Indian-origin CEO says Nandan Nilekani is 'awesome' but also 'wrong' for his stance on AI

The comments from Aravind Srinivas came in response to Nandan Nilekani’s earlier advice to Indian AI startups.

Aravind Srinivas, the Indian-origin CEO of Perplexity AI, publicly criticised Infosys co-founder Nandan Nilekani's stance on artificial intelligence (AI), asserting that India should prioritise both AI model training and practical applications.

Srinivas expressed his views in a post on X (formerly Twitter), calling Nilekani "awesome" for his unmatched contributions to India's technological advancements through Infosys and UPI. However, he challenged Nilekani’s recommendation for Indian AI startups to focus solely on building practical AI applications rather than training large language models (LLMs).

“To be clear: Nandan Nilekani 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,” Srinivas wrote.

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

The comments from Srinivas came in response to Nilekani’s earlier advice to Indian AI startups. At the Meta AI Summit in October, Nilekani had urged startups to avoid the costly endeavour of building large AI models, suggesting that resources should instead be directed towards creating practical AI solutions.

“Our goal should not be to build one more LLM. Let the big boys in the (Silicon) Valley do it, spending billions of dollars. We will use it to create synthetic data, build small language models quickly, and train them using appropriate data,” Nilekani had said. He further emphasised the need for scalable, frugal infrastructure and practical applications tailored to the Indian context.

Srinivas, however, disagreed, calling for a more ambitious approach to AI development in India. Drawing from his own experience, he argued that India is making a mistake by shying away from training AI models, assuming that the financial costs are prohibitive.

“I feel like India fell into the same trap I did while running Perplexity. Thinking models are going to cost a ton of money to train,” Srinivas said. He urged Indian startups to shift from relying on open-source models and focus on developing their own AI capabilities to be globally competitive across all benchmarks, including Indic languages.

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 MuskAravind Srinivas (@AravSrinivas) January 21, 2025

Srinivas likened India’s potential in AI to its achievements in space exploration, citing the Indian Space Research Organisation (ISRO) as a model for overcoming resource constraints to achieve global acclaim. He highlighted SpaceX CEO Elon Musk’s appreciation for ISRO’s cost-efficient yet successful missions.

“Elon Musk appreciated ISRO (not even Blue Origin) because he respects when people can get stuff done by not spending a lot. That’s how he operates,” Srinivas wrote, encouraging India to adopt a similar mindset in AI development.

Srinivas concluded by expressing hope for a paradigm shift in India’s AI strategy. He emphasised the importance of building expertise in training foundational AI models that could compete globally while also addressing local needs. He offered his support to anyone “obsessed enough” to take up this challenge, stating his willingness to help open-source AI models for broader benefit.