Here are the 8 New Tech Firms Scaling Under IndiaAI Mission’…

IndiaAI Mission has announced the selection of eight firms for the second phase of its foundation model initiative. As AIM exclusively reported on September 12, 2025,  the list includes BharatGen, Tech Mahindra, and Fractal, along with Avataar.ai, ZenteiQ.ai, Genloop, NeuroDX (IntelliHealth), and Shodh AI.

The government announced the list at the curtain raiser event of the IndiaAI Impact Summit. The event by the Ministry of Electronics & Information Technology (MeitY)  is scheduled for February 2026. With this, the total number of firms under the Mission’s foundation model program has gone up to 12.  It includes four previously selected players—Sarvam AI, Soket AI Labs, gnani.ai, and Gan.AI.

The mission’s goal is to build indigenous large language and multimodal AI models trained on India-specific datasets. Since its call for proposals on January 30, 2025, the mission received 506 proposals in a span of three months. 

The New Startups

Union information and technology minister Ashwini Vaishnaw felicitated the newly selected startups, beginning with Avataar.ai. The company is building a suite of domain-specific models, called Avataars, with up to 70 billion parameters. 

These will be optimised for Indian languages and contexts such as agriculture, healthcare, and governance, while keeping infrastructure costs low and scalable. The models will be shared on AI Coach. 

The IIT Bombay-led BharatGen Consortium, backed by the department of science and technology, has ambitious plans to develop multilingual and multimodal models ranging from 2 billion to 1 trillion parameters, integrating text, speech, and images. 

It also aims to build smaller domain-specific LLMs for agriculture, governance, finance, law, health, and education. BharatGen’s approach remains open—open source, open weights, and open recipes—to create sovereign Indian foundational models.

This adds to BharatGen’s recent milestones. In May 2025, the Consortium launched its first foundational LLM, Param-1, a 2.9 billion parameter bilingual model built entirely from scratch. Param-1 had 25% Indic data—far more than Meta’s Llama, which had only 0.01%. BharatGen’s new proposal expands this ambition with models scaling up to 1 trillion parameters.

Fractal Analytics, the company that is likely to go for an IPO soon, has proposed to build India’s first large reasoning model, with up to 70 billion parameters, focused on structured reasoning, deliberate problem solving, and agentic decision-making. 

The focus is on STEM and medical reasoning, with new benchmarks created for Indian contexts. 

Meanwhile, Tech Mahindra’s Makers Lab, too, joined the list. Its plan involves creating an 8 billion parameter model specifically tuned to Indic language groups, with a focus on Hindi dialects. Alongside, it is building an agentic AI platform ‘Orion’ for real-time intelligence to be deployed in government and beyond. 

ZenteiQ.ai (formerly Zentech AI Tech Innovations) has proposed BrahmAI, pitched as India’s first science-driven foundation AI model for engineering intelligence, scientific computing, and industrial innovation. 

The initiative will deliver multimodal models ranging from 8 billion to 80 billion parameters, with applications powered by a robust data infrastructure. 

Genloop is taking a different path with smaller models. Its project is to build 2 billion parameter language models designed natively for all 22 scheduled Indian languages, prioritising reasoning capabilities over translation-based approaches. The three models—Yukti, Varta, and Kavach—will also have in-built content moderation.

Another startup, NeuroDX (IntelliHealth), is working on a 20 billion parameter foundation model for EEG signal analysis, with the goal of enabling early screening of neurological disorders and building brain-computer interfaces. The plan is to create affordable, non-invasive diagnostic tools and integrate human-AI collaboration in neuroscience using transformer-based architectures. 

And finally, Shodh AI, which is developing a 7 billion parameter foundation model for material discovery, is now part of the Mission. Its automation framework integrates AI into every step of the discovery process, from data gathering to experiment planning and evaluation. 

What’s in the pipeline?

In less than a year, the IndiaAI Mission has turned into one of the largest GPU programmes. More than 34,000 GPUs are already empanelled, four times the original 10,000 target. Another 6,000 are in the pipeline, bringing the total to nearly 40,000 GPUs.

Sarvam AI is the biggest beneficiary so far. The Bengaluru startup landed 4,096 NVIDIA H100s through Yotta Data Services and nearly ₹99 crore in subsidies. Sarvam is expected to ship India’s first large language model by early next year, though the launch has been delayed from its original six-month target.

Soket AI Labs is planning a 120-billion-parameter Indic language model. It will start with a 7-billion model in six months, scale to 30 billion, and then 120 billion within a year. It has already released a 1-billion-parameter model earlier this year.

Gnani.ai is another key beneficiary. GPU and Cloud service provider E2E Networks recently secured a ₹177 crore order to supply GPU resources to Gnani.ai. The deal covers 1.3 crore GPU hours over a year, with H100 and H200 units allocated.

With these selections, IndiaAI has now created a 12-company cohort, tasked with building the backbone of India’s sovereign AI ecosystem.

The post Here are the 8 New Tech Firms Scaling Under IndiaAI Mission’s Phase 2 appeared first on Analytics India Magazine.

Generated by RSStT. The copyright belongs to the original author.

Source