India pushes to develop low-cost AI models to compete with ChatGPT and DeepSeek

Artificial intelligence (AI) has taken the world by storm with large language models (LLMs) gaining popularity for their versatility. ChatGPT is a prime example of a high-performance AI chatbot model, along with many new breakthrough models like DeepSeek from China. A tech powerhouse like India certainly doesn't want to be left out. The world's most populous country seems to be keen to compete head-to-head in the global AI space by developing its own advanced large language models, many of which could be released as early as this year.

Indian government to launch low-cost AI model

At a recent event on AI, India's Minister of Electronics and Information Technology Ashwini Vaishnaw said the country is developing its own set of basic AI models. He added that the model will work similarly to DeepSeek and ChatGPT, but at a much more affordable cost to develop and operate. The official said the new AI model could be ready in the next eight to 10 months.

At the Indian AI Mission event, Ashwini Vaishnaw revealed that researchers in India are developing an AI ecosystem framework to support the country's basic AI model. Such a model is developed to provide a tailored experience for Indian users. It will also understand the linguistic and contextual needs of Indian users, providing comprehensive coverage while eliminating biases.

Not much information has been revealed yet. However, according to Mr. Ashwini, the Indian government's AI model is being developed with a computing infrastructure using 18,693 GPUs. For comparison, ChatGPT was trained using about 25,000 GPUs, while DeepSeek used 2,000 GPUs.

Notably, a popular AI model like ChatGPT costs around $3 per hour to use, but India's AI model can cost as little as Rs 100 (around $1.15) thanks to government subsidies.

Samuel Daniel

Update 06 February 2025