IIT Bombay's BharatGen Leads India’s Big AI Build

News

1. Half of Indian Treasury Teams Put Automation First, EY Report Says

2. IIT Bombay's BharatGen Leads India’s Big AI Build

3. TCS to Modernise Vodafone Idea’s BSS with AI in 5-Year Deal

4. AI Turns Rivals Into Allies as Nvidia Bets $5 Billion on Intel

5. WTO Says AI Could Lift Global Trade Nearly 40% by 2040

6. Moody’s Puts AI Agents to Work Across Finance, Risk and Strategy

The BharatGen consortium led by IIT Bombay has landed the biggest grant under India’s new AI model-building program, with the government awarding ₹988 crore ($112 million) to develop multilingual, open-source foundation models tailored to Indian needs.

The consortium is among eight projects selected as part of the IndiaAI Mission’s push to build homegrown, culturally grounded artificial intelligence capabilities at scale.

Chosen from more than 500 proposals, the BharatGen project stands out for its ambition to create foundational models ranging from 2 billion to 1 trillion parameters that can support AI applications across agriculture, healthcare, education, finance, and law.

The models will be multilingual, multimodal, and open-source, and are expected to serve as the technological base for a range of public and private-sector use cases in India.

The other seven projects focus on highly specialized AI models for reasoning, healthcare, governance, and scientific research.

Avatar AI is building a 70B-parameter avatar model tailored for Indian languages and public-sector use cases, including agriculture, healthcare, and governance. The company, known for developing agentic AI platforms, focuses on making enterprise workflows more efficient by deploying contextual AI agents at scale.

Fractal Analytics is building what it describes as India’s first domain-specific reasoning model, geared toward structured problem-solving in science and medicine.

“This is not just a project,” wrote Srikanth Velamkanni, Co-founder and Group Chief Executive of Fractal, in a LinkedIn post announcing the selection. “It’s a national mission to place India at the forefront of AI innovation.”

“Fractal will build sovereign reasoning models from small to large scale, ranging from 2 billion to 70 billion parameters. In addition, we will develop specialised medical reasoning models to advance healthcare and create a global benchmark in reasoning, built on India’s toughest exams like JEE, NEET, Olympiads, CAT and GATE,” he said.

Velamkanni added that these models would be designed with India’s cultural and linguistic diversity in mind. “We’ll build models grounded in our linguistic and cultural context to ensure fairness, neutrality and sovereignty. Reasoning is the foundation of intelligence, and by investing in reasoning capabilities, we can build powerful, trustworthy, ethical, and inclusive AI,” he wrote.

Tech Mahindra’s Maker’s Lab is working on an 8B-parameter Indic language model and building a generative agentic platform called Orion, which will support digital public infrastructure and enhance AI applications in government service delivery.

Zenteiq is developing BrahmAI, a science-focused foundation model aimed at accelerating innovation in industrial and engineering sectors. The model will be trained on scientific data to help advance India’s R&D capabilities in applied technologies.

GenLoop is creating a family of compact 2B-parameter models that support all 22 scheduled Indian languages. These models will focus on safe reasoning, multilingual performance, and broad accessibility, addressing the needs of both enterprise and public platforms.

Intellihealth is building a 20B-parameter healthcare model that will process EEG signals and support brain-computer interface applications. The focus is on improving diagnostics and neuro-based healthcare delivery, especially in contexts where traditional models may fall short.

Shodh AI is creating a 7B-parameter AI model aimed at accelerating material science research. By embedding AI into lab experimentation workflows, the model is intended to support faster discovery and innovation in areas like advanced materials, electronics, and clean energy solutions.

The overall IndiaAI Mission, approved by the Cabinet in March 2024, has a total outlay of ₹10,371.92 crore over five years. It covers six major components: IndiaAI Compute Capacity, IndiaAI Innovation Centre, IndiaAI Datasets Platform, IndiaAI Application Development Initiative, IndiaAI FutureSkills, and the IndiaAI Startup Financing mechanism.

At the launch of the logo for the India-AI Impact Summit 2026 on Thursday, 18 September, IT Minister Ashwini Vaishnaw said India now has access to about 38,000 high-end graphic processor units used to train and run AI systems.

He also said the government has opened 30 training and research labs for data and AI across the country, including in smaller cities, and plans to scale this to 570 labs.

The first 27 labs were set up with NIELIT, a government body that runs electronics and IT training programs. The next three labs in Mokokchung, Mhow, and Mohali were opened with Intel.

These labs will deliver foundational courses in data annotation, curation, and applied AI under the IndiaAI FutureSkills initiative, with the goal of building a skilled workforce for the country’s growing AI ecosystem.

The India-AI Impact Summit will take place on 19–20 February 2026 at Bharat Mandapam, New Delhi.

The event will feature the AI Expo for Responsible Intelligence, showcasing more than 300 exhibitors from India and over 30 countries. Flagship programmes include UDAAN, a pitch festival for startups from Tier 2 and Tier 3 cities, with a spotlight on women founders and differently abled entrepreneurs. A Global Innovation Challenge and a research symposium will run in parallel.

The summit’s official Ashoka Chakra–themed logo is meant to reflect India’s constitutional values and commitment to ethical AI development.