India’s artificial intelligence (AI) race is entering a new phase. Until recently, the conversation was mostly about software, skills and innovation. Now it is moving to something far more physical: land, electricity, cooling systems and water.
This shift is visible in the competition among states to attract large data centres. Google’s planned $15 billion AI hub in Visakhapatnam and the expansion of Maharashtra’s data centre ecosystem show how quickly the foundations of an AI-led digital economy are being built.
But there is a challenge. Data centres are physical infrastructure that need uninterrupted electricity, efficient cooling and secure access to water. In a hotter and more uncertain climate, those requirements will become harder to guarantee. India’s digital ambitions, therefore, depend on solving a very traditional infrastructure problem.
A data centre differs from a factory. It has no smokestack and produces no visible goods. But in many ways, it behaves like an industrial facility. It uses large and continuous supplies of electricity and depends on the surrounding infrastructure to function.
The rise of AI poses a significant challenge. AI systems need far more computing power than many traditional digital services. More powerful chips consume more electricity and produce more heat.
The International Energy Agency estimates that global electricity demand from data centres could more than double by 2030, with AI an important driver of this increase. This impacts the rivalry between countries and Indian states seeking to become AI hubs.
Cheap land, subsidies and faster approvals will matter but won’t be enough, while a real advantage may come from reliable clean power, resilient electricity grids and cooling systems designed for a warming climate.
Andhra Pradesh and Maharashtra illustrate this new reality. Andhra Pradesh’s ambition to build a large AI and data centre ecosystem around Visakhapatnam reflects the idea’s potential. Coastal connectivity, renewable energy potential and electricity-sector reforms that improve power access for large consumers are all important advantages.
But the same geography that creates opportunity also requires caution. Coastal infrastructure built today must operate for decades in a world of higher temperatures, increasing climate stress and changing weather patterns. Maharashtra faces a different challenge. Mumbai has become India’s dominant data centre hub thanks to its financial industry, connectivity and established technology ecosystem.
But concentration creates its own vulnerabilities. A large share of critical digital infrastructure clustered in a single region increases exposure to pressure on electricity networks, land availability and water resources, apart from climate risks.
The answer is not to slow down investment, but to recognize that data centres require a different policy approach. Both the central and state governments have largely treated data centres as economic development opportunities, offering land, tax benefits and easier approvals. The next phase must focus equally on resilience.
Policies should require climate-risk assessments before major projects are approved. Location decisions should consider future heat, water availability and extreme weather exposure, not just current commercial advantages. Standards for energy efficiency, water consumption and backup infrastructure should be part of planning approvals.
Renewable power procurement also needs a more practical approach. A clean electricity contract helps reduce emissions but does not by itself solve the issue of keeping servers running 24/7. Large AI facilities will require investment in transmission networks, energy storage and dependable clean power.
A recent global analysis of planned data centres by risk consultancy XDI shows why resilience matters. The report examined the climate risks planned data centres are facing globally and warned that the question is no longer only of where new digital infrastructure should be built but whether these assets can remain operational and economically resilient through their lifetime. It assessed risks from hazards including extreme heat, flooding, cyclones and infrastructure disruption.
One of its key conclusions is that a data centre cannot be viewed in isolation. Even a highly protected facility depends on electricity networks, telecommunications, transport systems and water supply. A building may survive an extreme weather event but still suffer disruption if the systems around it fail.
Companies building these centres have a larger responsibility. Climate resilience cannot be treated as an engineering adjustment made after construction.
Operators need to choose locations carefully, invest in efficient cooling technologies, disclose energy as well as water use and design facilities for future climate conditions. This will increasingly become a competitive advantage. Customers relying on AI and cloud services will care not only about speed and cost but also reliability. Investors and insurers are also likely to examine whether expensive digital infrastructure is prepared for climate disruption.
India does not face a choice between AI and climate responsibility. Much of its AI infrastructure is being built now, giving the country an opportunity to avoid mistakes made elsewhere. Size and processing speed aren’t everything. The next winners in digital infrastructure are likely to be players that are equipped to keep data centres up and running reliably in a warming world.
The author is an independent expert based in New Delhi, Kolkata and Odisha. Twitter: @scurve Instagram: @soumya.scurve.