

1 Feb 2025
Ashwini Vaishnav, Union Minister of Electronics and IT, announced at the Utkarsh Odisha Conclave.
India's AI platform will host DeepSeek, the Chinese LLM platform, on local servers. The framework is ready and was launched during the India AI Mission event on Thursday.
The India AI Compute Facility, a common user computing setup, procured 18,693 GPUs to develop an LLM designed especially for Indian startups, developers, and researchers. Of those, 10,000 GPUs are operational. Tata Communications, Jio Platforms, NextGen Data Centre, and Yotta provided the GPUs.
ChatGPT was trained on 25,000 GPUs, while DeepSeek used 2,000. The government sought proposals from AI startups, and six developers are working on models, which are expected to be ready within 6-10 months. 18 apps for the learning disability, agriculture sector, and climate change will be a part of this model.
With a 40% government subsidy, the Indian AI Model computation will cost less than ₹100 per hour compared to other global models that cost $2.5-3 per hour.
Â
Â
Never miss another post from SalestorrsNews150. Follow Salestorrs on WhatsApp, LinkedIn, Facebook, X, and Instagram.