The Indian government has received 67 proposals under the Rs 10,370-crore IndiaAI Mission to develop an indigenous AI foundational model. As part of the IndiaAI Mission, the government gets 67 proposals for domestic AI foundational models, including 20 focused on large language models. Of these, 20 proposals focus on large language models (LLMs), while the remaining submissions cater to specific sectors, including healthcare.
The Union Ministry of Electronics and IT has initiated the process of forming a high-level technical committee to assess these proposals. The panel will include external experts who will evaluate the feasibility of each submission within a month. The ministry had invited proposals in January from India-based entities and researchers to create a domestic foundational model. Companies such as Sarvam AI, CoRover.ai, and Ola are among the key contributors in the LLM category.
China’s DeepSeek Sparks Urgency
India’s move comes in response to the rise of DeepSeek, a cost-effective AI model from China that has gained global attention. DeepSeek’s open-source nature, accuracy, and affordability compared to US competitors have disrupted the AI sector. Its emergence has also revived discussions on stricter export controls and India’s position in the AI race.
Electronics and IT Minister Ashwini Vaishnaw has expressed confidence in India’s ability to produce a world-class AI model within the next 9 to 10 months. He emphasized that the models will reflect India’s unique challenges and opportunities while ensuring fairness, bias mitigation, and adherence to ethical AI principles.
Funding Strategy and Intellectual Property Rights
The government plans to support multiple AI models based on scalability and potential impact. The intellectual property rights of these models will remain with the respective developers, while the government will secure a perpetual license for public use.
A two-tier funding approach has been outlined. The first involves direct financial assistance, with disbursements linked to project milestones. The second option includes equity-based funding to ensure sustainable growth of AI models.
Boosting Compute Power with GPUs
To facilitate AI development, the government has finalized 10 companies to supply over 18,693 graphics processing units (GPUs). These high-end chips will be crucial for training machine learning models.
The selected firms include Jio Platforms, Tata Communications, Yotta (Hiranandani Group), E2E Networks, NxtGen Datacenter, CMS Computers, Ctrls Datacenters, Locuz Enterprise Solutions, Orient Technologies, and Vensysco Technologies. This is a significant increase from the initial goal of procuring 10,000 GPUs under the IndiaAI Mission.
A centralized compute facility will soon be launched, allowing startups and researchers access to these GPUs. High-end GPU usage will cost Rs 150 per hour, while lower-end GPUs will be available at Rs 115.85 per hour. To make AI development more accessible, the government will provide a 40% subsidy on total costs for end users.
To enhance AI infrastructure, the Indian government gets 67 proposals for domestic AI foundational models, with a focus on India-specific challenges. The IndiaAI Mission aims to establish the country as a major player in AI. The focus is on creating models that cater to Indian contexts and ensuring local AI advancements keep pace with global trends. The upcoming months will determine how these AI initiatives shape the nation’s technological future.
As the AI race intensifies, the Indian government gets 67 proposals for domestic AI foundational models, aiming for self-reliance in AI. The IndiaAI Mission is a significant step towards making India a global leader in AI. By fostering indigenous AI development, the initiative aims to reduce dependence on foreign technology and address India-specific challenges.
