India's NASSCOM released its annual workforce report in January 2026 with a headline figure that surprised even optimists: the Indian technology sector created approximately 205,000 net new roles classified as AI-related during 2025. That figure exceeds the combined AI hiring of Singapore, Australia, and South Korea over the same period.
What counts as an AI role
NASSCOM's methodology counts any position where the primary function involves building, deploying, fine-tuning, evaluating, or managing AI systems. The breakdown is roughly 40% in machine learning engineering and MLOps, 28% in AI product management, 18% in data engineering that feeds AI systems, and 14% in AI safety and evaluation.
Geography of the boom
Bengaluru accounts for 47% of the new roles. But Hyderabad's share has grown from 18% to 24% as companies including Juspay, Darwinbox, and Freshworks have doubled their AI headcount. Pune is emerging as a hub for AI-assisted manufacturing applications.
The hiring split
Global multinationals absorbed roughly 55% of new AI hires. Domestic tech companies took 30%, and the remaining 15% went to roles at Indian startups. The ability to work with Indian language datasets — across Hindi, Tamil, Telugu, Bengali — commands a meaningful premium at domestic companies building for Indian consumers.
The gap nobody wants to talk about
Despite the volume, the highest-value roles — model research, architecture design, safety evaluation — are disproportionately going to IIT, IISc, and top IIM graduates. Tier-2 college graduates are being absorbed into implementation roles with shorter career horizons.
The risk is not that India will fail to create AI jobs. The risk is that benefits concentrate at the top, reproducing the stratification that has historically characterised India's technology sector.