TL;DR
- A debate is intensifying in India about AI development, with Aravind Srinivas advocating for indigenous large language models, akin to ISRO's space achievements.
- Nandan Nilekani counters, suggesting the focus should be on enhancing AI infrastructure rather than creating new models.
- Srinivas believes that India has the potential to innovate at lower costs, reflecting on developments like China's DeepSeek.
- The discussion highlights contrasting visions for India's role in the global AI landscape, underscoring the importance of both model development and application.
The debate over India’s approach to the rapid advancements in the artificial intelligence space is heating up, with key figures like Aravind Srinivas and Nandan Nilekani sharing opposing views.
Srinivas, reflecting on his experience as CEO of Perplexity, challenges India’s hesitance to develop its own models, arguing that the nation should demonstrate its capability in AI akin to ISRO’s achievements in space.
Highlighting India’s potential, he notes that recent developments like the Chinese LLM DeepSeek exemplify that innovation doesn’t always have to come with hefty investments.
“Elon Musk appreciated ISRO (not even Blue Origin) because he respects when people can get stuff done by not spending a lot,” Srinivas, whose company has earned the backing of the likes of Amazon founder Jeff Bezos, noted.
“I think that’s possible for AI,” he said, adding a wish that India seeks to compete on a global front by developing LLMs that can match benchmarks of top models.
Srinivas said he is “not in a position to run a DeepSeek-like company” for India but he is willing to help “anyone obsessed enough to do it and open-source the models.”
Re India training its foundation models debate: I feel like India fell into the same trap I did while running Perplexity. Thinking models are going to cost a shit ton of money to train. But India must show the world that it's capable of ISRO-like feet for AI. Elon Musk…
— Aravind Srinivas (@AravSrinivas) January 21, 2025
Srinivas’s comments came in reponse to Infosys co-founder Nandan Nilekani, a pivotal figure in India’s tech infrastructure, who cautioned against investing heavily in building India’s own large language models (LLMs).
Instead, Nilekani suggested focusing resources on AI cloud and computing infrastructure. His stance, shared last month, emphasized utilizing existing LLMs and channels efforts into creating applications for real-world scenarios.
To be clear: Nandan Nilekhani is awesome, and he's done far more for India than any of us can imagine through Infosys, UPI, etc. But he's wrong on pushing Indians to ignore model training skills and just focus on building on top of existing models. Essential to do both.
— Aravind Srinivas (@AravSrinivas) January 21, 2025
TCS CEO Krithivasan has supported Nilekani’s perspective, citing Indian companies’ history as system integrators. According to Krithivasan, leveraging available LLMs for regionalization, especially for Indian languages, can generate new market opportunities without the extensive costs of developing entirely new models.
Srinivas had met PM Modi in New Delhi late last month after earlier offering to chalk things out to offer free Perplexity Pro for researchers, students and faculty in India.
Planning your financial journey can be daunting but it doesn't have to be. Fire Fast by Dzambhala helps you understand and plan effectively.
Join the vibrant privacy-ensured Dzambhala community on 
Want to give feedback on this story? Write to us at [email protected]