In a significant move for India’s artificial intelligence landscape, the government is poised to select BharatGen and Fractal as key players in the second phase of the IndiaAI Mission. This initiative aims to bolster the development of sovereign foundational AI models tailored specifically for Indian developers and researchers. The official announcement is expected next week during an event in New Delhi, marking a pivotal moment in the country’s journey toward AI self-sufficiency.
BharatGen, which has garnered support from the Department of Science and Technology (DST), has already made strides in the AI domain with the launch of its first foundational large language model (LLM), Param-1, in May 2025. This model, boasting an impressive 2.9 billion parameters, is designed to cater to the unique linguistic and cultural nuances of India, featuring 25% Indic data. This is a substantial leap compared to global counterparts, such as Meta’s Llama, which incorporates a mere 0.01% Indic data. The emphasis on Indic data not only enhances the model’s relevance but also aligns with the government’s vision of promoting indigenous technology solutions.
The consortium behind BharatGen includes prestigious institutions like IIT Bombay, IIT Kanpur, IIT Mandi, IIT Madras, IIT Hyderabad, IIIT Hyderabad, and IIM Indore. This collaboration brings together some of the brightest minds in academia and research, ensuring that the foundational models developed under this initiative are robust, innovative, and capable of addressing the diverse needs of the Indian populace.
Fractal, a Mumbai-based firm, is another exciting addition to the IndiaAI Mission. The company recently launched its open-source LLM, Fathom-R1-14B, which has been making waves in the AI community. Fractal claims that this model excels in mathematical reasoning, outperforming other models like o1-mini and o3-mini, and approaching the capabilities of o4-mini—all at a remarkably low post-training cost of just $499. This affordability is crucial for democratizing access to advanced AI technologies, particularly for startups and smaller enterprises looking to leverage AI for their operations.
Fractal’s ambitions extend beyond model development; the company is also preparing for an initial public offering (IPO) later this year, signaling its growth trajectory and commitment to becoming a significant player in the AI sector. The inclusion of Fractal in the IndiaAI Mission underscores the government’s recognition of the importance of private sector innovation in driving technological advancements.
The selection of BharatGen and Fractal will place them alongside previously chosen firms such as Soket AI, Gnani.ai, Gan.AI, and Sarvam AI, all of which are working under the mission’s foundation model initiative. This collective effort is part of a broader strategy to create a robust ecosystem for AI development in India, one that emphasizes sovereignty and self-reliance in technology.
The IndiaAI Mission, spearheaded by the Ministry of Electronics and Information Technology (MeitY), has been ramping up its efforts to enhance both computational and model-building capacities. As of May 2025, the total GPU capacity under the IndiaAI Compute Pillar has expanded to an impressive 34,000 units, supported by a mix of public and private players. This infrastructure is vital for training complex AI models and ensuring that Indian researchers have access to the resources they need to innovate.
Union IT Minister Ashwini Vaishnaw has articulated a clear vision for the selected teams: each is expected to emerge as a global top-five player in their respective domains, whether that be multilingual foundation models, speech AI, or multimodal applications. This ambitious goal reflects the government’s commitment to positioning India as a leader in the global AI landscape.
The implications of these developments are profound. By fostering the creation of sovereign AI models, the IndiaAI Mission aims to reduce dependence on foreign technologies and promote homegrown solutions that resonate with local needs. This approach not only enhances national security but also empowers Indian developers and researchers to contribute meaningfully to the global AI discourse.
Moreover, the focus on multilingual capabilities is particularly significant in a country as diverse as India, where over 1,600 languages are spoken. The ability to develop AI models that understand and process multiple languages will facilitate better communication, accessibility, and inclusivity across various sectors, including education, healthcare, and governance.
As BharatGen and Fractal prepare to embark on this journey, they will face challenges typical of any cutting-edge technology initiative. These include ensuring data privacy, addressing ethical considerations in AI deployment, and navigating the complexities of regulatory frameworks. However, the collaborative nature of the IndiaAI Mission, with its emphasis on partnerships between academia, industry, and government, provides a solid foundation for overcoming these hurdles.
The success of BharatGen and Fractal will not only depend on their technical capabilities but also on their ability to engage with stakeholders across the spectrum—from policymakers to end-users. Building trust and transparency in AI systems will be crucial for widespread adoption and acceptance.
In conclusion, the selection of BharatGen and Fractal for the second phase of the IndiaAI Mission marks a significant milestone in India’s quest for AI sovereignty. With a strong focus on developing foundational models that cater to the unique needs of the Indian populace, this initiative has the potential to transform the AI landscape in the country. As these organizations gear up to make their mark, the eyes of the world will be watching, eager to see how India positions itself as a formidable player in the global AI arena. The journey ahead is filled with promise, innovation, and the opportunity to redefine what is possible in the realm of artificial intelligence.