
Alliance Electronics and Information Technology Minister Ashwini Vaishnaw announced on Thursday that India will build a locally generated artificial intelligence (AI) model within six to eight months. The plan was announced at Utkarsh Odisha Cendave, and the minister also revealed that the DeepSeek-R1 AI model has been implemented on local Indian servers and experts are currently evaluating the model’s technical reports. It is worth noting that the announcement was just one week after the end of the 55th annual meeting of the World Economic Forum held in Davos-Klosters, Switzerland, and India was also a participant.
India prepares to build local AI models
According to a report by MoneyControl, Vaishnav made several AI-focused announcements at the conference while talking to the press. The IT minister said India has acquired more than 18,600 high-end GPUs to build AI infrastructure and scale computing. This will become the backbone of the generative AI model to be developed this year.
“We believe that at least six major developers can develop AI models within six to eight months of external restrictions, while in more optimistic estimates, four to six months,” Vaishnaw quoted.
The union minister also stressed the urgent need to develop a common computing facility as it is the most important pillar of building a strong AI ecosystem. The program will be part of the Indian AI mission and will support computing requirements for researchers, startups and academic institutions.
Vaishnaw detailed the planned computing facility, a cloud-based server designed to run and process AI inference, which he said has exceeded the country’s initial expectations, with the country gaining more than 18,600 GPUs. These include 12,896 NVIDIA H100 GPUs, 1,480 NVIDIA H200 GPUs, and 742 AMD MI325 and MI325X GPUs. The IT minister said the initial goal was to buy 10,000 high-end AI chipsets.
“DeepSeek AI was trained at 2,000 GPUs, Chatgpt was trained at 25,000 GPUs, and now we have 18,000 high-end GPUs. India now has a powerful computing facility that can support our AI ambitions,” he added. It is worth noting that once deployed, the Indian AI model will have access to everyone. The Minister did not specify whether the model is open source or is available through the platform and application programming interface (API).
Vaishnaw said on DeepSeek that India has implemented the DeepSeek-R1 AI model and that experts are currently evaluating technical reports to better understand IT, such as CNN-NEWS18 video recording.
Vaishnaw focuses on AI security and security, and India will build an AI security agency with wheel hub and spoke models. Unlike countries where a single agency oversees AI regulations, India will allow multiple agencies to collaborate and establish security tools and frameworks.