Hemant Darbari

BharatGPT is a multilingual foundation model project built specifically for Indian languages, developed in partnership with IIT Bombay and backed by Reliance Jio. The model is trained on a large corpus of Hindi, Marathi, Gujarati, and other Indic language data, with the goal of making large language model capabilities accessible to India's 1.4 billion citizens in their native tongues.\n\nThe initiative has received approximately $50M in funding and operates under the Hanooman model family, with multiple model sizes targeting mobile and edge deployment. Early deployments span healthcare information access, government services, and vernacular content generation.\n\nAs global AI development remains English-dominated, BharatGPT represents a strategic national-level bet on Indic AI sovereignty. With Jio's distribution infrastructure and IIT Bombay's research depth, the project has a credible path to embedding AI into the daily digital lives of hundreds of millions of Indian users who are not served by English-first models. BharatGPT holds a unique position in the Indian AI ecosystem as one of the only foundation model initiatives backed by a major domestic conglomerate and a premier academic institution. It competes with Sarvam AI, Krutrim, and international multilingual models like Meta's Llama and Google's Gemini, though its deep focus on Indian languages and cultural context provides differentiation. The Reliance Jio backing gives BharatGPT potential distribution advantages across Jio's 450 million subscriber base. As India debates data sovereignty and the need for homegrown AI models that reflect local linguistic diversity, BharatGPT is positioned at the center of that strategic conversation.