PSA White Paper Outlines Need For Affordable, Inclusive Access To AI Compute, Data & Models

New Delhi, Dec 30 (KNN) The Office of the Principal Scientific Adviser (PSA) to the Government of India has released a White Paper on ‘Democratising Access to AI Infrastructure’, underscoring the need to make computing power, datasets, and AI model ecosystems broadly accessible, affordable, and inclusive to sustain India’s innovation and economic growth.

The paper notes that AI infrastructure—such as high-performance computing, large datasets, and foundational models—is currently concentrated among a few global firms and urban hubs, limiting equitable participation. 

For India, democratising access means treating AI infrastructure as a shared national resource that enables innovators across regions to develop local-language tools, assistive technologies, and solutions aligned with India’s diverse socio-economic needs.

Three Key Enablers for AI Access

Aligned with India’s AI governance vision, the White Paper identifies three core enablers—expanding access to high-quality and representative datasets, ensuring affordable and reliable computing resources, and integrating AI infrastructure with Digital Public Infrastructure (DPI)—while underscoring that broader access to foundational AI resources is essential for risk mitigation, improved accountability, and responsible AI deployment.

State of AI Infrastructure in India

The paper highlights that while India hosts nearly 20 per cent of the world’s data, it accounts for only about 3 per cent of global data centre capacity. However, rapid expansion is underway, with installed capacity expected to rise from about 960 MW currently to 9.2 GW by 2030.

Government initiatives such as the IndiaAI Mission, IndiaAI Compute Portal, IndiaAIKosh, Bhashini, and the National Supercomputing Mission are expanding access to datasets, GPUs, TPUs, and AI-ready tools at subsidised rates. IndiaAIKosh alone has onboarded over 5,700 datasets and 250 AI models across 20 sectors as of December 2025.

Treating AI as a Digital Public Good

The White Paper argues for treating data, compute, and AI models as Digital Public Goods (DPGs), similar to India’s existing digital public infrastructure. This approach would enable shared access through open data repositories, subsidised compute clouds, and open-source model hubs, lowering entry barriers for startups, researchers, universities, and smaller institutions.

Drawing on global examples such as the US National AI Research Resource, the paper suggests a phased, modular DPI-for-AI approach—starting with metadata standards and access protocols, and gradually moving towards federated data access and coordinated compute exchange mechanisms.

Role of States and Industry

State-level policies in Maharashtra, Tamil Nadu, Karnataka, Telangana, and others have played a key role by recognising data centres as essential infrastructure and promoting renewable energy use. 

Industry players such as Yotta, CtrlS, NTT, and AdaniConneX are also investing heavily in hyperscale and sovereign cloud facilities.

Policy Outlook

While cautioning that DPI alone cannot solve all ecosystem challenges, the PSA’s office concludes that democratising access to AI infrastructure is a critical policy priority. 

A balanced framework—combining public-good digital layers, private investment, and strong governance—will be essential to ensure that AI innovation in India is inclusive, sustainable, and aligned with public interest.

(KNN Bureau)