Artificial Intelligence is the most discussed buzzword in the world today. AI is increasingly influencing all our lives. To keep any AI running, you need data centres. So, along with the AI race, there is a race for building new data centres around the world, and India is also no exception. India has claimed that it envisages an increase in data centre capacity. There has been an increase in investment in this sector with the promise that India will be transformed with AI. By 2030, the government of India proposes to build about 250 data centres across the country. The Union Budget 2026-27 also provides a tax holiday for cloud service providers that operate through India-based data centre infrastructure. This can be justified using the logic that support to build data centres can induce foreign investment, which can boost our economy. On the other end of the world, New York has imposed the first statewide moratorium on new data centres, by executive order in July 2026, covering facilities of 50 MW or more. Texas has also paused pending grid interconnection requests for a project-by-project audit, after ERCOT found itself holding more than 474 GW of proposed new load, roughly 90% in data centres. Even in the tech hub of the world in California, people have protested against the construction of data centres. Many different states in the US have increased the barriers to set up data centres amidst concerns that these can be harmful for the environment. Amsterdam imposed a municipal moratorium on hyperscale facilities, and in Ireland the grid operator has refused new connections in the Dublin area until 2028, after data centres came to account for around a fifth of national electricity consumption.Also read: Full Text | India is Welcoming Huge Data Centres With Incentives. But What Do the People Get Out of It?We need to be more sceptical about the impact of data centres, which are required to fulfil the magic that AI conjures up on our screens every day. These massive centres work through the day and we require industrial scale cooling for them. There is a constant requirement of clean water in these places. Apart from clean water, the amount of electricity required by these centres is also massive. According to Council on Energy, Environment and Water (CEEW)’s estimate, Indian data centres used roughly 150 billion litres of water in 2025 and around half a per cent of national electricity consumption, with both set to double by 2030. The electricity requirement will cause significant strain on states like Andhra Pradesh, Maharashtra, Tamil Nadu and Telangana, which lead the race of data centre. Nationally, these are small shares. The problem is concentration and who pays. The load arrives in a handful of districts, and when a state reclassifies data centres from commercial to industrial tariffs, as Maharashtra has, it does not make their power cheaper to generate. It decides that other consumers will carry the difference. The constant requirement of water also begs the question if India can really afford to spare clean water in a situation where we have severe ground water crisis. In Andhra Pradesh, Google is building India’s largest hyperscale data centre, which is proposed to span across 480 acres and cover three sites in two districts, about 200 acres at Tarluvada, 120 at Adavivaram and Mudasarlova, and 160 at Rambilli. The centre will also be located in an area which is close to the Kambalakonda wildlife sanctuary and the catchment of the Mudasarlova reservoir. Google has received significant incentives and tax holidays from the Andhra government for 15 years. The package amount is about Rs 22,000 crore – 480 acres at a quarter off market price, stamp duty waived, electricity duty exempted, a discounted industrial power tariff for fifteen years, water at a 25% discount for ten, and full state GST reimbursement during construction capped at Rs 2,245 crore, against a stated investment of Rs 87,520 crore. Understandably, this will cause huge stress on electricity availability and will pollute natural resources.Capitalism has historically caused significant deterioration of natural resourcesPower relations across countries can influence the question of pollution in countries. While Google sets up data centres in the global south, which will significantly pollute the country and strain resource availability, we can analyse pollution through the lens of international trade. Peters and Hertwich’s work on emissions embodied in trade shows that a country’s recorded emissions do not necessarily correspond to the emissions generated by the consumption of its population. A substantial share of the pollution generated in developing countries is associated with the production of goods consumed elsewhere. This creates a space for ecologically unequal exchange, where global value chains allow wealthy economies to specialise increasingly in relatively clean, high-value activities while importing manufactured goods whose production involves substantial environmental costs. There is evidence that the relocation of production from the Global North to countries such as China and India contributed to an increase in global emissions because production was initially more carbon-intensive in these locations. The apparent greening of some rich economies, therefore, needs to be interpreted carefully: part of the improvement in domestic environmental indicators can coexist with the externalisation of environmental costs through international trade. We can use the pollution haven hypothesis to appreciate the economic impact of this process. Firms which produce pollution-intensive goods may have incentives to locate production in countries with weak environmental regulations and their enforcement. This leads to developing countries acquiring a comparative advantage in “dirty” industries since the overall cost including environmental and social costs are lower.Also read: India’s Data-Centre Expansion Needs a Water-Justice FrameworkThis dynamic also has a distinctly neo-colonial dimension. Colonialism historically organised production in the periphery around the extraction of resources and the supply of commodities to the industrialising core. Global value chains can replicate such a hierarchy through formal market transactions and trade. While capital, technology and high-value activities are concentrated in the global north, labour intensive and “dirty” stages of the production happen in the global south. The specialisation of global south in pollution intensive goods is a consequence of unequal trading patterns, and its requirement for foreign exchange.On employment, the constraint is particularly sharp, because a hyperscale campus of this size employs a few hundred people once it is running. The investment is being courted for what it signals more than for what it staffs. The raw materials of India, like cotton and indigo, were used to fuel the Industrial Revolution in England. Data centres trained on Indian data and running on Indian electricity and water are now supporting the AI boom in the global north.Capitalism has historically caused significant deterioration of natural resources. As capital became increasingly more mobile, the global north could outsource this depletion to the global south. The new age AI race has once again shown this. The political economy of “pollute here, regulate there” implies that consumption and economic rents are disproportionately captured in wealthy economies, while production-related pollution and regulatory burdens are geographically concentrated elsewhere.None of this is an argument against building a data centre. Latency, data localisation and India’s own AI programme all need computing on Indian soil. It is an argument about terms and accountability. The following three points can change the balance: i) make disclosure of water and power use a condition of claiming the incentives rather than a request afterwards; ii) charge full cost for water in basins that cannot meet the demand they already have; and iii) put a review clause in a concession that currently runs to 2047 without one. New York and Texas have paused to ask these questions. India has not asked them yet.Satyaki Dasgupta is an Assistant Professor at the Department of Economics, Christ University, Bangalore. Dr Tuhinsubhra Giri is an International Research Fellow at the Academy of International Affairs NRW, Germany and an assistant professor of economics at Christ University, Bangalore.