India Must Develop Its Own AI Model Instead of Relying on the US or China

New Delhi: At the AI India Summit 2026, experts emphasized that India needs to strengthen its IT sector by developing indigenous AI applications. While AI solutions can be built on existing platforms from the US or China, India requires a secure and tamper-proof AI framework for strategic and national security purposes. In other words, India must develop its own AI models.

Small Models vs. Large Global Models

The Economic Survey 2025–26 highlights the importance of small AI models for specialized applications. Indian startups like Sarvam have created models for Indian languages with around 2 billion parameters, whereas global giants like Google and OpenAI operate models with trillions of parameters, trained on vast datasets worldwide.

Smaller models can be trained without massive data centers, but India generates enormous amounts of data due to its large population and widespread internet access. To ensure this data remains within the country, India must invest in its own high-capacity data centers and develop large-scale AI models to maintain strategic technological autonomy.

Key Recommendations

1. Domestic Data Centers: Build centers capable of storing and processing Indian data securely. This does not mean storing all global data within India, as foreign data storage raises privacy and capacity issues.
2. Youth Training: Train Indian students and professionals in AI model development. Strong foundations in mathematics, especially linear algebra, calculus, and probability, are essential.
3. Chip Manufacturing: Encourage Indian startups to design and manufacture advanced chips, overcoming supply constraints and export restrictions imposed by countries like the US. Only high-end chips, often just 2–3 nanometers thick, meet the requirements for cutting-edge AI applications.
4. Strategic Use of Funding: Redirect subsidies currently provided to foreign companies for chip production in India toward nurturing domestic startups. A diversified ecosystem, where multiple startups work on each component, increases the chances of success.

Challenges

• Power Requirements: Large data centers require thousands of MWs of electricity. Even with renewable sources like solar and wind, supplemental thermal power may be necessary, which could impact pollution and public health.
• Chip Shortages: India currently lacks the advanced lithography machines and chip-making infrastructure required for cutting-edge AI development. Foreign controls on such equipment further complicate local production.

The summit concluded that India cannot remain dependent on the US or China indefinitely. Building a domestic AI ecosystem, supported by homegrown data centers, trained talent, and advanced chip-making capabilities, is crucial for the country’s technological sovereignty.

This roadmap mirrors strategies adopted by China and demonstrates that India can similarly establish a robust indigenous AI ecosystem.