Can India democratise AI by treating compute power as public good? What govt-backed white paper argues

India needs to treat access to artificial intelligence (AI) infrastructure as a public good and lower barriers to compute power, datasets and models to ensure equitable participation in the AI economy, according to a government-backed white paper released on Monday.

The paper, prepared by the Office of the Principal Scientific Adviser, said access to AI has become a critical determinant of innovation, competitiveness and governance, but remains concentrated among a few global firms and urban hubs.

“For India, democratising access means treating AI infrastructure as a shared national resource, empowering innovators across regions to build local-language tools, adapt assistive technologies, and create solutions aligned with India’s diverse needs,” the Office of the Principal Scientific Adviser said on X.

It argued that democratising access would enable startups, researchers, public institutions and smaller organisations across India to build and deploy AI solutions without owning expensive infrastructure.

“India’s priority to democratise access to AI infrastructure requires a scalable and transparent framework that lowers structural barriers while enabling innovation,” the paper said.

What is AI infrastructure?

AI infrastructure refers to the physical and digital foundations required to build and run artificial intelligence systems.

This includes data centres, high-performance computers and specialised chips such as graphics processing units (GPUs) that are essential for training large AI models.

At the digital level, it includes large datasets, software tools and pre-trained AI models.

How strong is India’s AI infrastructure today?

The document noted that while India hosts nearly 20% of the world’s data, it accounts for only about 3% of global data centre capacity. However, installed data centre capacity of around 960 MW is expected to rise sharply to 9.2 GW by 2030 amid growing AI workloads.

India has already taken steps to expand access through initiatives under the IndiaAI Mission, including a national compute platform that offers subsidised access to graphics processing units (GPUs) and a central repository of datasets and models.

The mission is backed by an investment of ₹76,000 crore and has facilitated the approval of 10 advanced chip-making projects, including domestic fabrication and packaging facilities.

The paper said the IndiaAI Compute Portal provides access to more than 38,000 GPUs at rates lower than prevailing global prices, allowing smaller firms and universities to train and fine-tune AI models without owning costly infrastructure.

It also pointed to platforms like IndiaAIKosh and Bhashini, which aim to expand access to datasets and language models tailored to Indian use cases .

Digital Public Infrastructure model for AI

The paper proposed using a Digital Public Infrastructure (DPI) approach for AI, building on India’s experience with population-scale digital platforms.

Such an approach, the paper said, could create standardised access layers for data and compute, reduce costs and administrative barriers, and allow innovators in smaller cities and institutions to participate more easily.

“The DPI for AI approach depends on consistent technical standards, high-quality metadata, and sustained institutional capacity for governance, oversight and auditability across multiple custodians and providers,” it said.

What are the risks and constraints?

The white paper, however, cautioned that democratising access would require careful sequencing, robust privacy safeguards, cybersecurity measures and sustained institutional capacity. It emphasised that public infrastructure alone would not be sufficient and called for public-private partnerships to expand regional data centres and AI-ready facilities beyond major cities.

The paper also raised sustainability concerns as data centres are energy-intensive, and their electricity consumption could rise sharply as AI workloads grow. Several states have begun linking incentives to renewable energy usage, but the paper notes that greener cooling and power systems will be essential as capacity expands.