India has chosen eight organisations, including IIT Bombay, Tech Mahindra, and Fractal Analytics, to lead the development of foundational large language models (LLMs) under the IndiaAI Mission. The announcement was made at the AI Impact Summit 2026 in New Delhi by IT Minister Ashwini Vaishnaw.
The most ambitious project will be undertaken by IIT Bombay through its BharatGen consortium, which has been tasked with building a trillion-parameter LLM. This initiative, backed by ₹988.6 crore in funding, aims to put India at the forefront of AI research and development, competing with global leaders.
In AI, “parameters” are internal variables that help models learn from data and identify patterns. The larger the number of parameters, the stronger the model’s ability to perform advanced language tasks. Developing a trillion-parameter model is a significant step toward India’s goal of AI self-reliance and technological leadership.
The selected organisations are: Avataar AI, IIT Bombay (BharatGen), Fractal Analytics, Tech Mahindra, Zeinteiq Aitech Innovations, Genloop Intelligence Pvt Ltd, NeuroDX (Intellihealth), and Shodh AI. Each will focus on building foundational models tailored to specific applications.
Tech Mahindra welcomed the move, highlighting its prior work on Project Indus, an Indic large language model developed at low cost, which has provided valuable experience for building sovereign AI systems.
Earlier this year, startups were chosen to create India’s first foundational AI models and specialised tools. Alongside this, the government is working with cloud and data providers to ensure access to GPUs and other critical resources for training large-scale models.
Minister Vaishnaw expressed confidence that by the next AI Impact Summit in February 2026, one or more of these models could be ready. He also confirmed that a framework for responsible AI is being developed by the Ministry of Electronics & IT in collaboration with the Office of the Principal Scientific Adviser.
