Reforms 3.0 — Towards Bharat Rate of Growth 

Syllabus: GS3/ Economy

Context:

• Increasing discussion about India’s AI strategy, together with the launch of IndiaAI Mission and talk of sovereign AI capabilities has renewed focus on treating access to AI technology like digital infrastructure such as Aadhaar and UPI.

From the Hindu Rate to the Bharat Rate of Growth

• India had low economic growth rates averaging at around 3% commonly referred to as the ‘Hindu rate of growth’ over a period of almost four decades after Independence.
• The balance of payments crisis in 1991 ushered in economic reforms which changed the growth path for India and unleashed entrepreneurship and productivity
• In today’s context, AI provides a turning point and can serve as the new driver of rapid growth for India with the potential of achieving sustained growth rates of 8% and more.

India’s Demonstrated Potential For Leapfrogging Technologies

• India has had a consistent record of leapfrogging technologies by using Digital Public Infrastructure (DPI) as follows:
  • Aadhaar: The biggest biometric ID platform in the world having 1.3 billion enrolments. It has helped achieve targeted delivery of welfare and financial inclusion.
  • Unified Payment Interface (UPI): It handles 50% of all real time digital transactions in the world and sets a global standard for digital payments.
  • Affordable Data Revolution: India emerged as one of the biggest consumers of mobile internet usage after the entry of Reliance Jio that made huge reduction in costs.
• These examples demonstrate that creation of an ecosystem can result in large scale adoption of digital services. Similarly, this can be done for AI.

Why India Needs Free or Affordable AI Tokens?

• AI tokens are the basic units through which users interact with generative AI systems. 
• Universal access to such capabilities can significantly enhance productivity in education, research and governance.

Need for Public Investment

• India spends merely about 0.65% of GDP on Research and Development (R&D), substantially lower than China (~2.4%), USA (~3.5%), South Korea (~4.9%), and Israel (~5.4%).
  • Limited R&D investment constrains innovation and technological competitiveness.
• Providing subsidised or free AI access to top national research institutions, universities, and selected schools can create an ecosystem where AI acts as a ‘cognitive teammate’ for students, scientists and innovators.
• The estimated expenditure of around 0.06% of GDP is relatively modest compared to existing subsidies on food, fertilisers and energy.
• Other challenges that India is facing are digital divide and unequal access, data privacy and ethical concerns, high energy requirements of AI infrastructure, and risks of misinformation and algorithmic bias.

Funding Models

• PPP Model: Government can collaborate with hyperscalers like Amazon Web Services (AWS), Google Cloud, and Microsoft Azure.
  • Instead of offering them land assistance, data centres, and regulatory clarity in exchange, they might offer concessional inference capacity.
• Cross-Subsidy Approach: Enterprise subscriptions can help pay for free access by education/research institutes.
• The government’s role in this process should be just to make the appropriate regulations rather than incur all expenses, as it happened in the telecom sector.

Building Sovereign AI Infrastructure

• Just consumption of foreign APIs can put a country in technological dependency.
  • India needs to build its capability for hosting and operating LLMs.
• Efforts made by IndiaAI Mission and indigenous projects like Sarvam AI fall into this category.

Advantages of Open-Source/Sovereign Models

• Strategic Independence: Helps decrease dependence on foreign suppliers imposing restrictions.
• Cost Effectiveness: Removes any recurring licensing fees and makes it possible to implement large scale educational deployment.
• Indic Language Adaptation: Enables creation of AI across all 22 scheduled languages.
• Auditability: Open weights make it possible in case of governmental applications.
  • But national level adoption will require knowledge in cybersecurity, data residency, low latency computation, multi-region redundancy and energy efficiency.
• That is why AI infrastructure must be seen as a national asset, same as the Indian space/nuclear programme.

Diversification of Computing Infrastructure

• Excessive dependence on a single provider, especially NVIDIA, poses a risk both financially and strategically.
• Such diversified model could consist of:
  • 40%: AWS Trainium and AMD hardware for affordable inference;
  • 30%: Google TPUs for academic purposes and training;
  • 30%: NVIDIA infrastructure for more complex workloads and compatibility.
• It helps lower prices, avoids vendor lock-ins, and increases technological robustness.

Way Forward: Plan for the National AI Token Policy

• Phase I: Launching the National AI Token Policy, setting up sovereign computer partnerships, and providing unrestricted access for research to premier institutes (IIT/IISc).
• Phase II: Extending access to universities and startups, implementing AI sandboxes for innovating and introducing AI literacy programs at schools.
• Phase III: Creating sovereign Indic AI benchmarks; extending deployments in healthcare, agriculture, judiciary, and education; and deploying AI across all major Indian languages.

Conclusion

• India’s experience with Aadhaar, UPI and affordable data demonstrates that digital public infrastructure can transform society at scale.
• Extending this philosophy to AI by promoting open models, sovereign infrastructure and affordable AI tokens can create a new era of innovation-led growth.
• Decisive policy support today may enable India not merely to participate in the global AI revolution, but to lead it.

Source: TH