Archive · 17 Aug 2025
India's Dual Approach to Becoming an AI Powerhouse
India is working hard to create its own artificial intelligence.
It's doing this in two ways: using existing AI models and also building brand new ones from scratch.
They're using current AI to solve problems right now, like helping farmers and students.
At the same time, they're building their own AI that understands Indian languages and culture better.
This approach helps India solve today's problems and have more control over its AI in the future.
It's like building a car using parts you have, while also working on making your own from scratch.
India is developing AI with a dual-track strategy: utilizing open-source models while building sovereign ones.
Startups are fine-tuning open-source models like Gemma to address immediate needs in sectors such as agriculture and education.
Simultaneously, initiatives like Project EKA are creating sovereign LLMs trained on Indian data and infrastructure.
The lack of high-quality training data in Indian languages is a key challenge for sovereign AI development.
India's approach offers a potential model for other nations in the Global South facing similar constraints.
- Who
- Indian government, AI startups, and tech experts.
- What
- India is pursuing a dual strategy to develop AI: utilizing open-source models and developing sovereign ones.
- When
- The initiatives are ongoing.
- Where
- India
- Why
- To address immediate needs while building long-term AI independence.
Reliance on Foreign Models
Building Sovereign AI
Dependencies
Reliance on Foreign Models
Over-reliance on foreign AI platforms can lead to dependencies and control issues.
Building Sovereign AI
Fine-tuning existing open source models enables faster deployment and practical solutions while indigenous models are developed.
Contextual Understanding
Reliance on Foreign Models
Foreign models may lack the ability to understand and process the nuances of local cultures and contexts.
Building Sovereign AI
Open source models allow access and innovation in data sets and languages.
- IndiaAI Mission
- ₹10,300 crore initiative
- Project EKA
- Sovereign LLM effort
- Gemma
- Google's open-source LLM
- Key Domains for EKA
- Agriculture, law, education and defence
- Digital languages with no datasets
- 72
