The country will soon be developing its own AI model like ChatGPT and DeepSeek. It has announced this decision by Union IT Minister Ashwini Vaishnaw, within the next 10 months. It is supported under the IndiaAI mission, which has a backing of ₹10,370 crore funding. The key objective of this AI mission is to augment the computing power required to run AI software, which also encompasses providing high-performance chips, called GPUs, necessary for AI processing. The government has shortlisted 10 companies that will supply these chips; among them, there are 18,693 graphics processing units (GPUs). These include Jio Platforms, Tata Communications, and E2E Networks, among others. A large portion of these chips will be supplied by YOTT, a company supported by the Hitachi Group. The government will provide access to this infrastructure to support startups and
researchers who need high computing capacity. A portal for accessing these resources will be ready in just two days, and the rates for using this hardware will range from ₹115 to ₹150 per hour, with a 40% subsidy from the government. IndiaAI CEO Abhishek Singh said that the new AI models would reflect India's languages and culture, incorporating local datasets into their development.
About DeepSeek, the Chinese AI model, Ashwini Vaishnaw appreciated how the AI model built itself at a modest cost. He referred to how it has been accused that the budget of the IndiaAI mission is too low but how DeepSeek could develop such a strong AI system at only $550,000 and had all its strengths because of the brain behind it.
Even an American cybersecurity company, Veeam, released a statement showing that DeepSeek had a leakage of some its data. What was leaked comprises chat logs featuring user recommendations-which can go online-but that was corrected one hour after learning about it; more on details at indiaai.gov.in.