India’s technology outsourcing sector, a $300 billion juggernaut fuelling 80 per cent of the nation’s services exports, faces an uncertain future as artificial intelligence disrupts its foundations.
Major company shares have plunged, with the Nifty IT index tumbling over 20 per cent so far in 2026, mirroring a global slide in tech equities triggered by innovations like Anthropic’s Claude agent tool launched in February to automate legal, compliance and data tasks.
United States visa hikes are piling on the pressure, set to inflate costs by $100-250 million annually for top firms, roughly one per cent of revenues.
This industry has underpinned India’s economic rise for more than three decades, nurturing a thriving middle class that has driven booms in housing, cars and restaurants across cities such as Bengaluru and Hyderabad.
Nasscom reports 2025 revenues hit $315 billion, but AI’s slice remained modest at $10-12 billion as companies shifted from experiments to rollouts, eyeing just 6.1 per cent growth next fiscal year. Hiring will creep up by a mere 2.3 per cent in 2026, while billing evolves from hourly logs to results-driven models, marking 2025 as a pivotal year for AI impact.
Dismal forecasts abound. Sun Microsystems co-founder Vinod Khosla warned at the AI Impact Summit, “IT services will vanish by 2030,” urging India to pivot its engineering talent towards AI innovation.
Growth could falter three per cent below par over five years before stalling post-2031, some executives even foresee half of entry-level jobs vanishing.
Yet optimism flickers. JPMorgan dubs IT firms the “plumbers of the tech world,” arguing AI lacks bespoke finesse, fostering partnerships for new roles. Infosys CEO Salil Parekh agrees, citing AI productivity gains of 5-15 per cent in coding; the firm plans 20,000 fresher hires in 2025 despite projecting 92 million jobs lost but 170 million created in AI fields.
HSBC’s report “Software Will Eat AI” insists enterprise software’s reliability, honed over decades—“Enterprise-class software has evolved over the decades to be almost error-free with high throughput and reliability. This critical and private IP is not trainable on the public internet”—positions IT as AI’s true enabler.