India is poised to make a significant leap in artificial intelligence with the imminent launch of PARAM-2, a sophisticated 17-billion-parameter multilingual AI model.

This development will be unveiled at the India AI Impact Summit 2026. PARAM-2 emerges from the BharatGen program, underscoring India's commitment to fostering sovereign foundational AI models tailored to its unique linguistic, governance, and cultural landscape.

BharatGen serves as India's national generative AI initiative, backed by the Department of Science and Technology (DST). Over recent years, the program has methodically established a robust foundation, positioning India among a select cadre of nations equipped to engineer large-scale AI models independently. This endeavour reflects a strategic pivot towards self-reliance in cutting-edge technology.

At its core, PARAM-2 accommodates all 22 Scheduled Indian languages, having been rigorously trained on datasets curated under the Bharat Data Sagar initiative. This India-centric data repository ensures the model resonates with local contexts, from regional dialects to societal nuances. Such localisation addresses the limitations of Western-dominated AI systems, which often falter in non-English environments.

Architecturally, PARAM-2 employs a Mixture-of-Experts framework, enabling it to process intricate multilingual tasks with remarkable efficiency. This design optimises performance by activating specialised 'expert' sub-models only when needed, conserving computational resources while delivering high accuracy. It marks a technical milestone in scalable AI deployment for resource-constrained settings.

Professor Ganesh Ramakrishnan of IIT Bombay, a pivotal figure in BharatGen's leadership, views the PARAM-2 launch as far more than a mere model debut. He emphasises its embodiment of collaborative synergy among researchers, academic institutions, government entities, and industry stakeholders. This collective effort aims to empower India to architect its AI destiny, free from external dependencies.

In stark contrast to commercial giants like ChatGPT or Gemini, BharatGen eschews a centralised business-to-consumer paradigm. Instead, it disseminates AI models as national public digital goods. This public infrastructure model democratises access, allowing seamless integration into diverse sectors without proprietary lock-ins.

Government departments, banks, hospitals, courts, and educational establishments stand to benefit immensely. They can deploy PARAM-2 locally, even in air-gapped, secure environments devoid of internet connectivity. This capability is crucial for sensitive operations, mitigating risks associated with cloud-based foreign services.

The approach prioritises transparency, trust, and data sovereignty—imperatives in an era of escalating geopolitical tensions over technology control. By keeping data and models within national borders, BharatGen safeguards against surveillance, biases, and undue influence from global tech overlords. It aligns with India's broader digital public goods philosophy, akin to successes in UPI and Aadhaar.

BharatGen's foundational support stems from the National Mission on Interdisciplinary Cyber-Physical Systems, under DST auspices. This mission injected an initial ₹235 crore to germinate the project, fostering early prototypes and talent pipelines. Its success has catalysed expansion, drawing further investment.

Scaling ambitions now intersect with the IndiaAI Mission, overseen by the Ministry of Electronics and Information Technology. This national strategy has earmarked ₹900 crore for BharatGen's maturation, embedding it within a comprehensive AI ecosystem. Additional funding will accelerate model iterations, dataset enrichment, and deployment infrastructure.

This public digital infrastructure model for AI carries profound implications for India's development trajectory. It could revolutionise public service delivery, from multilingual judicial transcription to AI-assisted healthcare diagnostics in remote areas. Educational tools powered by PARAM-2 might personalise learning in vernacular tongues, bridging urban-rural divides.

Economically, BharatGen promises to bolster indigenous innovation, curbing reliance on imported AI solutions that drain foreign exchange. It incentivises private sector participation, potentially spawning a vibrant AI start-up ecosystem attuned to Indian needs. Collaborations with firms in defence, aerospace, and manufacturing—key pillars of India's self-reliance drive—could yield sector-specific adaptations.

Geopolitically, the initiative enhances India's stature in the global AI race. As nations grapple with AI governance, India's open, sovereign model offers a compelling alternative to Big Tech hegemony. It positions New Delhi as a thought leader in ethical AI, advocating for inclusive frameworks at forums like the G20.

The trajectory is promising. BharatGen exemplifies how public investment in frontier technologies can yield multiplier effects, much like India's space program. As PARAM-2 debuts, it signals India's audacious bet on AI as a cornerstone of Viksit Bharat—a developed nation by 2047.

Agencies