BharatGen and Fractal are set to lead Phase 2 of IndiaAI Mission with open source AI models, boosting India’s homegrown AI capabilities and accessibility for developers.
BharatGen is likely to be selected for Phase 2 of the IndiaAI Mission, joining seven other firms including Fractal, Tech Mahindra, Avataar.ai, ZenteiQ.ai, Genloop, Intellihealth, and ShodhAI. This brings the total number of companies under the Mission’s foundation model initiative to 12, alongside existing players Soket AI Labs, gnani.ai, Gan.AI, and Sarvam.
The move significantly boosts India’s open source AI ecosystem. BharatGen is developing open-source foundational models for Indian developers and researchers. Its Param-1 model, a 2.9 billion parameter bilingual LLM, features 25% Indic data, far surpassing global counterparts such as Meta’s Llama. The consortium has also released 20 speech models across 19 Indian languages on AIKosha, targeting voice-first interfaces.
Fractal recently launched its open-source LLM, Fathom-R1-14B, offering high mathematical reasoning performance at a post-training cost of just $499. Fractal is also slated for an IPO later this year.
Prof Ganesh Ramakrishnan of IIT Bombay said, “We have made significant technical progress, and the announcement of our models has already come from the DST secretary.” BharatGen’s consortium includes IIT Bombay, IIT Kanpur, IIT Mandi, IIT Madras, IIT Hyderabad, IIIT Hyderabad, and IIM Indore.
The IndiaAI Mission, backed by the Ministry of Electronics and IT, has ramped up compute capacity, expanding GPUs under its Compute Pillar to 34,000 as of May 2025. Union IT Minister Ashwini Vaishnaw has emphasised that the goal for each team is to become a global top-five player in multilingual foundation models, speech AI, or multimodal applications.
The official Phase 2 announcement is expected next week in New Delhi, while Sarvam, the first selected under the Mission, is expected to launch its foundational model early next year.
