On a February morning in 2026, Anthropic quietly shipped an enterprise AI tool from San Francisco. Thousands of kilometres away in Mumbai, the Nifty IT index dropped nearly 6 per cent in a single day, its worst session since the pandemic crash of March 2020. Three months later, OpenAI’s 4bn-dollar-plus push into enterprise deployment knocked Indian IT stocks down another 3.7 per cent.
Every frontier release from Silicon Valley now lands on the Indian market like a precision-guided quake. Across the first half of 2026, more than 23bn dollars in foreign money fled, dragging foreign ownership to 14.7 per cent, a level last seen fourteen years ago. The Nifty IT index, India’s answer to the Hang Seng Tech index, has fallen 49 per cent over 18 months, wiping more than 19 trillion rupees off its ten largest firms.
The cause is structural, not cyclical. For thirty years India ate well on the population dividend, answering the world’s phones and patching its legacy code. But when a token costs less than an hour of human labour by the river, the outsourcing line that fed a middle class loses its meaning. India, the planet’s most crowded reservoir of carbon-based labour, missed the silicon dividend entirely and became the first to be harvested by the AI sickle.
The pain is concrete. Cognizant’s Project Leap set aside 200m to 270m dollars just for severance as it cut 12,000 to 15,000 roles, mostly in India. TCS posted its first annual revenue decline in years; Infosys grew just 3.1 per cent; and India’s five largest IT firms swung from a net hire of 12,718 to a net cut of 6,981. TCS alone shed over 23,000 people, its biggest reduction since 2008.
The crueller part is the upside India missed. Altimeter estimates 2026 global AI net profit at 637bn dollars: the US takes 49 per cent, South Korea 35 per cent, and the two split 84 per cent. The remaining 16 per cent goes to Taiwan, mainland China, Japan and Europe. India’s name is absent. A nation that skipped industrialisation and leapt straight to services now finds the service it specialised in is the one AI does best, and cheapest.
Read the original report (Sohu IT)
Translated and adapted from Sohu IT (it.sohu.com).