Rethinking India’s Sustainable AI Policy

India stands at a pivotal moment in its digital journey. Artificial intelligence is expanding rapidly, yet the environmental consequences of this growth are only beginning to surface in public debate. The decisions India makes now will shape not just its technological future, but also the resilience of its energy systems, water security, and ecological stability for decades.

Ram Ramprasad

January 6, 2026

954

Rethinking India's Sustainable AI Policy new — Image credit - Sify.com

Introduction

Artificial intelligence is fast becoming a foundation of modern economies. From healthcare and education to governance and climate modelling, AI holds enormous promise for India’s development. But the global push toward ever-larger AI systems is creating serious environmental pressures that are starting to undermine many of AI’s promised benefits.

At the heart of this challenge lies the soaring consumption of energy, water, and land by large language models (LLMs) and the hyperscale data centers that support them. For India — a country already under intense climate stress and resource constraints — this is no longer a technical side issue. It is now a central policy concern.

Land and Water Use by Current LLM Models

The environmental footprint of modern AI is no longer theoretical. Training and operating today’s LLMs requires enormous electricity for computation, massive quantities of water for cooling, and vast land parcels for data-center campuses. Recent studies show that global AI inference already consumes electricity comparable to tens of thousands of households and evaporates water volumes equivalent to the annual needs of over a million people (Jegham et al., 2025; Ecology of AI, 2025).

India’s own context makes this especially concerning. The country’s electricity demand is projected to nearly double by 2030, and AI-driven data-center expansion is emerging as one of the fastest-growing new loads on the national grid (CEA, 2024). Without careful planning, AI risks becoming a silent accelerator of India’s energy and water crises.

Current AI Policy and Lack of Strong Guardrails on Energy Use

India’s national AI framework, anchored by the IndiaAI Mission, rightly emphasizes innovation, democratization, ethics, and public benefit. However, it contains no binding limits or enforceable standards governing the energy, water, or land footprint of AI systems (IndiaAI Mission, 2024).

As global technology companies rapidly expand data-center infrastructure across India, development is racing ahead of environmental governance. In practical terms, this allows the environmental costs of AI to be passed quietly onto India’s power grid, water systems, and communities.

Vizag — A Case Study Where India’s AI Policy Demonstrated Lack of Guardrails

Visakhapatnam’s proposed transformation into an “AI City” illustrates both the promise and the risk of India’s current AI trajectory. While the initiative promises investment, employment, and global visibility, its strategy leans heavily on hyperscale computing and multinational partnerships, with insufficient emphasis on energy efficiency, local innovation, and sustainability alignment (Andhra Pradesh AI City Proposal, 2024).

Large AI installations planned for Vizag would significantly increase electricity demand and cooling-water consumption. Without enforceable sustainability standards, the city risks locking itself into decades of high-cost power contracts, growing water stress, and grid instability — problems that ordinary citizens are likely to end up paying for.

Researchers such as Dr. Sasha Luccioni of Hugging Face have long argued for a different path: small, task-specific, frugal AI models that deliver real public value at a fraction of the environmental cost (Luccioni, 2023; Luccioni et al., 2024). By prioritizing such models alongside necessary infrastructure, Vizag — and India — could foster inclusive innovation while protecting energy security and environmental resilience.

TED Talk by Dr. Sasha Luccioni

Lessons from Other Countries

Around the world, sustainability is beginning to enter AI governance. The European Union’s AI framework encourages environmental transparency and accountability (EU AI Act, 2024). Belgium is exploring energy-allocation limits for data centers to preserve grid stability (Belgium Energy Review, 2025). International initiatives such as the AI Energy Score, developed with Hugging Face, Salesforce, and Carnegie Mellon, now allow AI systems to be compared by environmental impact (AI Energy Score Initiative, 2024).

No country has yet imposed strict consumption caps, but it is becoming increasingly clear that AI policy and climate policy can no longer be treated as separate issues.

Rethinking a More Novel Sustainable Energy Policy for India

India now has a chance to lead the world by embedding sustainability directly into AI governance.

Baseline Energy Standard
Any AI tool launched in India should meet an energy-per-query benchmark comparable to a standard Google-type search.
Tiered Energy & Water Taxation
Systems exceeding this baseline should face a progressive tax on both energy and water use, calibrated to excess consumption.
Incentivize Efficient, Task-Specific AI
Public incentives should favour small, domain-optimized models, echoing the work of Dr. Sasha Luccioni and the broader Green AI movement (Luccioni, 2023; Green AI Consortium, 2024).
Learn from Indian Startups
Startups such as Takeme2space (Hyderabad, Telangana) are reducing land, energy, and water use by shifting heavy computation to satellite-based systems — a model of innovation that India should actively encourage.
Mandatory Environmental Disclosure
AI companies must publicly report energy use, water use, and carbon footprint.
Geographic Load Sensitivity
AI deployment should consider local grid and water stress, restricting intensive workloads in already fragile regions.
AI Hardware Circularity
Firms operating in India should commit to e-waste recovery and rare-earth recycling to reduce material extraction pressure.
Strategic Global Leadership
India’s sustainable-AI framework could be presented at COP summits as a global model.
Resisting Unsustainable Offshoring
India should resist becoming a dumping ground for energy-hungry data centers as Western grids tighten.
Reclaiming ESG in Big Tech
The current AI race increasingly resembles the early fossil-fuel industry — expansion first, responsibility later (IEA Digital Emissions Review, 2024). ESG must return as a design constraint, not a marketing slogan.

Conclusion

AI can become India’s greatest development engine — or its next environmental crisis. In the end, technological leadership should not be measured only by how fast we build machines, but by how carefully we protect the future those machines will influence.

With clear sustainability guardrails, India can build an AI ecosystem that is innovative, inclusive, and environmentally secure — and in doing so, offer the world a new blueprint for responsible digital growth.

References

Jegham et al. (2025). How Hungry is AI? Benchmarking Energy, Water, and Carbon Footprint of LLM Inference.
The Ecology of Artificial Intelligence (2025).
Central Electricity Authority, India (2024).
IndiaAI Mission (2024).
Luccioni, S. (2023). Frugal AI & Sustainable Machine Learning. Hugging Face.
Luccioni et al. (2024). Green AI: Reducing the Environmental Impact of Machine Learning.
EU AI Act (2024).
Belgium Energy Review (2025).
AI Energy Score Initiative (2024).
International Energy Agency (2024). Digitalisation and Emissions.

Ram Ramprasad is a sustainability advocate, author, and policy writer based in PA, USA. He has published extensively on environmental policy, climate resilience, and sustainable innovation in India and internationally.