IIT Bombay Makes History: Launches India's First Institute-Owned AI Company

In a groundbreaking development that marks a watershed moment for India's artificial intelligence landscape, the Indian Institute of Technology Bombay has officially incorporated its own AI company. This isn't just another startup incubated within campus walls it's a fully owned corporate entity that signals a dramatic shift in how premier educational institutions are approaching the AI revolution.

On November 7, 2025, the BharatGen Technology Foundation was registered with the Registrar of Companies in Mumbai, with its registered office at IIT Bombay itself. This historic move makes IIT Bombay the first Indian academic institution to directly own and operate an AI-focused company, moving beyond the traditional role of merely supporting student startups and faculty ventures.

Why This Matters: Breaking New Ground in Academic Innovation

The decision by IIT Bombay to step directly into the corporate arena represents more than just an administrative change. It reflects a growing recognition that building world-class AI infrastructure requires the autonomy, flexibility, and operational framework that only a corporate structure can provide.

Professor Ganesh Ramakrishnan, the Founder Director of BharatGen, emphasized this point clearly when he stated that transitioning from research prototypes to real-world deployment demands "the autonomy of a corporation rather than an academic project." This pragmatic approach acknowledges the limitations that traditional academic structures face when trying to scale cutting-edge technology for national deployment.

BharatGen: Building India's Sovereign AI Infrastructure

At the heart of this initiative lies an ambitious goal developing India's first truly sovereign Large Language Model that understands and serves the country's remarkable diversity. Unlike existing AI models that are predominantly trained on Western data and English language content, BharatGen is being designed from the ground up to work seamlessly across India's linguistic and cultural landscape.

Technical Capabilities and Scale

The numbers behind BharatGen are staggering. The initiative has set its sights on building a foundational AI model with one trillion parameters placing it in the same league as the world's most advanced language models. To put this in perspective, this would make it one of the largest AI models ever developed specifically for a non-English speaking market.

The model already has an impressive foundation. BharatGen has previously released Param-1, a bilingual language model with 2.9 billion parameters that was trained on approximately 5 trillion tokens in English and Hindi. This initial version demonstrated the technical feasibility of building India-centric AI systems and laid the groundwork for the much larger model now in development.

A True National Consortium

BharatGen isn't a solo effort. The initiative brings together some of India's most prestigious technical institutions in an unprecedented collaboration:

• IIT Bombay (Lead Institution)

• IIT Madras

• IIT Kanpur

• IIT Hyderabad

• IIT Mandi

• IIT Kharagpur

• IIIT Hyderabad

• IIIT Delhi

• IIM Indore

This consortium approach pools expertise from across India's top technical minds, creating a collaborative ecosystem that can tackle the massive challenge of building sovereign AI infrastructure.

Unprecedented Funding Support

The Indian government has backed this initiative with substantial financial resources, recognizing its strategic importance for the country's technological sovereignty. The funding structure includes:

• ₹235 crore from the Department of Science and Technology under the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS)

• ₹1,058 crore from the Ministry of Electronics and Information Technology (MeitY) as part of the IndiaAI Mission

• An additional ₹988.6 crore approved specifically for building the one trillion parameter foundational model

This brings the total government investment to well over ₹2,200 crore, making it one of India's most heavily funded public AI programs.

Multilingual and Multimodal by Design

What sets BharatGen apart from existing AI models is its fundamental design philosophy. The system is being built as a multimodal platform that can understand and generate content across three key dimensions:

Language Coverage

Currently, BharatGen supports nine Indian languages: Hindi, Marathi, Tamil, Malayalam, Bengali, Punjabi, Gujarati, Telugu, and Kannada. The roadmap is aggressive but achievable:

• By December 2025: Expansion to 15 Indian languages, including Assamese, Nepali, Odia, Sanskrit, Sindhi, and Maithili

• By June 2026: Complete coverage of all 22 scheduled Indian languages

Multimodal Capabilities

BharatGen integrates three major modalities that work in harmony:

1. Text Processing: Advanced natural language understanding and generation

2. Speech Technology: Comprehensive speech-to-text and text-to-speech systems

3. Vision-Language Models: Document understanding and visual content interpretation

This multimodal approach ensures that the AI can serve users regardless of their literacy levels or preferred mode of communication a crucial consideration for a diverse country like India.

