Earlier this year, BharatGen unveiled Param-1, a bilingual LLM with 2.9 bn parameters
Earlier this year, BharatGen unveiled Param-1, a bilingual LLM with 2.9 bn parametersThe Ministry of Electronics and Information Technology (MeitY) has announced the third tranche of funding under its Rs 10,000 crore IndiaAI Mission, allocating Rs 1,500 crore to accelerate India’s artificial intelligence capabilities. Union Minister of Electronics & IT Ashwini Vaishnaw made the formal announcement.
Among the eight selected companies, BharatGen has emerged as the largest beneficiary with Rs 989 crore. Launched just 10 months ago, BharatGen is one of India’s first indigenously developed multimodal generative AI initiatives, focused on creating foundational models tailored for Indian languages and contexts.
Earlier this year, BharatGen unveiled Param-1, a bilingual large language model (LLM) with 2.9 billion parameters, pretrained on 5 trillion tokens in English and Hindi. “The intent is to create a national sovereign foundational model ecosystem for generative AI—bringing together top-class academia and industry, with operational grounding,” said Prof. Ganesh Ramakrishnan, IIT Bombay & Principal Investigator, alongside Rishi Bal, Executive Vice President, BharatGen, in a conversation with Business Today.
Of the funds allocated, BharatGen noted that the bulk will be directed towards GPU and compute infrastructure, the most resource-intensive aspect of building AI systems. Less than a quarter will cover manpower, data, and overheads.
Calling AI “a once-in-several-decades transformational opportunity,” Prof. Ramakrishnan highlighted BharatGen’s vision around three pillars:
“Speech is critical given India’s literacy gap, so we focus on dialects, accents, and inclusivity,” he added. The next phase aims to expand BharatGen’s systems to cover all 22 scheduled Indian languages, ensuring equitable access for every citizen.
The BharatGen consortium includes leading academic and research institutions such as IIT Bombay, IIT Madras, IIIT Hyderabad, IIT Kanpur, IIT Hyderabad, IIT Mandi, IIM Indore, IIIT Delhi, and IIT Kharagpur.
So far, BharatGen has released 3B parameter models, speech-to-text, and text-to-speech systems, which are available on Hugging Face and AI Kosh India. The goal is to strengthen the AI ecosystem so that startups and enterprises can build on these foundations.
Recently, BharatGen also partnered with IBM to scale its models for Indian and global markets. “Our ecosystem builds large models that can be distilled into smaller, domain-specific ones—for banking, agriculture, healthcare, and more. When you build for India, you build for the world, because our cost-efficient, culturally aware models can serve markets with similar needs,” Bal explained.