India to Develop Local Language AI Foundation Model Amid Global AI Race

India's AI Ambitions: Developing a Local Language Foundation Model

In a significant move to bolster its position in the global artificial intelligence (AI) race, India is set to back an industry-academia consortium to develop a local language AI foundation model. The Indian government will soon issue a Request for Proposal (RFP) to advance this initiative, placing India among select nations like the US and China that are building prominent AI models 1

1

Driving Factors and Government Strategy

The decision comes in the wake of recent US government restrictions on AI compute infrastructure exports and limitations on the use of language models dominated by American tech firms. A senior official stated that the debate over whether India should have its own foundational model or work on open-source models has been resolved: "We need to do both" 1

1

To facilitate this endeavor, the government plans to form a technical committee that will engage with startups and academic institutions to assess the technical and financial support required. The initiative aims to create a consortium-based approach, with the resulting model based on Indian languages and datasets to be open-sourced for broader industry use 2

2

Global Context and Competitive Landscape

The global AI landscape is rapidly evolving, with significant developments in the US and China:

1. The outgoing Biden administration recently imposed restrictions on GPU imports and AI model usage for countries like India.
2. The new US president, Donald Trump, has backed a $500 billion AI investment initiative led by OpenAI.
3. In China, the AI startup DeepSeek launched an open-source AI reasoning model that reportedly outperforms some American counterparts at a lower cost 1

   1

   .

Related Stories

India's AI Ambitions: Government Receives 67 Proposals for Indigenous Foundational Models

India's AI Ambitions: Balancing Innovation, Accessibility, and Cultural Relevance

India's AI Race: Government Launches IndiaAI Mission to Develop Sovereign Foundation Models

Sarvam AI and Others Selected to Build India's Indigenous AI Models Under IndiaAI Mission

Industry Perspectives and Challenges

Srikanth Velamakanni, co-founder of Fractal.ai, emphasized the importance of India developing its own foundational models, not only to avoid being left behind but also to advance Artificial General Intelligence (AGI) 2

2

However, challenges remain:

1. The cost of training and housing such models is substantial, often running into millions of dollars.
2. High-powered compute resources, particularly GPUs, are expensive and scarce.

Call to Action from Tech Leaders

Indian-origin technology leaders are rallying support for this initiative. Aravind Srinivas, co-founder of AI startup Perplexity, urged India to build its own globally competitive models rather than relying on open-source options. He offered a personal investment of $1 million and his time to support qualified teams in this endeavor, with a potential increase to $10 million if the resulting model outperforms DeepSeek R1 on all benchmarks 1

1

2