Nandan Nilekani reiterated his stance that India should not allocate resources towards developing another large language model (LLM). Responding to comments made by Google Research India’s Director Manish Gupta, who suggested that India could benefit from creating a foundation model, Nilekani emphasized that investing in LLMs may not be the most efficient use of funds.

“If India has $50 billion to invest, it should prioritize building computational infrastructure, AI cloud systems, and other foundational resources,” Nilekani said. He argued that foundation models, like those developed by OpenAI and Meta, require enormous financial investments due to the vast data sets and high-end infrastructure needed for training.

Nilekani has previously stressed that India’s focus should be on developing AI use cases that can be built on existing LLMs available globally. In contrast, Gupta, during his speech at the Bengaluru Tech Summit, disagreed with Nilekani’s viewpoint. Gupta highlighted that while Nilekani’s approach with Aadhaar began with foundational infrastructure, India should similarly focus on establishing solid foundations for AI innovation, leveraging its unique constraints as advantages.