Can India be an AI winner?

The Indian government’s ambitions stretch further. Narendra Modi, the prime minister, wants the country to lead in AI and insists that any global progress in the field is incomplete without India. In February Ashwini Vaishnaw, the minister in charge of technology, pledged that India would release an indigenous foundational model within ten months and develop domestic AI chips in three to five years.

So far progress is modest. Sarvam’s system is built atop technology from Mistral, a French firm. Bharat Gen is based on other open-source variants. No Indian model features among the top 200 LLMs ranked by LMArena, a benchmarking website. According to the AI Index compiled by researchers at Stanford University, in 2024 India attracted just $1.2bn in private AI investment, placing it 12th globally, behind Austria and Sweden. In America the figure was $109bn and China $9bn.

India’s digital infrastructure also lags woefully: its datacentres account for just 3% of global capacity. While America and China build foundational platforms from scratch, India remains an adapter, repurposing foreign models for domestic needs. Whether this will be a transitional phase or a lasting weakness is an open question.

Policymakers are now playing catch-up. Last year the government announced a $1.2bn programme to support local AI efforts. Some of the funds have gone towards buying more than 34,000 chips, which are being made available to researchers and startups. (Meta, an American social-media giant, owns ten times as many.) Sarvam will have access to 4,000 chips for six months. Abhishek Singh, who leads the IndiaAI Mission, the agency overseeing the project, says the goal is to kick-start a domestic ecosystem. The agency has made AI processors available at roughly a third of the cost globally. Four companies, including Sarvam, have also been selected to develop a fully homegrown AI model.

Booting up

Not everyone is convinced. Anirudh Suri of India Internet Fund, a venture-capital firm, argues that policymakers are too fixated on chips and code. Rather than trying to pick winners, he says, they should focus on building the infrastructure that allows AI development to thrive.

One missing piece is data. India has one of the world’s largest bases of smartphone users, internet subscribers and digital transactions. Its collection of public-facing digital platforms, such as a national biometric ID system and real-time payments, generates troves of data. But much of it remains siloed and inaccessible to researchers and developers. Although the IndiaAI Mission has begun releasing datasets, progress has been sluggish. Without open, high-quality data, developers in India will struggle to break new ground.

Talent is another problem. The country trains large numbers of coders, but few specialists in cutting-edge AI research. Of those, many head abroad, mostly to America. A study in 2022 by MacroPolo, an American think-tank, found that only one-fifth of India’s top AI researchers remain in the country. Harshit Joshi, an AI researcher of Indian origin at Stanford University, says the appeal of working abroad lies in easier access to resources and the chance to collaborate with the world’s best minds. Had he stayed in India, he says, “the fear of missing out" would have loomed large.

Those who remain often work at the Indian outposts of American tech giants such as Google and Microsoft, which operate some of the best-funded AI labs in the country. Yet their presence has not catalysed a thriving local ecosystem. One reason, suggests Kailash Nadh, chief technology officer at Zerodha, a fintech firm, is that India lacks a culture of patient, long-term research. “There is a lack of cohesion between industry and academia," he says. As a result, original research remains sporadic and siloed.

Electric sheep

India’s AI push has also leaned heavily on the domestic market. Policymakers often speak of making AI “work for India". That seems sensible. Many of the country’s challenges—from health care to education—could benefit from AI. But a narrow domestic focus may also limit ambition. To truly seize the moment, argues Mr Suri, India needs a “big bet" mindset. AI, he argues, is more than just a tool; it is also a powerful engine for economic growth.

India has heard this story before. Previous waves of digital change, from the internet to the smartphone, were dominated by American tech firms. India is one of the largest users of products from Google, Microsoft and Meta. These shifts did give rise to some local champions, mostly in e-commerce and fintech. But few of these companies have expanded abroad. For a country that has ambitions of moving up the innovation ladder, the opportunity offered by AI is too big to be missed.

Correction (June 17th 2025): This article has been updated to correctly identify Harshit Joshi, an AI researcher at Stanford University.

Stay on top of our India coverage by signing up to Essential India, our free weekly newsletter.