Tata Consultancy Services reported FY26 revenue of $30.0 billion, down 0.5 per cent year-on-year, the first annual decline in the company's modern era, and ended the year with 584,519 employees1. Infosys cut headcount by 8,400 in Q4 FY26 to 328,594 and guided FY27 revenue growth at just 1.5 to 3.5 per cent, with shares down 22 per cent in 20262. Wipro hit its zero-fresher hiring target for FY27, the advance signal flagged in mid-April.
The three companies together employ over a million people and supply the application-management, testing and customer-support backbone of most Fortune 500 IT estates. A coordinated turn negative across all three in the same quarter is the first hard evidence that the substitution is now quantifiable in the offshore services market that built itself on bounded, structured, repeatable work, the exact substrate that AI agents perform best on. TCS's annualised AI revenue crossing $2.3 billion is the partial offset. The headline FY26 decline, despite that AI revenue, is the net signal: the substitution is running ahead of the new product line.
The Indian IT trio's results are the closest thing to a pure-play empirical reading of AI substitution at scale. They lack the diversified revenue mix that lets US hyperscalers absorb the displacement inside cloud and product growth. They serve global enterprise IT estates that are themselves cutting headcount; the demand-side compression and the supply-side substitution are pointed in the same direction at the same time. Wipro's zero-fresher hiring target is the most explicit version of that compression: an entire annual graduate cohort that previously fed the sector is no longer being absorbed.
For Indian engineering graduates, the market that absorbed roughly 350,000 fresh entrants a year through the late 2010s and early 2020s is no longer doing so at scale. For US, UK and EU enterprise IT departments that depended on the Indian services trio for cost-effective bulk technical labour, the pricing and capacity assumptions of the next contract round will look different.
