India is on the brink of a significant AI breakthrough with the development of its first large language model (LLM), similar to ChatGPT and DeepSeek. According to Union Minister of Electronics and Information Technology Ashwini Vaishnaw, work on the AI model is progressing steadily and is expected to be completed within 10 months. This marks a crucial step toward India's self-reliance in artificial intelligence.

What’s Driving India’s AI Ambitions?

Speaking at the Utkarsh Odisha Conclave, Vaishnaw emphasized that extensive groundwork has already been laid. The AI model's framework is in place, and the remaining challenge is refining the system to meet India's specific needs.

"For the past 1.5 years, our teams have collaborated closely with startups, researchers, and academic experts. Today, we are inviting proposals to build our own foundational model. This AI model will be tailored to the Indian context, encompassing regional languages and cultural diversity, while ensuring neutrality and fairness," Vaishnaw stated.

India’s AI Infrastructure: More GPUs, Greater Computational Power

A robust computing infrastructure is key to India’s AI aspirations. The country has exceeded its initial goal of securing 10,000 GPUs, now boasting 18,600. These powerful processors will serve as the backbone for training India's AI model.

For comparison, DeepSeek AI was trained using 2,500 GPUs, while ChatGPT utilized approximately 25,000 GPUs. India currently possesses over 15,000 high-end GPUs, including NVIDIA H100, H200, and MI325 models. This places India in a strong position to develop a globally competitive language model.

Democratizing AI: A Common Computing Facility for Startups and Researchers

To ensure broader participation in AI development, the Indian government has launched a shared computing facility equipped with 18,000 GPUs. This initiative aims to remove financial barriers that often hinder smaller players from accessing cutting-edge AI infrastructure. Of these, 10,000 GPUs are already operational, with the remaining units expected to be activated soon.

Vaishnaw highlighted that this initiative will enable Indian innovators to contribute to AI advancements without facing prohibitive costs.

AI Model Expected Within 4 to 10 Months

India's AI model is being developed by six major teams, with the first versions expected to be ready in the next 4 to 10 months. The model is being designed to accommodate India's linguistic and cultural diversity, ensuring efficiency and inclusivity.

Vaishnaw expressed confidence in the project, stating that advancements in algorithmic efficiency will help India develop a competitive AI model within a shorter time frame. With strong government support, state-of-the-art infrastructure, and a focus on inclusivity, India is poised to make a significant impact on the global AI landscape.