Ashwini Vaishnaw announces 8 firms for IndiaAI Mission, IIT Bombay to head trillion-parameter model

by

New Delhi: India is turning the dial up on artificial intelligence research. At the AI Impact Summit 2026 in New Delhi, the government revealed the next set of organisations chosen to build large language models (LLMs) under the IndiaAI Mission. IT Minister Ashwini Vaishnaw made the announcement on September 18, calling it the most ambitious phase yet of the country’s AI push.

Among the eight names, IIT Bombay stands out. Through its BharatGen consortium, the institute has been handed the challenge of developing an LLM with one trillion parameters. That number might sound abstract, but in AI circles it’s a sign of scale, sophistication, and national ambition. For context, India has approved Rs 988.6 crore in support for this project alone.

BharatGen and the trillion-parameter challenge

In simple terms, parameters are the building blocks of an AI model. They are the hidden variables that let the system understand patterns, relationships, and language. More parameters usually mean a deeper understanding, though it also means far higher costs in training and hardware. A trillion-parameter model puts IIT Bombay in a race alongside global giants working on similar scales.

For India, this is more than just a technical milestone. It is seen as a statement of intent, proving that the country can build home-grown AI systems at a level that competes globally.

The eight chosen organisations

Apart from IIT Bombay’s BharatGen, the government named:

  • Avataar AI
  • Fractal Analytics Limited
  • Tech Mahindra Ltd
  • Zeinteiq Aitech Innovations
  • Genloop Intelligence Pvt Ltd
  • NeuroDX (Intellihealth)
  • Shodh AI

Each of these teams will contribute to building foundational models, aimed at different applications and sectors.

Tech Mahindra highlights its Indic LLM

Tech Mahindra, which has been working on Project Indus, welcomed the recognition. In its statement, the company said it was “proud to be recognised” as part of the IndiaAI Mission. The note added, “This announcement comes on the back of building our own and India’s own Indic LLM, Project Indus. Built completely in-house and at frugal cost, the journey of having Project Indus as open source to creating sovereign LLMs has been a learning and rewarding experience.”

Project Indus has already gained attention in the AI community for focusing on Indian languages, a gap global models often overlook.

Earlier selections and GPU access

This isn’t the first round of announcements. Back in May 2025, three startups — SoketAI, Gnani.ai, and Gan AI — were selected to create India’s first home-grown foundational models. A month before that, four other startups, including Sarvam AI, were identified to focus on specialised systems.

Training large AI models isn’t just about algorithms. It needs massive computing muscle, usually in the form of GPUs. The government has worked with cloud and data providers to make this access easier for the selected companies.

What’s next for IndiaAI Mission

Minister Ashwini Vaishnaw sounded upbeat about the progress. “The models that were selected earlier are progressing really well and I am confident that by the time the AI Impact Summit gets underway in February 2026, India will have a model or models ready,” he said at the event.

The next big step will be a formal AI framework, currently being drafted by the Ministry of Electronics and IT along with the Office of the Principal Scientific Adviser. It is expected to outline how AI should be built and used responsibly across the country.