India, highlighted that AI and data centers cause irreparable environmental havoc, displace marginal farmers (mostly Dalits), and result in widespread human rights violations. Under Article 48A of the Constitution of India, the MEFCC cannot afford to abandon its constitutional obligation of “protecting and improving” the environment. Mr. Sama further calculated that AI and data centers consume enormous quantities of water and electricity. The cumulative global capacity of data centers will reach 200 GW over the next ten years, of which 50% will be located in the USA, 30% in the Asia-Pacific region, and the remainder in other regions, including India.
A comprehensive study by a UN agency states that one of the most consequential yet comparatively unexamined dimensions of AI is its environmental footprint and the justice implications that follow from where and how AI infrastructure expands. AI is not “just code”; it also involves physical infrastructure and supply chains, including data centers, chips, electricity generation, cooling systems, water withdrawals, land occupation, critical minerals, and eventual e-waste.
In the Indian context, the environmental footprint of a 1 GW data center would consume 11.4 TWh per year of electricity and 8.1 billion liters of water per year, generate 8.1 billion tonnes of carbon, and occupy more than 600 acres of land. Thus, if 17 GW of data center capacity comes up in India by 2030, these centers would consume 194 TWh of electricity—four times the projected electricity demand of Delhi by 2030. Additionally, these centers would require 138 billion liters of water, equivalent to 15–18 times the projected water demand of Delhi by 2030. The carbon footprint of 17 GW of data center capacity would be comparable to that of Delhi by the same year. Regarding arable land requirements, 17 GW of data center capacity would displace more than 30,000 marginal farmers, mostly from Scheduled Castes, disrupting their lives and livelihoods. It may be noted that in India, per capita arable land use is only 0.3 acres.
Mr. Sarma has also cited the example of two sites in Visakhapatnam that were handed over to a foreign AI and data center conglomerate. These sites are located on forest lands, with one falling within an eco-sensitive zone notified by the MEFCC. Here, environmental injustice would encourage the displacement of forest dwellers, tribals, endangered species, and biodiversity. To borrow a Gandhian concept—”water famine,” which he coined more than a hundred years ago—such a crisis could be a natural consequence of establishing data centers in eco-sensitive zones.
Angala Huyne Zhang, in an article titled “Democracy vs. Data Centres,” points out that a Gallup survey shows seven out of ten Americans oppose building AI data centres in their local communities. The race for AI supremacy between the United States and China has pushed for massive construction of data centres in the U.S. According to one estimate, there are over 4,000 data centres in the U.S., located in clusters such as Ashburn, Virginia; Silicon Valley; Dallas; Sulphur Springs, Texas; and Chicago. These centres impose severe environmental and public health burdens on surrounding communities.These centres mostly depend on fossil fuel power plants and massive diesel backup generation. They emit 200 to 600 times more nitrogen oxides than a natural gas plant. The worst affected are rural, low-income, Black, and Latino communities, exacerbating high rates of asthma, respiratory illness, and air pollution.
AI servers require millions of gallons of water per day for cooling. In the drought-stricken western and southwestern parts of the U.S., water scarcity for residential and agricultural use is acute. The staggering electrical draw of AI clusters crowds out the grid’s baseload, driving up utility bills for communities and disproportionately burdening low-income households. These centres generate 24/7 low-frequency noise pollution and require expansive land footprints for buildings and high-voltage transmission lines.
In the U.S. and Europe, there has been stiff resistance from working-class people, peasants, and local marginalized communities. The pressure on electricity supply, water scarcity, and thermal stress is too heavy to bear. There is also an imminent threat to data sovereignty from data centres run by U.S.-based companies like Google. Finding it difficult to face this resistance, IT companies are shifting their projects to developing countries like India, where data protection laws are fragmented, environmental regulation is fragile, and crony capitalism is rampant.
The outcome of these centres would lead to the elegant evasion of the present misery of the working class. Their history would be altered by algorithm. Selective memory would engineer confusion. The causes and consequences of working-class movements would lose their stable meanings. Democracy would be substituted by electionocracy. Governance would acquire immunity from accountability, as the right-wing governing class acts as the henchmen of power-guzzling behemoths. It is high time that people realize the environmental implications of data centres and the threat they pose. AI adoption is not the sole measure of a country’s strength. Crony capitalists will accelerate job displacement, widen inequality, and cause limitless harm to the environment. Who will use these facilities? And at what environmental cost?
