What Ambani’s deep-tech ambitions reveal about India’s AI limits

{"text":[[{"start":7.55,"text":"As New Delhi hosted an AI summit in February, Narendra Modi declared his ambition for India to rank among the top three superpowers developing the advanced technology, outlining a vision of homegrown models deployed worldwide that would create millions of “high-quality jobs”."}],[{"start":25.45,"text":"Among those who have got on board with that goal is Mukesh Ambani, one of India’s most dominant billionaires, who echoed the prime minister’s message with his own heady dose of AI nationalism last week. During Ambani’s much scrutinised annual address to Reliance Industries’ shareholders, he laid out the role his petrochemicals-to-telecoms conglomerate would play as “India’s sovereign AI backbone”. "}],[{"start":50.75,"text":"“India should not be a mere consumer of AI created elsewhere,” Ambani said. “It must become a creator, adopter and global leader in AI.”"}],[{"start":60.45,"text":"Like many of India’s largest business houses and IT groups, Reliance is eager to demonstrate that it will not be left behind in the AI arms race. Yet Ambani’s plans also reveal why India remains far from becoming a global tech leader."}],[{"start":76.45,"text":"For all the rhetoric around Indian AI sovereignty, Reliance’s strategy so far is largely built upon western deep-tech innovation. Like tech groups in other countries, it is investing heavily in data centres powered by Nvidia hardware. Reliance is also integrating large language models developed by Google and Meta into its digital and telecoms business Jio."}],[{"start":100,"text":"Distribution has made Reliance a formidable domestic powerhouse. Jio, which has more than 500mn subscribers, will underpin Reliance’s AI push."}],[{"start":111.2,"text":"Tellingly, Jio is described in the prospectus for its upcoming IPO as the preferred “Gateway to Digital India”, alluding to Ambani’s role as either a feared competitor or an access-easing partner to foreign corporates seeking to break into the world’s most populous nation. "}],[{"start":null,"text":"

"}],[{"start":127.10000000000001,"text":"So far, global tech giants appear to see India as a market rather than a source of AI technology. After receiving a rock-star welcome at the end of 2024 when he appeared on stage with Ambani in Mumbai, Nvidia chief executive Jensen Huang was later asked whether the chipmaker would start sourcing from India’s nascent semiconductor industry. Huang did not express any desire to reduce dependence on Taiwan Semiconductor Manufacturing Company. “They’re the world’s best at what they do — like we’re the world’s best at what we do — by a mile.” "}],[{"start":160.60000000000002,"text":"Those comments reflect scepticism about India playing a major role in the industry compared with Asian rivals including China, Taiwan and South Korea, all of which have cornered part of the global AI play. "}],[{"start":172.55,"text":"Can India catch up? Despite its huge IT sector, India rarely sits at the centre of global tech breakthroughs. The country spends a paltry 0.6 to 0.7 per cent of GDP on research and development, compared with roughly 2.5 per cent in China and 3 to 3.5 per cent in the US. India also lacks large, well-organised datasets needed to train advanced AI models. "}],[{"start":197.60000000000002,"text":"Still, India’s tech-savvy workforce is often cited as a source of optimism. The Stanford Institute for Human-Centered Artificial Intelligence found India accounted for the second-largest share of identified AI researchers and developers last year (the US was top). But many make a beeline for better opportunities and pay overseas, with India recording a near 17 per cent net outflow of such talent last year, the world’s largest. "}],[{"start":222.95000000000002,"text":"Does it matter for India if it remains just one of many nations without a thoroughbred AI horse in the race? Some of its most influential tech leaders argue the country should focus on its strengths, such as developing applications rather than attempting to create expensive LLMs. In this view, India should look for practical AI services tailored to local needs. "}],[{"start":245.8,"text":"That might be the more realistic path. But the recent US ban on exports of Anthropic’s latest models to foreign nationals highlights future risks: Bernstein analysts believe cutting-edge versions could become the equivalent of nationally restricted “fighter jets”. "}],[{"start":262.2,"text":"“India can’t build its AI future on borrowed models,” they wrote in a recent report. “While renting compute to power foreign LLMs may look like progress, it leaves the country dependent.” "}],[{"start":273.59999999999997,"text":"The likes of Reliance may reap commercial success in using and adapting AI tech from elsewhere. But the country might need much more if it wants to meet Modi’s ideal of Atmanirbhar Bharat, or self-reliant India, in AI."}],[{"start":294.49999999999994,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1782362357_4299.mp3"}