The Telangana Data Exchange (TGDeX) represents a significant step for India as its first digital public infrastructure (DPI) designed for artificial intelligence (AI), developed by the Telangana state government. This initiative aims to amalgamate government, academic, and private sector data sets into a unified platform, with a target of generating 2,000 AI-ready datasets over a five-year period from 2025 to 2030. By integrating these diverse data sources, TGDeX seeks to facilitate collaborative AI development while ensuring data sovereignty.
A white paper titled “Democratising Access to AI Infrastructure,” published on December 29, 2025, by the Office of the Principal Scientific Adviser to the Government of India, emphasizes the importance of this convergence between DPI and AI. The document outlines how TGDeX enables stakeholders to collaborate securely and in compliance with privacy regulations, allowing for the sharing of datasets without the necessity of moving raw data. This approach aims to reshape India’s AI policy and governance landscape.
The white paper identifies democratising access to AI infrastructure as a crucial policy priority for India. It asserts that for the nation, this democratization involves treating foundational AI components as shared resources, thereby allowing innovators across the country to play a role in shaping the AI landscape. The paper raises concerns about the centralized control of AI resources by global firms, which has led to disparities in AI participation and use.
One of the pivotal recommendations from the subcommittee was to foster the integration of AI with DPI, similar to the framework established by TGDeX. The adoption of a DPI approach for AI suggests the establishment of a modular set of public goods designed to address existing gaps within the AI ecosystem. A critical step towards this goal involves identifying technical pathways to construct an integrated stack that connects essential AI infrastructure layers, including computing power and data management.
The next stage would be to create a cohesive technical architecture that links these layers to mitigate fragmentation, thereby making the utilization of computing power and datasets more streamlined. Initial governmental efforts could focus on implementing “lighter weight elements,” such as metadata standards, access protocols, and registries, before progressing to more advanced systems like federated data access and consent-based data flows. The paper emphasizes that the development of DPI for AI should occur in a phased and modular manner.
This approach to DPI could standardize interfaces for interoperability, establish common governance norms, and create transparency, ultimately lowering the costs associated with using AI infrastructure. It is particularly valuable for smaller firms, research institutions, and startups, which frequently encounter significant barriers to entry in the AI field. The white paper reiterates the necessity of rolling out a DPI approach in modular phases, ensuring it complements other initiatives such as government investments in AI capacity building and infrastructure expansion.
Furthermore, the paper highlights the importance of creating reusable, open layers, or digital public goods, that lower entry barriers and empower innovators nationwide to build, test, and deploy AI solutions. These digital public goods comprise open data repositories, subsidized compute clouds, and open-source model hubs, all contributing to a more accessible AI ecosystem.
As India embarks on this ambitious journey toward democratizing AI access, the TGDeX initiative and the supporting white paper outline a strategic framework that reflects not only the nation’s aspirations but also the challenges posed by a rapidly evolving technological landscape. By fostering collaboration among various stakeholders and emphasizing the importance of equitable resource distribution, India aims to position itself as an active participant in the global AI arena, paving the way for innovations that could benefit its diverse population.
See also
Canadians Demand AI Regulation as Government Prepares New Strategy Amid Public Skepticism
OpenAI Launches ‘OpenAI for Countries’ Initiative to Bridge Global AI Application Gap
Gobind Singh Deo Featured in Apolitical’s 2023 Government AI 100 List for Digital Leadership
CGI Partners with Google Cloud to Enhance AI-Driven Government Workflows Amid Market Challenges
OpenAI Partners with Leidos to Enhance AI Deployment for Federal Agency Missions
