<\/p>\n
On Saturday, finance minister Nirmala Sitharaman announced a fiscal outlay of \u20b9<\/span>2,000 crore ($230 million) for the Mission. The move succeeded the Ministry of Electronics and Information Technology (Meity) announcing last Thursday that a repository of 18,693 graphic processing units (GPUs), provided by private sector companies, was ready to be used as an on-cloud supercomputing service to help train AI models. Startups and academia members will get access to the computing platform after being vetted by Meity-appointed officials on merit.<\/p>\n A senior government official said, requesting anonymity, that the platform that enables access to GPUs to startups is ready, and the framework to decide who gets access to the central compute infrastructure under the AI Mission \u201cwill be decided soon.”<\/p>\n \u201cWe\u2019re taking all steps to ensure that foundational AI work happens at the fastest possible pace. Beyond the \u20b9<\/span>2,000-crore allotment, if we need more capital, there\u2019s room for that as well,” the official said.<\/p>\n Also read | <\/i>Computers unleashed economic growth. Will artificial intelligence?<\/i><\/a><\/p>\n IT minister Ashwini Vaishnaw said during the Mission’s press briefing on Thursday that Meity is \u201calready in contact with six major startups, which will be able to build a foundational AI model with Indic language datasets in the next six to 10 months.”<\/p>\n Industry stakeholders largely lauded the move but cautioned that a prompt phase of on-ground development should be key for the Mission to adequately rival moves such as the Stargate entity in the US and China\u2019s push through its conglomerates.<\/p>\n \u201cIndia has always been a nation that has offered technology services to the world, while the likes of the US have held the fundamental technologies,” said Ajai Chowdhry, co-founder of HCL Technologies<\/a>. \u201cThis meant that India\u2019s margin from its technology sector has remained slim, even as we made up for it by building scale in our IT sector. However, now is the time for India to conduct research and develop core IPs and patents in every field that we\u2019re in\u2014AI, semiconductors and electronics products. This will start generating greater value, give us geopolitical independence, and make sure that we build our own technologies by avoiding the mistakes of the past.”\u00a0<\/p>\n In electronics,\u00a0Mint\u00a0<\/i>reported in December that India\u2019s electronics economy generated just short of $150 billion in revenue. However, India\u2019s gross margin, or the actual value earned through its services, remains around 10%\u2014driven by heavy reliance on assembling components rather than manufacturing them. In comparison, China\u2019s margin in technology businesses is over 30%.<\/p>\n Also read | <\/i>Five artificial intelligence trends to track in 2025<\/i><\/a><\/p>\n At Saturday\u2019s budget, Sitharaman exempted key components required in television and mobile phone manufacturing to draw in the component supply chain to the country. Vaishnaw said on Saturday that the move will help India create \u201cpermanent establishments” for component makers in India, which, in the long run, will keep more parts of the revenue generated from businesses in India, within India.<\/p>\n Jayanth Kolla, partner at tech advisory firm Convergence Catalyst, added that what\u2019s necessary for India now \u201cis for the government to invest in and build a risk-taking capital among academia and entrepreneurs.”<\/p>\n \u201cThe US has become what it is by investing in big risks at scale\u2014such as the $500-billion quantum for AI through Trump\u2019s Stargate. In future, taking such a risk is what will help India be a key stakeholder in the future of the world\u2019s economy\u2014where US and China are the clear leaders,” Kolla added.<\/p>\n Also read | <\/i>From chips to Stargate: Can India weather the storm of Trump’s AI policy changes<\/i><\/a><\/p>\n Developing such core capabilities can bring greater value to the industry. Kashyap Kompella, AI analyst and founder of tech consultancy firm RPA2AI Research, said that the AI Mission\u2019s key goal should be to find strategies that do not leave India at the risk of other nations.<\/p>\n \u201cThe US has brought major muscle to the table, with a $500-billion budget to maintain its dominance in AI,” he said. \u201cIndia cannot really think of keeping up with such a scale of spending. China, meanwhile, has always grown with strong government backing\u2014wherein they first develop scale in an industry by offering the lowest cost of tech services and then use its profits to build core competencies, such as what we saw with DeepSeek<\/a>. In the long run, a favourable India-China tie can see the latter sell its innovations to the West through India as a marketing vehicle, which favours all parties involved. But, this could leave India at a sensitive juncture of geopolitical tensions, the early shoots of which can be seen with US president Donald Trump\u2019s executive orders and sanctions against nations.”<\/p>\n Both Chowdhry and Kompella warned that to avoid having its economy swayed by global geopolitics, India must build its core technology competencies\u2014such as its foundational model backed by its proprietary Indic language datasets.<\/p>\n A senior official with Meity said, requesting anonymity, that the Centre is officially establishing a local language dataset repository different from the consumer-facing Bhashini approach. \u201cEveryone within Meity agrees that the current Mission\u2019s contours are at least a good place to start in terms of core AI innovation. A lack of consensus on the value addition of the Indian government\u2019s technology incentives would only hold the industry back. The GPU access on-cloud is set to go live shortly, and by some time next year, the early advances on India\u2019s own, government-backed foundational AI model could be available for peer reviews and startups to work on,” the official said.<\/p>\nTech sector margins slim<\/h2>\n
Geopolitics shield with core tech<\/h2>\n