By Narendra Kumar Mishra, Public Policy Professional
Data centres can strengthen digital sovereignty and attract investment, but their electricity and infrastructure costs must not be transferred to ordinary consumers.
Google’s proposed $15-billion AI and data-centre hub in and around Visakhapatnam is a striking sign of India’s growing importance in the global digital economy. Planned as a one-gigawatt-scale ecosystem across three campuses, the project could bring advanced computing infrastructure, international connectivity and new technical opportunities to Andhra Pradesh and India.
But its scale also highlights a question that receives far less attention than the investment figure: how will India reliably supply electricity to a new generation of power-intensive data centres without placing additional costs on households, smaller businesses and an already stretched grid?
Data centres are the physical backbone of the digital economy. They store and process information, operate cloud services, and provide the computing power required to train and run artificial intelligence models.
Despite the popular language of the “cloud”, these are large industrial facilities dependent on round-the-clock electricity, cooling systems, water, land, advanced chips and high-capacity fibre networks. As AI expands, India must therefore plan not only how many data centres it wants to attract, but also how their energy and infrastructure requirements will be met, who will bear those costs, and what lasting capabilities they will create within the country.
The urgent question, therefore, is not whether India should welcome data centres. It should. The real question is how India can build an AI economy without becoming merely a provider of subsidised land and power for technologies designed and owned elsewhere.
A recent argument on India’s AI future raises an important question: how can India become a producer of AI rather than merely a user or host of global AI infrastructure? Hosting global infrastructure can attract investment and build capability, but the real strategic opportunity lies in developing the next layer: domestic research, energy-efficient technologies, skilled talent and competitive Indian enterprises.
India’s data-centre policies have so far been highly effective at attracting investment. Maharashtra and Tamil Nadu, which together account for much of the country’s installed data-centre capacity, offer single-window clearances, electricity-duty waivers, land support and concessions for renewable-energy procurement. Mumbai and Chennai have consequently emerged as important digital infrastructure clusters.
However, policies designed to attract data centres have not paid equal attention to the power systems needed to sustain them. Data centres need uninterrupted, high-quality electricity. Their demand does not fall when the sun sets or the wind slows. Renewable energy is essential to their future, but installed solar and wind capacity cannot automatically be treated as round-the-clock supply. Batteries currently used at many facilities can manage short interruptions or balance demand for a few hours, but they cannot always support a large data centre through extended shortages.
Suggestions
The first priority should therefore be a national framework for data-centre resource planning. India’s draft national policy has remained predominantly investment oriented. A more comprehensive framework should establish baseline standards for electricity, water, land, emissions and energy efficiency. States may continue to compete for investment, but they should do so within common national guardrails.
Before approving a major facility, authorities should conduct a regional resource assessment. Can the local grid support the proposed load? What transmission upgrades will be required? Is the area already experiencing water or electricity stress? Who will pay for additional substations, transmission lines and backup capacity? Approvals should be linked to credible answers, rather than investment commitments alone.
Second, electricity tariffs must follow the principle of cost causation. Data centres require exceptional reliability, redundant connections and substantial grid capacity. The additional cost should not be quietly transferred to households, small businesses or rural consumers. Data-centre operators should pay proportionately for the infrastructure and power quality they require. Where public incentives are offered, they should be transparent, time-bound and linked to measurable outcomes.
Third, India must treat transmission as seriously as generation. Renewable-energy projects are often located far from the urban centres where data centres are concentrated. New solar panels alone will not solve the problem if clean electricity cannot reach Mumbai, Chennai or other digital clusters. Investment in interstate transmission, high-capacity substations, storage and grid-management systems must precede, or at least accompany, large data-centre projects.
Fourth, future capacity should be distributed more carefully. Continued clustering around Mumbai and Chennai may provide connectivity and commercial advantages, but it also concentrates pressure on land, water and electricity. Carefully selected sites outside congested metropolitan regions can spread investment and reduce local stress. This transition must be planned because less-populated regions may lack fibre networks, skilled workers and reliable power infrastructure. Location-based incentives should therefore reward projects that help create new, fully serviced digital corridors rather than merely adding load to existing hubs.
Fifth, policy should reward efficiency, not simply megawatts and capital expenditure. Companies seeking incentives should disclose their power-usage effectiveness, water consumption, carbon intensity and backup-fuel use. Benefits could be linked to improvements in server efficiency, advanced cooling, recycled water, closed-loop systems, waste-heat utilisation and reductions in diesel dependence.
India should also invest in research on efficient chips, cooling technologies, long-duration storage and power-flexible computing. Not every AI task must run at the same time. Some workloads can potentially be shifted to periods when renewable energy is abundant or overall grid demand is lower. Universities, start-ups, utilities and data-centre operators should be encouraged to test such models.
Finally, major projects must create capabilities beyond construction and facility operations. Incentives should support Indian research laboratories, local suppliers, workforce training and affordable computing access for start-ups and universities. A data centre becomes strategically valuable when it strengthens the surrounding innovation ecosystem, not merely when it occupies land and consumes electricity.
India does not have to choose between AI leadership and energy security. But it must recognise that digital policy is now also power policy, water policy and industrial policy. The objective should not be to slow the data-centre boom, but to govern it intelligently.
India’s success will not be measured by the size of its data centres, but by the strength of the innovation ecosystem built around them. The goal must be clear: not simply to power the world’s AI infrastructure, but to build the capability to shape, own and lead the AI future.
Disclaimer: This article is for educational purposes only and is based on publicly available material. It does not constitute legal, forensic, or cybersecurity advice. Tools and case examples mentioned should be used ethically and within legal boundaries. Misuse may result in legal consequences. Readers are advised to consult qualified professionals or legal authorities before undertaking any forensic investigation.
