India is making a big move in Artificial Intelligence (AI) by creating its own large language model (LLM), similar to ChatGPT and DeepSeek. Union Minister of Electronics and IT, Ashwini Vaishnaw, announced that the AI model will be ready in 10 months. This is a major step in India’s goal to become self-reliant in AI technology.
Speaking at the Utkarsh Odisha Conclave, Vaishnaw shared that the groundwork for the AI model is complete. The government is now focused on building a system that meets the specific needs of Indian users. Over the past 1.5 years, teams have worked with startups, researchers, and professors to prepare for this project. The new AI model will be designed to understand India’s languages, culture, and context while avoiding biases.
India has strong infrastructure to support its AI ambitions. The country has exceeded its initial target of 10,000 GPUs (graphics processing units), now having 18,600 GPUs. These GPUs, including high-performance models like NVIDIA H100, H200, and MI325, are crucial for training the AI model. For comparison, DeepSeek AI was trained on 2,500 GPUs, and ChatGPT on 25,000 GPUs. India’s 15,000 high-end GPUs put it in a strong position in the global AI race.
The government has also set up a shared computing facility with 18,000 GPUs. This facility will help AI startups, developers, and researchers access advanced computing resources without high costs. Already, 10,000 GPUs are operational, with more coming soon. Vaishnaw highlighted that this facility will make AI development more accessible to everyone.
Currently, six major developers are working on India’s AI model. The first versions are expected to be ready in 4 to 10 months. The model is being designed to handle India’s diverse languages and cultures. Vaishnaw is confident that India can create a world-class AI model quickly, thanks to improvements in algorithmic efficiency.
Source: msn
