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

TeamLease data shared with CNBC TV18 shows professionals earning above Rs 20 lakh getting roughly 18% year-on-year hikes, with premium skills in the Rs 50 lakh-plus bracket commanding 40-45%. Broad-based roles below Rs 20 lakh are stuck at 6-6.8%. Overall hiring rose 14% year-on-year, with IT services, GCCs and BPO adding headcount after nearly two years of weakness. Demand is concentrated in cloud management, semiconductors, cybersecurity, AI/ML and data roles, and it is coming from non-tech sectors too. Mid-level coding and testing roles are being actively cut as AI tools absorb that work.

Teksands take: 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.

Wipro's CTO has stated that the company's AI initiatives have freed up capacity equivalent to 20,000 workers, according to Reuters. The disclosure underscores how India's IT services majors are using generative and agentic AI to automate routine delivery work, boost productivity and rethink headcount-linked revenue models. For a sector that has historically grown by adding people, the implication is that AI is not just a new service line but a lever that changes the economics of existing contracts. The capacity gain does not automatically mean layoffs, but it does mean hiring will be increasingly selective, with demand shifting toward AI engineering, data, cloud and platform skills rather than traditional support and maintenance roles.

Our take: The headline number will make every delivery head ask why their bench isn't shrinking. But the real story is not layoffs; it is the quiet reallocation of talent toward AI, data and platform roles. Recruiters who keep selling generic Java and testing profiles will find the market thinning, while those who can source and assess AI engineering talent will command a premium. Map internal capacity freed by AI and redeploy it before competitors do.

GCCs have the budget, the mandates and the leadership backing. What they don't have is people.

India's Global Capability Centres are hitting a talent wall. Demand for advanced skills, particularly AI engineering, data, cloud and cybersecurity, is outpacing the supply of candidates who can actually do the work at the depth GCCs now need. The story lands as GCCs have become the country's most aggressive hirers of senior technology talent, competing directly with product firms and IT services for the same narrow pool. The signal is not that GCCs are slowing down. It is that the constraint has shifted from budget and mandates to people, which changes how GCCs hire, where they hire, and what they are willing to pay for proven specialist capability.

The GCC hiring story has been about scale and mandates. The real constraint now is depth of skill. Companies that keep competing on brand and CTC will keep losing candidates to faster movers. The advantage goes to firms that build capability internally, hire on demonstrated skills, and move into Tier-2 talent pools before everyone else does.

Four signals worth watching

The lesson: The chairman of India's second-largest IT services firm publicly describing large-company AI efficiency as a headcount compressor is the most honest thing said about this market all month. It also reframes TA strategy at big employers: if large firms are net job compressors in the AI era, volume hiring pipelines lose their purpose. Hire a handful of AI-native engineers who make ten average hires unnecessary, and stop pretending the pyramid is coming back.

Two markets, one message: scarcity at the top, oversupply everywhere else

Hire for the skills the title doesn't name

If half of AI requisitions are mislabelled and GCCs are hitting a talent wall on advanced skills, the bottleneck is assessment, not sourcing. Teksands works with GCCs, IT services firms and product companies to define skill-based role architecture, build AI and platform talent pipelines, and run hiring processes that test demonstrated capability rather than keyword matches.

Run a title-versus-skills audit on your last 50 AI requisitions

Take your last 50 AI, ML and platform requisitions and put the job title next to the actual skills listed. Where the title says 'AI Engineer' but the skills demand LLM orchestration, retrieval design, evaluation and vector databases, you are advertising a role that does not exist and screening out people who do the work. Rewrite those titles around the skill bundle, re-open the pipeline, and change the interview loop to ask about production retrieval failures, evaluation regressions and cost-latency trade-offs instead of model theory. Do it before the next campus cycle, not after.