The Indian government, under the Ministry of Electronics and Information Technology (MeitY), announced on September 18, 2025, the selection of eight organizations to develop foundational large language models (LLMs) as part of the IndiaAI Mission. Led by Union Minister Ashwini Vaishnaw during the curtain-raiser for the AI Impact Summit 2026 in New Delhi, the initiative includes IIT Bombay’s BharatGen consortium, Tech Mahindra, Fractal Analytics, and five others. These entities will create sovereign AI models tailored for Indic languages and sector-specific applications like agriculture and healthcare. With a total outlay of ₹10,372 crore over five years, the mission provides substantial funding—₹988.6 crore for IIT Bombay alone—to foster India’s AI ecosystem, announced at the Bharat Mandapam convention center.
A Leap Toward Sovereign AI
The IndiaAI Mission, approved by the Cabinet in March 2024, aims to democratize AI by building homegrown capabilities, reducing reliance on foreign models like GPT. The eight selected players represent a mix of academia, startups, and corporates, each tasked with specialized LLMs. IIT Bombay’s BharatGen consortium—comprising IITs from Kanpur, Mandi, Madras, Hyderabad, and IIIT Hyderabad—will develop a trillion-parameter model, one of the world’s largest, focusing on Indic use cases for government sectors. Tech Mahindra, already advancing its Indic LLM “Project Indus,” will leverage its expertise in multilingual AI. Fractal Analytics, a leader in analytics, will build reasoning-focused LLMs for healthcare and policy-making. The full list includes Avataar AI (virtual try-on tech), Zeinteiq Aitech Innovations (edge AI), Genloop Intelligence (generative AI), NeuroDX (Intellihealth, healthcare AI), and Shodh AI (research AI). This follows earlier selections like Sarvam AI and Gnani.ai, with the first models expected by February 2026.
Funding and Strategic Focus
The government has allocated significant financial support to accelerate development. IIT Bombay receives ₹988.6 crore for its trillion-parameter behemoth, which uses internal variables to capture language patterns, enabling applications in agriculture, finance, and education. Fractal Analytics gets ₹34.58 crore for reinforcement learning models in national security. Vaishnaw emphasized, “These models will be India’s own, frugal, and open-source where possible,” highlighting Project Indus as a blueprint. The mission also plans 500+ data labs for skilling, addressing India’s talent gap in AI. With global LLMs like GPT-4 boasting trillions of parameters, this push positions India as a sovereign AI contender, backed by 1,000+ GPUs under the mission.
Industry Reactions and Challenges
Tech Mahindra hailed the selection, stating, “Proud to be recognized… Project Indus, built in-house at frugal cost, paves the way for sovereign LLMs.” Fractal Analytics echoed enthusiasm for domain-specific innovations. However, challenges loom: data privacy under the Digital Personal Data Protection Act, ethical AI biases in Indic languages, and competition from global giants like OpenAI. Experts like Prof. Pushpak Bhattacharyya (IIT Bombay) note, “A trillion-parameter model demands massive compute—India must scale infrastructure.” In India’s diverse 780-language landscape, the focus on Indic LLMs could bridge digital divides, empowering rural users in agriculture and health.
A Bold Step for India’s AI Future
The IndiaAI Mission’s eight selections mark a pivotal moment, transforming ambition into action. As Vaishnaw envisions “AI for all,” these LLMs could drive ₹10 lakh crore in economic value by 2030. Yet, success hinges on collaboration and ethics. With IIT Bombay leading the charge, India’s AI sovereignty feels within reach—frugal, innovative, and truly homegrown.
-By Manoj H