India’s AI ambitions need a sustainable powerhouse - Firstpost

An AI data centre is a facility filled with powerful processors, storage systems and networking equipment. It trains and runs AI models, handles huge data volumes and supports digital services, often around the clock. That makes it a physical foundation of the AI economy.

The processors generate substantial heat while working. Cooling systems must remove that heat to keep equipment reliable. Conventional cooling can use considerable electricity and, in some designs, large amounts of water. Direct-to-chip liquid cooling and immersion cooling can reduce resource use where suitable.

India’s data-centre capacity rose from approximately 375 MW in 2020 to around 1,500 MW in 2025. As more facilities are built, electricity, water availability and local infrastructure will face greater pressure. Planning must therefore consider climate, ecological sensitivity, water resources and possible failures, not only land and connectivity.

India’s data-centre capacity expanded rapidly as AI, cloud services and digital activity grew. Government figures cited in the article show an increase from approximately 375 MW in 2020 to around 1,500 MW in 2025. That is roughly four times the earlier capacity in five years.

The country is also expanding shared computing access. The IndiaAI Mission is widening access to computing resources. At the India AI Impact Summit in February 2026, the government announced plans to add 20,000 GPUs to its existing pool of roughly 38,000. GPUs are important because they power demanding AI training and operating tasks.

These figures show strong infrastructure momentum, but capacity alone is not enough. India must turn investment and computing power into an AI advantage that is reliable, affordable and sustainable. That requires matching data-centre growth with electricity, transmission, water and cooling plans.

Artificial intelligence is moving beyond a future technology. It is changing how economies function across major sectors, including healthcare, agriculture, banking, education and manufacturing. For India, this creates a chance to improve economic strength while building capabilities in a strategically important technology.

The mechanism is broad. AI can support services, industrial activity and digital infrastructure, while the organisations and countries with strong AI systems may gain greater influence. The article links this opportunity to India’s ambition to become the world’s third-largest economy. A competitive domestic ecosystem is therefore treated as an economic and national priority.

The opportunity also brings a condition. India must build AI without placing unsustainable demands on energy, water and natural resources. International technology companies are committing billions to Indian infrastructure, and the IndiaAI Mission is expanding computing access. The challenge is converting that investment into durable advantage.

AI data centres need dependable electricity because powerful processors train and run models and support services around the clock. If facilities expand faster than generation and transmission, the power system may struggle to serve both data centres and other users. Reliability and affordability could suffer.

The article cites rising demand, reduced hydropower output and greater reliance on thermal generation. It also reports that India may need additional coal-fired capacity beyond previously planned additions. Data-centre growth can therefore increase pressure for short-term fossil-fuel solutions, especially when large facilities cluster geographically and require strong grid connections.

That path would create tension with climate commitments and long-term sustainability goals. India needs diversified supply, including renewables, nuclear power and storage, alongside stronger transmission and efficient electricity use. Developers can help by securing clean power, investing in storage and improving demand management as the grid expands.

A data centre can be commercially attractive while creating difficult local trade-offs. Its electricity demand may add pressure to the grid. Its cooling systems may draw on water. Its construction may require land and affect surrounding ecological conditions. These impacts can shape whether communities accept a project.

The article reports that residents resisted proposed investments because of concerns about electricity demand, water stress, land use and environmental effects. Communities therefore experience the immediate consequences, while developers and governments face questions about approvals, infrastructure capacity and long-term viability. Existing users may also compete for local resources.

Planning should assess more than land prices, connectivity and electricity access. It should consider water availability, temperatures, ecological sensitivity and existing infrastructure, while avoiding one-point failures. State governments should examine cumulative impacts and present and future allocations, rather than reviewing every project in isolation.

India’s sustainable alternatives involve both energy supply and cooling. A diversified power strategy can combine renewables, nuclear power, energy storage, stronger transmission networks and more efficient electricity use. This reduces dependence on expanding fossil-fuel generation for every new data-centre demand.

Data-centre developers can secure additional clean-power supplies, invest in storage and improve demand management. Cooling can also become more efficient. Direct-to-chip liquid cooling, immersion cooling and recycled water are identified as options where they are technically and economically appropriate. These approaches target the heat produced by high-performance chips while limiting electricity or freshwater use.

The article does not present one universal solution. Suitability depends on local resources, infrastructure and project conditions. Developers should disclose meaningful water and energy data. Governments should assess water availability, temperatures, ecological sensitivity and cumulative impacts before approving facilities, while planning transmission and avoiding one-point failures.

An electricity grid moves power from generating sources through transmission networks to users. Generation must match demand closely as electricity is consumed. Storage can hold electricity and release it when needed, while demand management can shift or reduce use. This balance keeps service dependable.

For AI data centres, the issue is especially important because powerful processors and digital services may operate around the clock. If generation is insufficient, or transmission cannot deliver power to concentrated facilities, reliability can weaken. If demand rises faster than infrastructure, pressure can spread across the wider system. This explanation draws on established electricity-system principles; the article focuses on their relevance to AI.

India’s article identifies rising demand, reduced hydropower output and increased thermal reliance as current pressures. It calls for renewables, nuclear power, storage, stronger transmission and efficient electricity use. Coordinated planning is essential so computing growth does not outpace the grid’s ability to serve it.