A big sell-off of Indian stocks by foreign institutional investors (FII) was triggered by the recent outbreak of war in West Asia and its impact on our economy. Then a narrative emerged of India being ‘nowhere in AI’ (artificial intelligence).
A big sell-off of Indian stocks by foreign institutional investors (FII) was triggered by the recent outbreak of war in West Asia and its impact on our economy. Then a narrative emerged of India being ‘nowhere in AI’ (artificial intelligence).
The fall of Indian market indices, led by IT stocks, even as AI-rich indices surged across the globe, bolstered this narrative.
The fall of Indian market indices, led by IT stocks, even as AI-rich indices surged across the globe, bolstered this narrative.
In AI, India is currently a fast follower; it is not a leader (unlike the US and China) but not a laggard either. Globally, the buzz around an ‘AI bubble’ is getting louder.
If the benefits and profit potential of AI , as reflected in the stock prices of AI-related companies, have been overestimated, then inflated valuations point to the existence of an AI bubble that might burst.
Domestic narratives suggest that the earnings of listed Indian companies are not dependent on AI. Logically then, Indian stocks could act as an investment hedge against an AI bubble.
What unites tech bubbles: The last 200 years have seen several tech bubbles get inflated by investor enthusiasm. The British railway bubble of the 1840s, the US automobile bubble of 1880-1910 and the internet bubble of the late 1990s are all examples. Such bubbles reflect Amara’s Law: People overestimate the benefit of a new technology in the short-run and underestimate its long-term impact.
Tech bubbles also exhibit three basic characteristics: overestimation of the speed of adoption that would drive earnings up, outsized capital expenditure and underestimation of competitive pressure. However, the underlying technology has always been transformative and AI may prove more so than the railways, automobile and the internet, irrespective of a bubble.
How a tech bubble bursts: A recent Gartner study suggests that AI-driven automation could raise operating costs, driven up by the cost of computation and a specialized workforce. Even at today’s discounted compute costs, AI projects fail to scale meaningfully.
This is an issue for organizations with a constrained ability to re-imagine existing processes and redesign operations. They require auxiliary capabilities to make the best use of AI and need to experiment. This takes time. Investors buying stocks at sky-high prices could turn impatient if AI returns take too long.
Spending on AI infrastructure has surged. Earlier, hyper-scalers like Amazon and Microsoft were using profits for capex while AI labs with large language models (LLMs) relied on equity funding. Of late, however, debt has also been financing AI capex. The leverage ratios are assuredly low right now. But history shows that debt finance used during the bubble phase of railway, electricity and automobile expansion ended in a pile-up of bankruptcies.
Chinese AI labs with LLMs are drastically reducing their compute costs. Their experience raises questions about the training and inferencing architecture of their US peers and the valuations they command. And history also shows that tech bubbles burst once interest rates spike.
India’s AI play: The Nifty IT Index’s poor returns (about 2% since November 2022 when OpenAI launched ChatGPT and some 25% year-to-date) are indeed worrisome. The stocks of two of India’s largest IT majors that have performed badly have a weight of about half in this index.
However, some mid-tier IT companies have delivered 80%-plus returns. Some of them have already plugged themselves into the global AI ecosystem. Meanwhile, market investors overlook some of India’s most intently AI-focused firms because they are unlisted.
That India is a fast AI adopter is beyond dispute. Github, a leading global cloud platform for AI projects, has identified India as among its fastest growing markets. Indians account for about 15% of Github’s user base and its database indicates these 27 million Indian users are contributing to 7.5 million projects.
Moreover, as per BCG’s AI Radar 2026, a global survey of over 2,300 CXOs, with over 200 from India, revealed that 76% of Indian CEOs believe that their AI investments will pay off. The corresponding proportion for US-European CEOs is under 60%. As per Nasscom, 1,200 of India’s 2,300 global capability centres are focused on AI. Thus, India being ‘nowhere in AI’ is an ignorant statement.
Why a burst bubble could suit India: If the Indian rupee’s exchange rate is kept largely stable to protect the dollar value of investments in India, then at some point the theme of Indian stocks acting as a hedge against an AI bubble may play out.
A bubble burst could put the spotlight on India’s mid-tier IT companies. But most importantly, in the aftermath of such a burst, a reduction in global AI capex could grant India the time it needs to leap into the AI leadership cohort.
On current trends, ‘intelligence’ is set to become a commodity.
Several countries, India included, will probably come up with new foundational models. But that’s just Round One. For ‘sovereign AI,’ the ability to keep upgrading ‘intelligence’ generating technology in terms of precision, cost of delivery and ease of adoption, we would require huge commitments by the private sector.
The quest should go well beyond tinkering with AI capabilities that already exist or playing the role of AI distributors in India. Else, from today’s oil shock, the country might graduate to an AI-token import shock tomorrow. The government should initiate a plan to attract AI talent from abroad and focus on creating an environment of ‘ease of doing research.’ In the AI era, business gains follow research efforts.
The author is a quantitative risk management professional and a visiting faculty of risk management at IIM Ahmedabad and IIM Calcutta.