India is set to create its own generative AI model, joining global players like OpenAI’s ChatGPT and China’s DeepSeek. The Union IT Minister, Ashwini Vaishnaw, announced this plan at the Utkarsh Odisha Conclave. The new AI model will be developed using the India AI Compute Facility, which has secured nearly 19,000 powerful GPUs (computer processors). These include 12,896 Nvidia H100 models and 1,480 Nvidia H200 models, designed for high-performance AI tasks.
The facility initially aimed for 10,000 GPUs but has surpassed that goal with almost 19,000. Around 10,000 GPUs are already operational, with the remaining being rolled out gradually. The AI model will be specifically built for India, addressing the country’s unique linguistic and cultural diversity and aiming to reduce biases in AI datasets.
Vaishnaw announced plans to build AI data centers in Odisha as part of India’s efforts to strengthen its AI ecosystem and digital infrastructure. The government has been investing heavily in AI research and infrastructure to reduce dependence on foreign AI models and create a self-reliant digital ecosystem. A key component of this initiative is the India AI Compute Facility, which will provide powerful computing resources to support the country’s growing AI projects.
Ashwini Vaishnaw, has emphasized the country’s commitment to developing its own world-class AI model. He announced that India is calling for proposals to create a foundational AI model that can compete globally. In the first round of funding, 18 projects have been selected to develop AI models focused on key areas such as agriculture, healthcare, disaster management, and learning disabilities. Vaishnaw also highlighted the importance of a techno-legal approach to establish AI safety institutions, ensuring responsible and secure development of AI in India.
Vaishnaw mentioned that developers would begin creating the AI models soon, with an expected development timeline of four to eight months. Some developers are optimistic they can finish their work in just four to six months. The government plans to invite proposals from developers to start the AI model creation process immediately.
Union Minister Ashwini Vaishnaw announced that India now has 15,000 high-performance GPUs, with 10,000 of them currently available for creating AI models, algorithms, and other processes. To put it into perspective, DeepSeek was trained using 2,000 GPUs.
In addition to building the AI model, India’s AI Mission focuses on using AI to solve major societal challenges. The government has selected 18 AI-driven projects for the first round of funding, with a focus on three key areas: agriculture, climate change, and learning disabilities. This initiative highlights India’s commitment to using AI for the greater good.
India is also prioritizing responsible AI development. To ensure safe and ethical use of AI, the government will set up an AI safety institution. Unlike many countries that rely on a single regulatory body, India will implement a hub-and-spoke model, enabling multiple institutions to collaborate and provide the necessary safety frameworks for AI technology.
Through this ambitious initiative, India aims to become a major player in the AI landscape, ensuring that its AI systems are both powerful and aligned with its societal needs.
Union IT Minister Ashwini Vaishnaw announced on January 30 that India will soon host Deep Seek, an open-source model similar to LLama, on Indian servers. This move aims to address data privacy concerns, as hosting such models locally can ensure greater control over data. The initiative will be supported by the India AI Compute Facility, which has secured 18,000 GPUs to develop a Large Language Model (LLM) designed specifically for India’s needs.
Vaishnaw highlighted that the facility will be affordable, costing less than $1 per unit, with the government covering the expenses for the next four years. He emphasized the importance of algorithmic efficiency and high-quality datasets in unlocking the real potential of AI models. The project will be powered by key partners like Jio Platforms, CtrlS Datacenters, Locuz Enterprise Solutions, E2E Networks, and NxtGEn DataCenter, who will supply the necessary GPUs. The government is also collaborating with tech partners to ensure the data center is efficiently managed and accessible to all.
Ashwini Vaishnaw stated that India’s foundational AI model will rival the best globally under the India AI Mission.
What is the India AI Mission?
The “India AI Mission” is an initiative aimed at enhancing AI infrastructure in India and supporting research, startups, and businesses. The mission has made remarkable progress, having already surpassed its initial target of acquiring 10,000 GPUs by achieving 18,693 GPUs. This includes 1,480 NVIDIA H200 GPUs, 12,896 NVIDIA H100 GPUs, and 742 MI325 and MI325X GPUs, which are considered some of the most powerful AI chips available today.
For comparison, China’s AI startup DeepSeek was trained on 2,000 GPUs, while ChatGPT was trained on 25,000 GPUs. This highlights the scale of the India AI Mission. Companies such as Jio Platforms, Tata Communications, Yotta, and Nextgen Data Centers are involved in providing these GPUs. Approximately 15,000 high-end GPUs have been purchased, with around 10,000 ready for immediate use, and the remaining ones will be added gradually.
The goal of the mission is to provide resources to researchers, startups, and businesses across India. A common AI computing facility will be operational within two days, serving as a vital resource for AI projects in India.
A key feature of the mission is that it will provide the world’s most affordable AI computing facility. The average rate for AI computing has been set at ₹115.85 per GPU hour, while the global benchmark is $2.5-$3 per GPU hour. Students and academics will receive a 40% subsidy, bringing the cost to less than ₹100 per GPU hour.
Stay informed with our financial updates, stocks, bonds, commodities. Get global & political insights. Follow us & enable notifications for the latest updates.