India's pay curve inverted: 45% hikes at top, 6% at bottom

TeamLease's August 2026 hiring data, shared by CFO Ramani Dathi with CNBC TV18, shows India's salary curve has inverted. Professionals earning above ₹20 lakh are seeing roughly 18% year-on-year hikes, and premium skills in the ₹50 lakh-plus bracket are commanding 40-45%. Broad-based roles below ₹20 lakh are stuck at 6-6.8%. Overall hiring rose 14% year-on-year, with IT services, GCCs and BPO all adding headcount after nearly two years of weakness. The demand is concentrated in cloud management, semiconductors, cybersecurity, AI/ML and data roles, and it is coming from non-tech sectors too. Meanwhile, mid-level coding and testing roles are being actively cut as AI tools absorb that work. Key signal: TeamLease data shows ₹50 lakh-plus premium-skill talent getting 40-45% hikes and ₹20 lakh-plus roles ~18%, while sub-₹20 lakh roles get just 6-6.8% and mid-level coding and testing roles are being cut. For hiring leaders, this matters because this is the clearest evidence yet that India's hiring market is not recovering evenly, it is splitting. If your compensation bands are built on the old assumption that seniority buys you a discount on percentage hikes, you are about to lose your best AI, cloud and cybersecurity people to a competitor willing to pay 40% more. The flip side is equally important: mid-level coding and testing capacity is now a cost problem, not a talent problem. CHROs and delivery heads need two separate playbooks, one for scarce premium skills where you pay to retain, and one for AI-displaceable roles where you redeploy or exit fast. Teksands view: Stop running one increment budget for the whole company. The 40-45% premium for AI, cloud, cybersecurity and semiconductor skills is the real clearing price, and the 6% for broad-based roles reflects work AI is already doing. Two bands, two budgets, two retention strategies. Anything else means you lose the scarce people and overpay for capacity you no longer need.