Source: Government of India (2)
With robust and high end Common computing facility in place, India all set to launch its own safe & secure indigenous AI model at affordable cost soon: Shri Ashwini Vaishnaw
Compared to global models costing 2.5 to 3 dollars per hour usage, India’s AI Model will cost less than less than 100 rupees per hour after 40% government subsidy; Attractive half yearly & annual plans will make it more affordable
Multiple foundational models for Indian context, in Indian languages, likely to be ready later this year will help researchers, students & people at large for its low cost, fast computing & prompt results.
To begin with, 18 citizen centric applications pertaining to agriculture sector, learning disability & climate change to be part of this AI Model
DeepSeek will get hosted on Indian servers after security protocol checks so that users, coders, developers can benefit from its Open Source Code
Posted On: 30 JAN 2025 6:07PM by PIB Delhi
India is all set to launch its own safe & secure indigenous AI model at an affordable cost. Union Minister for Electronics & Information Technology, Railways, Information & Broadcasting, Shri Ashwini Vaishnaw today announced this in New Delhi at Electronics Niketan. Interacting with the media, the Union Minister said that the Indian AI model is a timely step as India is a trusted nation among the comity of nations & therefore it will help India emerge as a more reliable technological powerhouse of ethical AI solutions in the days to come. Backed by a high-end common computing facility, the India AI mission is now closer to customising indigenous AI solutions for the Indian context using Indian languages. He said that scientists, researchers, developers and coders are working on multiple foundational models in this regard & with the given pace, the Union Minister expressed hope that the Indian AI model is likely to be ready within 6 months.
The AI model is beginning with the computation facility of roughly 10000 GPUs. Soon the remaining 8693 GPUs will be added. It will largely benefit researchers, students & developers in the beginning. The Technical partners who are participating in the mission have expressed a lot of confidence in the ability of the mission to deliver its objective of democratising access to computing & that too at a very competitive rate. Government has decided to give it to the users for less than 100 rupees per GPU after subsidising the 40% of the cost. Compared to global models costing 2.5 to 3 dollars per hour usage, India’s AI Model will cost less than less than 100 rupees per hour after 40% government subsidy. The attractive half yearly & annual plans will further make it more affordable.
Within 10 months of the launch of India AI Mission, Ministry of Electronics & Information Technology, is able to get an unprecedented response & create a high end & robust common computing facility of about 18,693 Graphic Processing Unit, GPUs ready for use. It is about nine times of what Open Source Model DeepSeek has & about two third of what ChatGPT has. Answering queries of the media, the Union Minister said that DeepSeek can get hosted on Indian servers after security checks so that coders, developers & designers can take benefit of its Open Source code.
Safety and ethical deployment of AI Model remains top priority for the government. Expressing this commitment, the Union Minister announced that India is establishing an AI Safety Institute, adopting a techno-legal approach.
The Key safety-related projects in this regard include following 8 simultaneous efforts to ensure the privacy of data along with ethical auditing of algorithmic efficiency.
- Machine Unlearning (IIT Jodhpur)
- Synthetic Data Generation (IIT Roorkee)
- AI Bias Mitigation Strategy (NIT Raipur)
- Explainable AI Framework (Defence Institute of Advanced Technology, Pune & Minecraft Technologies)
- Privacy Enhancing Strategies (IIT Delhi, IIIT Delhi, IIT Dharwad & Telecom Engineering Centre, TEC)
- AI Ethical Certification Framework (Tool Nishpaksh being developed at IIIT Delhi & TEC)
- AI Algorithm Auditing Framework (Tool Parakh being developed by Civic Data Labs )
- AI Governance Testing Framework (Amrita Vidyapeetham & Telecom Engineering Centre)
In addition, Watermarking and labeling, Deep fake detection tools, Real time identification and mitigation, AI risk management are five other projects AI Mission is running to make India’s AI model safe & secure from a privacy point of view. This initiative aims to benefit researchers, students, and the public with its low-cost, fast computing capabilities and its prompt results. The Indian AI Model will help promote innovation & develop citizen centric better governance tools including many industrial uses having humongous potential to harness technological benefits for the people at large.
Common Compute Facility is the strong foundation for Democratic AI Development
Under the India AI Mission, a huge computing infrastructure has been developed, surpassing global benchmarks in a short time. The facility now houses 18,693 GPUs, including 12,896 H100, 1,480 H200, and 7,200 MI 200 300 units, significantly exceeding the initial target of 10,000 GPUs. To put this capacity into perspective, DeepSeek was trained on 2,000 GPUs, while ChatGPT required 25,000 GPUs. This vast computing power will not only accelerate research, model training, help ethical AI algorithm development and foster innovation in India’s AI ecosystem.
A common compute platform has been created, ensuring accessibility to all stakeholders. Approximately 10,000 GPUs are already available, and technical partners have expressed confidence in the mission’s capability to deliver world-class AI solutions. Following approval, this facility will soon be operational for widespread use.
India’s Own AI Model: Built for Local Context
Over the past one & half year, India has been developing a robust AI ecosystem framework to support its own foundational AI model. This model will address Indian linguistic and contextual needs while eliminating biases, ensuring inclusivity and promoting fairness. Leading developers and researchers are working towards completing multiple foundational models within 8 to 10 months, leveraging algorithmic efficiency to achieve cost-effective and timely development.
Experts believe that India’s AI model will meet the country’s diverse requirements, bringing forth a high level of innovation tailored for Indian users.
AI Applications to Benefit Citizens
The India AI Mission focuses on developing AI applications in critical sectors such as agriculture, healthcare, weather forecasting, and disaster management. Eighteen applications have been identified in these domains to harness AI for societal benefits. The initiative will address challenges such as climate change, learning disabilities, and agritech solutions, ensuring AI contributes to the well-being of millions.
Affordable Compute Facility for AI Development
India’s Compute facility is being offered at highly competitive rates. The cost per GPU hour is approximately ₹115.85, significantly lower than the global benchmark of $2.5–$3 per hour. High-end computing will be available at ₹150 per hour, with a 40% government subsidy reducing the cost to less than ₹100 per hour for common Compute access. This affordability ensures democratized AI access, empowering startups and researchers alike.
The initiative offers competitive six-month and annual compute rate packages. The facility has garnered international appreciation, including recognition at Davos, reinforcing India’s position as a trusted global AI hub.
Future Roadmap and Sustainability
The India AI Mission operates under a four-year sunset clause with long-term sustainability. As India advances in its semiconductor journey, the government is strategically developing its ecosystem with clarity and systematic planning. With over $30 billion invested in the semiconductor mission, India’s AI aspirations align with its broader technological vision.
The Union Minister said that DeepSeek and other foundational models can be hosted on Indian servers, similar to previous initiatives like LLaMA. He added, AI’s real impact lies in industrial applications beyond chatbots and image generation, It will address real-world challenges such as:
- Health of oil drilling rigs
- Railway ticketing optimization
- Soil health monitoring for agriculture
- Weather and cyclone prediction
The mission also emphasizes AI safety through real-time detection tools, deep-fake mitigation, and robust AI risk management strategies. Stanford ranks India among the top nations in AI education, with 240 universities offering AI courses and 100 universities equipped with 5G labs.
With a focus on democratic, inclusivity, affordability, and innovation, India is set to emerge as a global AI powerhouse, shaping the future of artificial intelligence for societal and industrial advancements.
*****
Dharmendra Tewari/Shatrunjay Kumar
(Release ID: 2097659) Visitor Counter : 81