The hugely hyped AI Impact Summit in February turned out to be all that it was promised to be—a jamboree with thousands thronging the venue to get a peek at the future, epic traffic jams that paralysed New Delhi and an event that was overshadowed by Prime Minister Narendra Modi, his philosophy of what AI or artificial intelligence should be and the role India would play globally in this cutting-edge technology. It did not matter that Bill Gates pulled out as the keynote speaker, or that Sarvam AI, the front-runner in India’s race to have its own foundational AI model, failed to evoke any excitement. What the summit did deliver was a trademark Modi acronym and a disastrous business plan to make India a hub of the world’s data centres. It revealed a deep schism between the two objectives.
So we got MANAV, a “framework based on human values”. Described as the PM’s vision for AI, it stands for moral and ethical systems; accountable governance; national sovereignty; accessible and inclusive; and valid and legitimate principles. The main aim is to transform AI from being machine-centric to being human-centric and to “make it sensitive and responsive” and to ensure “the welfare of all, the happiness of all”, as he explained in an interview to ANI news agency. It is not such an original idea. Since 2017, the UN has been running an initiative called AI for Good—Unlocking AI’s Potential to Serve Humanity. This is a collaborative effort spearheaded by International Telecommunication Union (ITU), the specialised agency for digital technologies, along with 53 other UN organisations. Not many here appear to be aware of this project.
But before thinking of the global good, India needs to work urgently on its own economic welfare. A crisis of enormous proportions has been building that could deal a lethal blow to the critical tech services sector which stands on the frontline of AI disruption, says an official analysis. The AI juggernaut could wipe out 2.7 million jobs by 2031 if the government does not take urgent steps, warns a report put together by its top think tank NITI Aayog. Released in October 2025, the report is blunt about the underlying problems of the sector and the consequences in store for the country if immediate action is not initiated. The time frame before disaster strikes is narrow— just five years.
NITI Aayog examined the tech services sector which it sees as strategically vital to the economy because it employs 13 per cent of the total workforce and over 30 per cent of the white-collar talent. Here, the headcount could go down from 7.5 to 8 million to 6 million by 2031. Similarly, the number of employees in the customer services sector could go down from 2-2.5 million to 1.8 million. There are more worrying forecasts. The Centre of Advanced Study in India says over 60 per cent of formal sector jobs in the country would be vulnerable to AI automation by 2030.
The report takes a knife to the complacency that has marked the tech industry and exposes the huge fault lines running through it. The report uses data published by the 2025 AI Index Report to show just how poorly India lags in AI publications, citations and AI patents. India’s share of granted AI patents fell from 8-10 per cent in 2010 to under
5 per cent in 2023, it points out. A sense of urgency runs through the report which lists three major challenges that need to be tackled immediately: the sheer scale of the jobs at risk, the fundamental shortcomings in education and skills programme and the growing shortage of AI talent in the country. It is incredible that India, which has world’s largest pool of young digital talent, finds itself in such a dire predicament.
NITI Aayog’s action plan calls for the establishment of an India AI Talent Mission, a nationally coordinated programme involving all of the government that is focused on just one objective: equipping India’s workforce for the AI disruption. AI, it warns, “is advancing faster than policy, curriculum, and skilling cycles can adapt to”. If India does not act in time, it risks not just irreversible job losses in its flagship tech services, but worse, it will suffer “broader societal disruption, economic marginalisation and a weakening of its global competitiveness”.
Can such fundamental flaws be set right in five years? Are we hoping for a miracle? Perhaps, this is one reason that even before the AI Impact Summit opened, the PM had pitched the country as the hub for the global AI industry’s data centres. The grand invitation to “all of world’s data to reside in India” appears to be the government’s best chance to push its foot into the AI world where it has remained on the edges so far because it has neither the infrastructure—specialised chips which are the basic hardware for AI—nor the capacity yet to develop a foundational LLM or Large Language Model.
The plain fact is that the US $200 billion in investment India is expecting in data centres here—Google, Amazon and Microsoft have made huge commitments—will not help India in its AI enterprise. The AI behemoths do not share data or technology. Nor will they provide huge employment as the government claims. Developing foundational LLMs is vital for our AI sovereignty and the data centres are irrelevant to that pursuit.
What the centres will do is to put tremendous pressure on our already strained natural resources, water and power in particular. A 100 MW data centre can consume over 2 million litres of water daily and here we have Google alone setting up a massive 1 GW centre in Visakhapatnam spread over 243 hectares.
The centres are projected to need 2.5 GW capacity by next year and this would result in high carbon emissions because India still relies on coal-based power. Besides contending with the emissions and a bigger carbon footprint, Indians will also have to brace themselves severe grid disruptions. Since more than half our districts are at high risk from extreme heat, data centres in the vulnerable areas would require higher cooling loads. While the use of renewable energy corridors or liquid cooling processes are emerging, the UN Environment Programme cautions that the immediate strain on resources remains a major worry.
Given this, would the PM apply his MANAV framework to the data centre enterprise; a framework that puts the human being at the centre along with accountable governance?
This column was originally published in the March 1-15, 2026 print edition of Down To Earth