India's Strategic Push to Catch Up in the Global AI Race

India bets on 'frugal innovation' to catch up in global AI race

India is betting on the tradition of "frugal innovation" and its huge tech talent pool to catch up in the global artificial intelligence arms race, as it seeks a share of the fast-developing industry. Prime Minister Narendra Modi's government, start-up founders and policymakers believe the world's most populous country can be competitive in AI by creating cheaper large language models trained on Indian languages and by building AI "applications" to solve specific problems. "If DeepSeek has been built at a reasonably low cost, we should be able to as well", said Abhishek Singh, the bureaucrat picked by Modi to lead the government's AI mission in 2024. Singh is sifting through 67 bids from tech start-ups and research labs seeking funding for domestic AI models. Those approaches came after the government issued a rallying call for ideas in January in a push to create a homegrown rival to China's DeepSeek, which claims to have built a competitive model at a small fraction of the usual cost. The government had taken a "strategic call" to build Indian models and was "helping to build the ecosystem" with its five-year $1.2bn AI strategy that it launched last year, said Singh. It has made available 10,000 graphic processing units (GPUs) -- the chips needed to power advanced AI -- to Indian researchers and start-ups and plans to provide initial funding for a few promising ideas. But for India to have a chance to catch up in a field that has been dominated by the US and China, Indian companies need to take larger bets to fund innovation, according to executives and policymakers. Earlier this month commerce minister Piyush Goyal revealed his frustration at a start-up event in New Delhi, asking if India would only produce delivery and quick commerce companies instead of innovating in deep tech. India's tech giants have announced relatively little major investments in AI compared with their global peers. Google, Microsoft, Meta and Amazon plan to spend as much as $300bn on AI this year, while Chinese eCommerce giant Alibaba has said it will invest $53bn over the next three years. Of the $43bn worth of AI investments made globally in 2024, India garnered just $179.3mn, compared to China's $3.3bn and US's $34.2bn, according to data compiled by Tracxn. "I think capital is there," said Pratyush Kumar, co-founder of Sarvam AI, one of India's leading AI start-ups. "It's the intent to deploy, and intent to innovate, intent to lead, these are the things that are required." The Bengaluru-based start-up is working on a larger LLM, after the launch of its first model last year, Sarvam 1, which it hopes will come close in its capability to leaders like OpenAI's ChatGPT and DeepSeek in relation to everyday tasks. "We want to be very realistic about usage of technology" for day-to-day solutions, said Kumar, noting that its model would not solve harder problems such as advanced math and complex philosophical questions. It has raised $40mn from Peak XV Partners, which was previously Sequoia's India and south-east Asia business, as well as US-based Lightspeed and Khosla Ventures. But a top venture capitalist in Bengaluru said that while there was a lot of activity and noise around AI in India, he did not see enough original innovation. "Are 10 fundable ideas coming to us every month? Probably not," he said. While a 2024 Stanford University AI Index report ranked India at the top in AI skill penetration, the country is far behind in terms of the total AI patents granted globally between 2014 and 2022 -- 60 per cent were from China, 20 per cent from the US and only 0.22 per cent from India. Unlike US and Chinese tech companies that have poured billions of dollars into frontier research, India's private tech groups are too narrowly focused on immediate profits to take the risk of supporting original research, according to analysts. "India's IT services firms . . . don't have the ambition to jump from being services firms to product firms," said Anirudh Suri, a venture capitalist and scholar at Carnegie Endowment for International Peace. "These companies don't believe their moat lies in innovation." Keen to bridge this gap for frontier research, Paras Chopra, an Indian tech entrepreneur who sold his software company Wingify to private equity group Everstone in January for about $200mn, has started to use some of his own money to fund AI research in the country. However, Infosys' chief technology officer Rafee Tarafdar insisted it was becoming an "AI-first company". To do so, the $70bn group is focused on three things: upskilling its 340,000-plus workforce to "amplify their potential", use AI to "reimagine" its services and help their customers to leverage AI. Infosys's co-founder and chair Nandan Nilekani added India could succeed by focusing on building applications on top of leading models and "will lead the way in showing people how to actually use AI for useful things" at scale. Nitin Rakesh, chief executive of IT firm Mphasis, said as global companies felt the need to adapt to AI, groups like his would benefit from the years of insight they had about their clients. India's IT services industry "was born out of labour arbitrage. We are moving towards tech arbitrage," he added. Analysts believe New Delhi wants to be indispensable by leveraging its skilled demographic, even if it is still searching for exactly where it can lead in the fast-paced AI sector. "The world can do anything for AI," Modi told American podcaster Lex Fridman last month. "AI is incomplete without India."

From 'hopeless' to a hotspot: India has second-mover advantage in the global AI race: Amitabh Kant

ET reported that globally, India led the way in AI hiring growth in 2024, recording a year-over-year increase of 33.39%. The country has also seen the highest increase globally in AI talent concentration between 2016 and 2024, at 252%, according to a recent report by Stanford.Artificial intelligence (AI) will define the next global power shift, and India must act swiftly to stake its claim, former NITI Aayog CEO Amitabh Kant said, calling AI the "decisive force" of our time. Speaking at the All India Management Association (AIMA) 10th National Leadership Conclave, Kant stressed that AI holds the potential to revolutionise sectors such as health, education, and productivity -- particularly in the developing world. However, he also warned of the social risks posed by unchecked algorithms and called for international governance frameworks. Tuning into Sam Altman's shift Referencing OpenAI founder Sam Altman's earlier remark, where he dismissed the idea of developing a competitive AI model on a $10 million budget as "hopeless", during an Economic Times Conversations event in June 2023, Amitabh Kant highlighted the dramatic shift in Altman's stance, first reported by ET in an exclusive interview. In doing so, Kant made a broader comment on the country's growing influence on the global AI ecosystem. ET reported that globally, India led the way in AI hiring growth in 2024, recording a year-over-year increase of 33.39%. The country has also seen the highest increase globally in AI talent concentration between 2016 and 2024, at 252%, according to a recent report by Stanford. In fact, OpenAI's Sam Altman himself posted on X this month, acknowledging that India is rapidly emerging as a global leader in artificial intelligence (AI) adoption, marked by an unprecedented surge in innovation, talent, and creative energy. OpenAI vs DeepSeek Putting facts into place, Kant compared Microsoft-backed OpenAI, a proprietary, closed AI ecosystem backed by massive capital, with China's DeepSeek, an open-source model built at a fraction of the cost. "This is a classic case where disruption doesn't guarantee victory. It's the second-mover advantage that counts -- and India is well positioned for that." ET reported in January that China's low-cost model was developed within two months and with an investment of less than $6 million, in contrast with the $100 million spent by OpenAI on training its GPT-4 model. The DeepSeek arrival also impacted the stock market, particularly in the tech sector. Companies such as Nvidia, Broadcom, Microsoft, and Alphabet saw sharp declines during that period with Nvidia losing $593 billion overnight. Second-mover advantage According to CPA Australia's Business Technology Report 2024, released in February, 23% of Indian businesses have already adopted AI -- the highest among surveyed countries -- and 73% plan to scale up AI usage in 2025, significantly above the global average of 52%. Having an AI adoption rate of 92%, over 500,000 engineers, and being the world's second-largest developer community, Kant insists India can play a key leadership role. Echoing a similar sentiment, Infosys co-founder Nandan Nilekani recently pointed out that India's prior digital transformation will significantly shorten the gap between global AI advancements and domestic adoption. The country's digital public infrastructure, he said, will act as a key enabler for deploying AI at scale to benefit a billion people.