India’s AI White Paper lays out plan to expand compute, data access

India is moving to widen access to artificial intelligence infrastructure as it looks to prevent computing power, data and advanced AI models from becoming concentrated in the hands of a few large companies and major urban centres.

In a White Paper series released by the Office of the Principal Scientific Adviser, the government said democratising access to AI infrastructure would be a key policy priority as India builds its long-term AI governance framework. The papers are aimed at guiding policymakers, industry and researchers at a time when access to compute and data increasingly determines who can innovate, and who cannot.

The challenge is scale. India generates nearly 20% of the world’s data but accounts for just 3% of global data centre capacity. With AI workloads rising fast, the country’s installed data centre capacity of about 960 megawatts is expected to jump to 9.2 gigawatts by 2030, according to the paper.

Data centre capacity remains heavily concentrated. Mumbai and Navi Mumbai together account for more than 25% of live capacity, helped by dense subsea cable networks and supportive local policies. Chennai follows with roughly 13%, while Bengaluru and Hyderabad each hold about 22%. The Delhi NCR region accounts for around 14%, with Pune and Kolkata making up smaller shares.

To bridge the gap, the government is pushing a model that treats core AI resources--compute, datasets and models—as shared digital infrastructure rather than proprietary assets. Under the IndiaAI Mission, a secure GPU cluster housing 3,000 next-generation processors is being built for sovereign and strategic use. Platforms such as AIKosha are also being developed as central hubs for datasets and pre-trained models, including AI tools trained on Indian data.

The expansion, however, comes with rising costs. Scaling data centres could require an additional 45 to 50 million square feet of real estate by 2030. Data centres currently consume about 0.5% of India’s electricity, a figure that could climb to nearly 3% by the end of the decade. The data centre cooling market, valued at $2.1 billion in 2024, is projected to grow to $7.13 billion by 2030.

The government is urging greater use of renewable energy, energy-efficient cooling systems and hybrid power sources as capacity expands. States such as Maharashtra and Tamil Nadu have already introduced renewable energy mandates for data centres.

While sectors like telecom, media, pharmaceuticals and manufacturing are adopting AI at scale, usage remains uneven in agriculture, healthcare, education and public services due to limited access to infrastructure and data.

Officials say widening access to compute and datasets will be critical if AI is to deliver economic gains beyond big tech firms and metro hubs—spreading benefits to startups, researchers and public institutions across the country.