Nifty22,552.507.8

English EditionEnglish Editionहिन्दी ગુજરાતી मराठी বাংলা ಕನ್ನಡ മലയാളം தமிழ்తెలుగు

| Today's ePaper

My Watchlist

Infosys cofounder Nandan Nilekani on AI commoditisation, data, and the rise of Indian startups

SECTIONS

ETtechLast Updated: Mar 08, 2025, 10:05:00 PM IST

Font Size

Save

Comment

Synopsis

Nandan Nilekani discussed the declining returns of building large AI models, highlighting China’s affordable, open-source approach. He emphasised the importance of high-quality data, such as AI4Bharat’s work, and the growing number of Indian startups using scalable technologies.

India need not spend resources on another large language model, says Nandan Nilekani

Infosys cofounder Nandan Nilekani on Saturday said that people have been trying to build frontier models and spending billions of dollars building the latest models which are more gigantic, with more tokens and more parameters, but clearly that has declining returns.

“They keep doing more of that, but it is not that it is going to get any better. They are all typically kept within the four walls of a company. It is like a black box for the rest of us,” he said, speaking at the People+ai Mela in Bengaluru.

In some sense what has happened in China is interesting, he said, because not only have they shown you can build a model cheaper, faster, better with lesser chips but they have also been very open about publishing the techniques that they have used. “They have done the world a service. So that goes back to commoditisation,” he said.

“I have always believed that models are a commodity. This is proving to be true. Hundreds of models are getting built. China’s DeepSeek demystified model-building. Last week they tweeted all their open-source practices. The knowledge to build models is getting rapidly commoditised,” Nilekani said.

He said we need computing capability and, most importantly, data. If you can create high quality data, like AI4Bharat is, in Indian languages and share it with everyone, it will accelerate by years, the pace at which things get rolled out, he said, adding, “The hard part is uses. How do we use this to deliver value for people? That is where we can lead. We can always replace the model.”

The number of startups in India has increased to 150,000 from 1,000 startups in 2016, and it will keep growing, Nilekani said. “They will use mass scale technologies that we have deployed in interesting use cases. There is also going to be a lot of capital coming in,” he said.

Discover the stories of your interest

Rohini Nilekani, cofounder of EkStep Foundation, said, “All the private infrastructure in Bengaluru is fantastic. One cm outside, what do you see in the public sphere? The opposite needs to happen in AI. The private and public should not be different.”

She further said, “The quality, service, access, availability, trust and inclusion should not be different in a gated community and outside in the public sphere.”

Nandan Nilekani said that he had started supporting Chennai-based AI4Bharat three or four years ago and will continue to support it for three more years till its activities can be further ramped up.