Real-World Applications Already in Motion

While many AI projects remain confined to research labs, BharatGen is already demonstrating practical utility through sector-specific applications:

Krishi Sathi (Agricultural Assistant)

This application is designed to provide farmers with AI-powered guidance in their local languages, covering everything from crop selection to pest management. The goal is to democratize agricultural knowledge and make expert advice accessible to millions of farmers across India.

e-VikrAI (Seller's Platform)

Aimed at small businesses and individual sellers, e-VikrAI leverages AI to help entrepreneurs navigate e-commerce, manage inventory, and communicate with customers more effectively all without requiring sophisticated technical knowledge.

Defense and Governance Applications

BharatGen has also developed specialized applications for defense and governance sectors, with pilots already conducted in select regions. Once deployment is complete, these applications are planned for rollout across all states and districts.

The Bharat Data Sagar Initiative

A cornerstone of BharatGen's strategy is the Bharat Data Sagar project one of the most ambitious data collection initiatives ever undertaken in India. This initiative recognizes a fundamental problem: AI systems can only be as good as the data they're trained on, and currently, the vast majority of AI training data reflects Western contexts, languages, and cultural perspectives.

Bharat Data Sagar aims to build comprehensive datasets that capture India's lived realities, regional diversity, and cultural nuances. This involves large-scale data collection and curation from individuals, institutions, and organizations across all sectors. The goal is not just better AI performance, but genuine digital sovereignty ensuring that India controls its own digital knowledge resources.

Strategic Partnerships for Scale

To ensure widespread adoption and deployment, BharatGen has forged partnerships with major technology companies and government entities:

• IBM: Providing enterprise integration and cloud infrastructure support

• Zoho: Facilitating business application integration

• Multiple State Governments: Enabling localized deployment and customization

These partnerships are crucial for moving BharatGen from an academic project to a national infrastructure that can serve hundreds of millions of users.

Democratizing Access to Sovereign AI

One of the most significant aspects of BharatGen's approach is its commitment to accessibility. Professor Ramakrishnan has emphasized that the foundation will release distilled versions of its models to Indian developers and companies. This democratization strategy means that startups, legacy enterprises, and public institutions can build AI-powered products without bearing the enormous costs of training large-scale models from scratch.

As Ramakrishnan put it, BharatGen is "doing the heavy lifting so the country's innovators can get straight to building." This approach could catalyze an entire ecosystem of India-centric AI applications and services.

The Governance Structure

The BharatGen Technology Foundation operates as a Section 8 company a non-profit corporate structure under Indian law that allows it to pursue social objectives while maintaining the operational flexibility of a private company. The Board includes several ex-officio directors from IIT Bombay:

• Prof. Shireesh Balwant Kedare (Director, IIT Bombay)

• Prof. Ganesh Ramakrishnan (Professor in Charge)

• Prof. Milind Diwakar Atrey (Deputy Director)

• Prof. Ravindra Dheerendra Gudi (Deputy Director)

The foundation has also appointed Shashi Shekhar Vempati, former CEO of Prasar Bharati and IIT Bombay alumnus, as an Independent Director. Vempati brings valuable experience from both the public broadcasting sector and the deep technology space through his work with AI4India.Org.

Addressing the Foreign Dependency Challenge

One of the driving forces behind BharatGen is the recognition that India's current reliance on foreign AI systems creates several problems:

1. Language Limitations: Models like ChatGPT, Google Gemini, and others are optimized primarily for English and other Western languages. Their performance on Indian languages is often suboptimal.

2. Cultural Context: AI trained predominantly on Western data often misses or misinterprets cultural nuances, idioms, and context specific to Indian society.

3. Data Sovereignty: When Indian users interact with foreign AI systems, their data is processed on foreign servers, raising questions about privacy and digital sovereignty.

4. Customization Constraints: Foreign AI systems are built for global markets and may not address India-specific use cases or regulatory requirements.

