The boom we see around us with the recent Global South expansion of AI data centres, although signifies humanity heading towards a future that is incomparable in technological growth, hides the ugly truth of overexploitation of resources. 2026 was the year the saw an aggravated expansion of data centres into the global south. If one were to look at the ASEAN, Malaysia has rapidly emerged as a major data-centre hub in Asia, with substantial growth in both operational and planned facilities. This expansion is increasing pressure on electricity, water and other infrastructure resources. South Korea plans to develop up to 18.4 GW of AI data centres, a project split across the nation to account for grid congestion and support regional economic development, by 2035. Expansion into countries that continue to rely heavily on coal powered energy, strains the exisiting infrastructure and locks in long term dependence, while the global north is working towards phasing out of coal power plants.

India is emerging as a popular destination for data centres in the Asia Pacific region, with heavy investment by International corporations. While it looks good for the influx of FDI would place India in the AI race with major players, which necessitates a conversation about the impact on the country's natural ecosystem.

1. The Carbon Intensive Life Cycle of a data centre:

   
   

If one were to consider the manufacturing process of semiconductors, which is responsible for around 50 megatons of CO2 emissions annually, in addition to 75% of mobile carbon footprint coming from chip manufacturing. It is not difficult to assume that the entire life cycle from manufacturing, usage to disposal would be highly carbon intensive. The advancement that we see around us, apart from the speculation of a bubble, comes at a cost.

The commentary of production would be left unfinished if the extraction of rare earth minerals are not discussed. Extraction of minerals significantly damage land, waster, health and livelihoods of indigenous people, which is reflective of the colonial pattern of land dispossession and wealth extraction. The economic and social injustice is an added cost to the creation of advanced human intelligence. The carbon life cycle of extraction methodologies and transportation would further add to the embodied carbon emission of chip manufacturing. One of the methodologies of extraction requires removing the top soil and creating a leaching pool where rare earth elements from the ore are seperated using chemicals, these chemicals can seep into ground water and cause erosion or air pollution. Another technique would be to directly pump chemicals into the ground and use PVC pipes and hoses, which produces a mix, which is pumped into leaching ponds for seperation, essentially creating the same problems, according to a report published by Columbia Climate School.

2. Health and after concern:

3. 
   

The massive demand for electricity by data centres is usually run in fossil fuelled power plants or backup generators. These data centers emit hazardous pollutants such as nitrogen oxides and fine particulate matter, which results in respiratory diseases and cardiovascular conditions leading to the possibility of 1300 deaths annually in 2030.

These pollutants are a health risk even at the extraction stage, as discussed earliers, where chemicals and extraction methodologies pose critical risk to human health and environmental sustainability in general.

Rare earth elements enter human body through multiple pathways, resulting in organ and system dysfunctions through changes in genetics, epigenetics and signaling pathways. Peole located in areas with REE contamination can absorb them through inhalation and skin, which is detected in human blood, urine and hair, resulting in risk for neurotoxicity and adverse effects on reproductive system.

3. Water and energy being the more overexploited resources:

   
   

The already parched Global South would be stripped of all its natural water sources and would have to make up for the surging electricity demands of these data centres at the cost of its people. Now, if we were to come to India and specifically the South of India, the heavy influx of FDI and big names thrown across media, presents a beautiful facade to an already overpopulated, under-resourced region. Open AI partners with Tata to build a data centre with 100MW capacity and Google has invested 15 billion dollars on building a full stack AI hub. This is coming at a time when the retreating ground water table is reaching alarming levels of water scarcity, wherein borewells are having to be installed deeper and rainwater replenishment for the year 2026 is below average at 11-12%.

Although, the transition towards more green data centres are growing, one must question the infrastructural capabilites of the region and country in general to produce clean electricity to meet demand. Isolated transition of green energy reliant data centres does not lead to transition of the larger infrastructural capability. The scale at which large corporations invest into developing AI capabilities, would necessitate a larger discussion regarding their commitment to the the country's sustainability agenda. The energy demand in India to support the AI data centre boom will grow from 1.5 GW in 2025 to 13.5 GW in 2032, although the push for green energy has been discussed, there was no discussion on the regulatory framework that would govern data centres and a possible systainability standard specific to the sector.

4. Chinese innovation to support data centre expansion:

   
   

It would be difficult to gain a complete picture of the Asia Pacific region without commenting on the Chinese progress to support data centre expansion. China is pushing the limits to build green data centres that operate underwater powered by wind. In addition to this, the availability of cheap electrucity to power hyperscale facilties places it in the lead of the AI race, with their commitment to provide cheap stable and low carbon electricity.

Amongst the few big players in the race of AI, the transition to meet the unquenchable needs of the data centres are reliant on rapid green energy innovation and low carbon extraction methodologies.