India Lays the Roadmap for Democratizing Access to AI
The Government of India has unveiled a whitepaper aimed at establishing a robust AI infrastructure that is accessible to all citizens. This initiative particularly focuses on individuals who primarily use regional languages and those residing in non-metropolitan areas. Titled Democratising Access to AI Infrastructure, the document outlines strategies for equitable participation in India’s digital transformation and economic growth.
Current AI Infrastructure Landscape
According to the whitepaper, prepared by a senior policy fellow at the Office of the Principal Scientific Adviser to the Government of India, access to AI infrastructure—including compute power, data repositories, and model ecosystems—is crucial for innovation, competitiveness, and governance in the digital economy. The report highlights that these resources are predominantly concentrated among a few global firms and urban hubs, which limits equitable participation.
Wider and More Equitable Access
To democratize AI infrastructure, the government emphasizes the need for institutional governance frameworks that treat the essential components of AI systems as Digital Public Goods (DPGs). This approach allows stakeholders to utilize data, compute, and models without being restricted by geographical limitations.
Despite hosting nearly 20 percent of the world’s data, India only accounts for 3 percent of global data center capacity. As the demand for data infrastructure increases, the current installed capacity of 960 MW is projected to rise to 9.2 GW by 2030.
Development of AI Infrastructure
The whitepaper outlines the IndiaAI Mission, which includes the construction of a secure GPU cluster to house 3,000 next-generation GPUs for sovereign and strategic applications. Additionally, the India Semiconductor Mission (ISM) aims to bolster the foundational layer of processing units through a significant investment of Rs. 76,000 crore, facilitating the approval of ten advanced chip-making projects.
AI Data and Compute Access
Launched in 2025, IndiaAIKosh serves as a national repository for AI datasets, models, and tools, organizing data across 20 sectors. As of December 2025, it has onboarded 5,722 datasets and 251 AI models from various entities. The platform allows for permission-based access, enabling contributors to control data usage while facilitating AI development.
The Bhashini initiative aims to create language datasets and models for India’s diverse languages. This is particularly important given the country’s over 20 official languages and thousands of dialects.
Regarding compute access, the national GPU pool is being enhanced by the IndiaAI Mission and is available through a government-supported cloud infrastructure. The IndiaAI Compute Portal operates over 38,000 GPUs and 1,050 TPUs, facilitating research, startup activities, and governmental projects. This initiative allows smaller cities to train models at subsidized rates and supports universities in conducting advanced AI research.
Regulatory Environment and Policies
The report highlights the MeghRaj (GI Cloud) initiative, supported by the Ministry of Electronics and Information Technology (MeitY), which establishes a foundation for AI-oriented public storage. Several states, including Maharashtra, Tamil Nadu, Karnataka, and Telangana, have developed their own data center policies, with a national policy under consideration.
Recommendations for Action
To effectively democratize access to AI infrastructure, the whitepaper advocates for a scalable and transparent framework that diminishes structural barriers while fostering innovation. This involves making foundational AI resources, such as compute capacity and high-quality datasets, widely available. The goal is to enable a diverse range of actors to build, test, and deploy AI responsibly.
The report emphasizes a Digital Public Infrastructure (DPI) approach to advance AI access. This involves establishing shared, standards-based layers that enhance access, interoperability, accountability, and trust. DPI should not be seen as a singular platform but as a set of modular public-good enablers that address specific coordination gaps in the AI ecosystem.
Encouraging Private Sector Involvement
Finally, the document encourages the private sector to enhance India’s digital transformation by developing edge infrastructure in non-metro areas. It also stresses the importance of resource efficiency, noting that data centers currently consume approximately 0.5 percent of India’s total electricity, a figure expected to rise significantly by 2030 due to increased capacity and workloads.
While sectors like telecom, media, and pharmaceuticals are rapidly adopting AI, areas such as agriculture, education, healthcare, and public services lag behind due to inadequate infrastructure and resource access.