BharatGen directly tackles these challenges by building AI infrastructure that is fundamentally aligned with India's needs and priorities.

The Timeline Ahead

The BharatGen initiative is operating on an ambitious but realistic timeline:

• 2024-2025: Foundation building, initial model releases, and pilot deployments

• December 2025: Support for 15 Indian languages

• June 2026: Complete coverage of all 22 scheduled languages

• Beyond 2026: Full-scale national deployment and continuous model improvements

Global Context: India's AI Ambitions

This development comes at a time when countries worldwide are recognizing the strategic importance of sovereign AI capabilities. China has been aggressively developing its own AI models, the European Union is investing heavily in AI infrastructure, and now India is making its move.

The launch of BharatGen Technology Foundation signals that India is not content to be merely a consumer of AI technology developed elsewhere. Instead, it's positioning itself as a key player in the global AI landscape, with the potential to set standards for how AI serves diverse, multilingual populations.

What This Means for India's Tech Ecosystem

The ripple effects of this initiative could be profound:

For Startups

Indian AI startups will gain access to powerful, India-optimized models without having to build foundational infrastructure from scratch. This could accelerate innovation and reduce time-to-market for AI-powered products.

For Established Companies

Large Indian enterprises will have sovereign alternatives to foreign AI systems, potentially reducing dependency on international tech giants and addressing data localization concerns.

For Researchers

The academic community gains a world-class platform for AI research that's specifically relevant to Indian contexts, potentially attracting more students and researchers to work on India-centric AI problems.

For Citizens

Ultimately, everyday Indians stand to benefit from AI services that truly understand their languages, contexts, and needs making AI more accessible and useful across all sections of society.

Challenges Ahead

While the vision is compelling, building and deploying a trillion-parameter AI model comes with significant challenges:

Computational Requirements

Training a model of this scale requires massive computational resources. The infrastructure needs are enormous, and maintaining and updating such models is an ongoing challenge.

Data Quality and Bias

Collecting diverse, high-quality data across 22 languages and numerous dialects is a monumental task. Ensuring that the data is representative and free from harmful biases will require continuous vigilance.

Adoption and Integration

Building the technology is only half the battle. Ensuring that BharatGen is actually adopted and integrated into real-world applications will require sustained effort in developer relations, documentation, and ecosystem building.

Competition with Global Giants

BharatGen will compete for mindshare with well-established global AI systems that have significant head starts in terms of capabilities and user base.

The Road Ahead: A New Chapter for Indian AI

The incorporation of BharatGen Technology Foundation marks more than just the creation of another AI company. It represents a fundamental shift in how India approaches technological sovereignty. By having a premier academic institution directly own and operate an AI company, India is creating a new model for bridging the gap between research and deployment.

As Union Minister Dr. Jitendra Singh noted during his visit to IIT Bombay, BharatGen represents India's first sovereign multilingual and multimodal AI effort one that truly reflects the nation's linguistic, cultural, and social diversity. This isn't just about building another language model; it's about ensuring that as AI transforms the world, India's voice, languages, and perspectives are fully represented.

The success of BharatGen could pave the way for similar initiatives in other domains, creating a new paradigm where academic institutions don't just educate and research, but also directly build and deploy critical national infrastructure.

For now, all eyes are on IIT Bombay and its partners as they work to deliver on this ambitious vision. If they succeed, BharatGen could become the foundation for an entire generation of India-first AI applications, truly making AI work for every Indian, in every language, across every context.

Key Takeaways

• IIT Bombay has created India's first institute-owned AI company, BharatGen Technology Foundation, incorporated on November 7, 2025

• The initiative aims to build a trillion-parameter sovereign AI model covering all 22 Indian languages

• Government funding exceeds ₹2,200 crore through multiple programs

• A consortium of top Indian institutions is collaborating on development

• Real-world applications in agriculture, governance, and commerce are already in pilot stages

• The model is designed to be multimodal (text, speech, vision) and culturally aligned with Indian contexts

• BharatGen represents a strategic move toward AI sovereignty and reduced dependence on foreign AI systems

This development signals a new era in India's technological journey one where the country is not just consuming AI technology but actively shaping its future direction.