The Indian government has received 67 proposals to develop indigenous artificial intelligence (AI) foundational models under its Rs 10,370-crore IndiaAI Mission, aimed at creating solutions suited to the country’s linguistic and cultural landscape.
Among these proposals, 20 are for large language models (LLMs) from entities such as Sarvam AI, CoRover.ai, and Ola, according to a senior government official familiar with the matter.
Focus on local context
The IndiaAI Mission, announced earlier this year, seeks to strengthen India’s position as a global leader in AI by encouraging collaboration among government agencies, academic institutions, and the private sector. The goal is to build AI models that tackle challenges unique to India, ranging from agriculture and healthcare to public service delivery.
According to the reports, the indigenous development will ensure better data security and privacy, while also making AI more accessible and relevant for India’s diverse population.
Evaluating proposals
The Union Ministry of Electronics and Information Technology has formed a high-level technical committee to review the feasibility of the submitted proposals, and a decision is expected within a month.
Some of the proposals that do not focus on LLMs are sector-specific or smaller language models, including one intended to assist medical professionals in diagnosing and treating breast cancer.
Electronics and IT Minister Ashwini Vaishnaw has expressed confidence that India can produce a “world-class” AI model within the next nine to 10 months. He has stressed the need for diverse models that align with Indian requirements, address local challenges, and adhere to ethical AI principles.
The government is prepared to provide both direct financial support and equity-based funding to developers, depending on each model’s scalability and viability.
Impact of global competition
India’s push for AI self-reliance comes at a time when global competition in the field has intensified. The emergence of DeepSeek in China, an open-source foundational model built at relatively lower costs, has generated fresh debate about the need for affordable and locally developed AI solutions.
DeepSeek’s entry has also rekindled concerns about export controls, cost-effective development, and the availability of high-end hardware necessary for training AI models.
GPU procurement and subsidies
To support these domestic AI efforts, the government has selected 10 companies to provide 18,693 graphics processing units (GPUs), which are critical for machine learning and AI research. This is an increase from the initial plan to procure around 10,000 GPUs for the IndiaAI Mission.
Jio Platforms, Yotta, Tata Communications, E2E Networks, NxtGen Datacenter, CMS Computers, Ctrls Datacenters, Locuz Enterprise Solutions, Orient Technologies, and Vensysco Technologies are among the companies that will supply these GPUs.
A common computing facility is set to launch soon, giving startups and researchers access to these GPUs at a cost of Rs 150 per hour for higher-end chips and Rs 115.85 per hour for lower-end ones.
The government will offer a 40% subsidy, cutting the effective cost to around $1 per hour from the global average of $2.5 to $3 per hour. Officials hope this reduction in costs will boost innovation and support more research projects, making India a competitive hub for AI development.
Related:- 25 Great Business Ideas for Entrepreneurs 2025
Related:- How To Start A Business When You Have No Money
Related:- Difference Between Cash and Profit 2024
Follow Startupcolleges on Facebook, Instagram, and Twitter, for the latest updates from the startup ecosystem. Subscribe to our YouTube Channel for more entrepreneurship, business, and corporate solution videos.
