ewsba.se
On 5 November, US technology stocks saw their worst day since April 2025. The Nasdaq composite index fell 5.5% after Democrats swept the US midterm elections, raising fears that new limits on data centres could slow AI spending. Indian stocks moved the other way. Foreign investors bought $1.6 billion of Indian shares that day, their largest one-day purchase in nearly six years, according to exchange data.
Analysts at JPMorgan and elsewhere now call India a hedge against the AI trade. However, India is not completely insulated from the risks—the exposure goes beyond the stock market as it runs through equipment suppliers, IT services firms and a planned data centre build-out.
India's appeal during the rout comes from its low exposure to AI. Companies related to AI account for about 16% of MSCI India's value, against 70-80% in Taiwan and Korea, according to Goldman Sachs. Chipmakers dominate those markets. On 6 November, South Korea's Kospi closed 6% lower as SK Hynix fell 10% and Samsung Electronics 6%, while the Nifty 50 rose about 2%.
The same low exposure hurt India earlier in the year. Its lack of AI-linked stocks kept foreign investors away, Reuters reported, and they sold a record $30 billion of Indian shares between January and October. The Kospi was still up 20% for the year on 6 November. The Nifty 50 was down about 6%.
Some investors now treat that gap as protection. JP Morgan describes India as an "AI hedge" for funds cutting crowded AI positions. Goldman's Timothy Moe expects markets less exposed to AI, such as India, to see "some relative outperformance". Bernstein says India is "neither an AI beneficiary nor an anti-AI trade", pointing to its valuation of 18.6 times forward earnings, higher than in South Korea and Taiwan.
While the benchmark indices rose, India's own AI-linked stocks fell with global markets. On 6 November, shares of CG Power, MTAR Technologies, Hitachi Energy India and cloud providers such as E2E Networks dropped by up to 8%. These companies belong to a group of 42 that Goldman Sachs identified in September as India's "AI Enablers", worth $670 billion. The group rose about 58% between January and mid-September, led by data centre hardware and power equipment makers, and traded at an 85% premium to the market.
Their outlook depends on how much data centre capacity gets built. In the US, voters backed limits on data centres in 40 of 41 local ballot measures, according to Ballotpedia. A US slowdown would reduce global demand for power and data centre equipment.
It could also move some building to India, which has about 1.5GW of capacity against about 50GW in the US, according to Sterlite Technologies' Ankit Agarwal. Mumbai is the second-cheapest major data centre hub after Shanghai, at about half the cost of Silicon Valley, according to Turner & Townsend data cited by the Rhodium Group.
AI is lowering the price of the work that India's IT services firms sell and creating new revenue for them. Persistent Systems' chief executive told Reuters that clients want the same work at 25-30% lower prices. ICICI Direct estimates that AI could cut traditional IT services revenue by 2-3% a year. AI services make up about 7.5% of revenue at Tata Consultancy Services and 8% at Infosys, according to the Rhodium Group.
Infosys reported a 29% quarter-on-quarter rise in AI revenue in July-September. When AI is 8% of revenue, AI revenue must grow 23-35% a year to offset a 2-3% annual fall in traditional work.
AI-driven cyberattacks could bring IT firms new work. On 4 November, India's cybersecurity agency CERT-In named a hacking group that uses AI tools to breach non-bank lenders, fintech firms and cryptocurrency exchanges, Reuters reported.
McKinsey expects the global market for outside cybersecurity products and services to grow from $223 billion in 2025 to $346 billion by 2029. Nasscom president Rajesh Nambiar told Nikkei Asia that IT exporters must open new opportunities in cybersecurity and AI consulting.
India has only a few companies building AI models, and the rout could make them harder to fund. Of the 84 leading private AI companies identified by Goldman Sachs in October, only seven focus on models, model training, or AI integration. Together, they are worth about $2 billion of the group's $55 billion.
Sarvam AI's $105 billion is India's first sovereign model with over 100 billion parameters, according to the Rhodium Group. Sarvam's Indian-language AI earned under $5 million in FY26.
Building models needs long-term capital. The Rhodium Group argues that the market for Indian-language AI is too small to fund India's model builders. Most large Indian companies work mainly in English, and Rhodium puts India's voice AI market, the most common Indian-language use, at about $200 million in 2025. Sentiment is also turning against AI developers. After the selloff, prospective investors began rethinking the valuation of Anthropic's planned listing, Bloomberg reported.
In October, before the selloff, Goldman estimated that new listings could add $50-100 billion to India's listed AI market over two to five years. Model developers would account for about 3% of that.
India's data centre build-out will need domestic customers, and demand from Indian users is weak. Investment commitments in Indian data centres passed $300 billion over the past year, according to the Rhodium Group. Its report finds that hyperscalers lease about 60% of operational capacity, while capacity used only by Indian entities is under 13%. About 80% of the GPUs subsidised under the IndiaAI Mission reportedly sit unused. Yotta, India's largest AI compute provider, gets nearly 80% of its revenue from foreign clients, by its chief executive's estimate.
Foreign demand is becoming less certain. Microsoft is winding down a deal to rent AI computing from Oracle and has tightened oversight of other data centre projects, Bloomberg reported. Bain's survey of more than 250 technology leaders found that Indian companies expect to raise technology spending by 6-8% in 2026, against 4-6% globally. About 90% of business leaders told Bain their data and AI capabilities cannot support use across the company, and CIOs cite legacy systems, skill shortages and unproven returns as the main obstacles. India's new capacity will need these companies as its main users.www.howindialives.com is a database and search engine for public data
Howindialives.com is a Delhi-based venture set up by former business journalists to combine public data and technology for decision-making. We cover the entire data chain: collect data; process, interpret and visualize it; and design outputs. We have been a data partner to Mint since 2015, narrating data stories and more.
Catch all the Business News , Market News , Breaking News Events, JIO IPO Listing Today LIVE and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.