India’s IT services industry is not shrinking. It is being repriced. NASSCOM projects the technology sector to cross $315 billion in FY26, up 6.1%. The bigger story sits underneath the growth: AI is changing what these firms sell, how they charge, and who they hire.
What is changing in IT services
For thirty years the model was simple. Sell people by the hour, staff projects with a pyramid of juniors, and bill for the effort. AI compresses the effort.
NASSCOM’s 2026 review puts it plainly. Providers are “moving away from FTE delivery to outcome-based, risk-sharing constructs as AI-driven productivity materialises”. Technology spending is expected to stay range-bound at 5% to 7%, with hiring shifting from volume to skill mix.
The largest firm in the sector has already crossed the line. In Accenture’s third quarter of fiscal 2026, managed services revenue of $9.39 billion overtook consulting at $9.33 billion. Managed services grew 5%; consulting grew 1%.
The billable hour is the problem
If you bill by the hour, making the work faster reduces your revenue. That is the trap AI has set for the whole sector.
Reuters reports TCS, Infosys, Wipro, HCLTech and Cognizant tying fees to outcomes as clients demand steep price cuts, shift work in-house and shorten contracts. Cognizant says fixed-price and outcome-based contracts crossed 50% of its revenue for the first time.
The deflation is measured. Hexaware’s chief executive has said AI could cut the price of some routine IT projects by 20% to 25%. Everest Group has cited productivity gains of 30% to 70% in some services. HFS Research found 40% of organisations expect AI-led services to cost 10% to 30% less, yet only 19% have redesigned how they buy them.
Why the shares fell anyway
Through 2026, share prices across the major IT services firms fell sharply. Some of the firms it marked down were growing and raising guidance.
The clearest evidence came from disclosure itself. TCS reported about $2.3 billion in annualised AI revenue, HCLTech about $620 million, Infosys about $1 billion, and Wipro disclosed nothing. All four were de-rated by roughly 35% to 39%.
The market was not paying for the AI line. It was discounting the base underneath it: the hundreds of thousands of billable people whose hours the technology is learning to compress.
The AI revenue number nobody can compare
There is a reason the market discounted the disclosures. AI revenue has no standard definition.
GAAP does not define it, so companies choose what to include and how to count it. That makes two similarly named figures hard to compare. Accenture stopped separately disclosing advanced AI bookings and revenue from the first quarter of fiscal 2026, having reported cumulative figures of about $11.5 billion in bookings, 11,000 projects and $4.8 billion in revenue.
NASSCOM estimates AI revenue across IT services at $10 to $12 billion in FY26, about 5% to 6% of average company revenue. Greyhound Research argues that number is a useful signal but not a full measurement model, and that operational indicators matter more.
The mid-tier is winning
The gap between the largest firms and the mid-tier is now wide and consistent. Mid-tier firms have outpaced large caps on growth and margins for two straight years.
Coforge reported FY26 revenue of $1,870 million, up 29.2% in dollar terms, and said 86% of its revenue comes from AI-led engineering, data and cloud services. Persistent Systems grew 38.4% year on year in the first quarter of FY27. LTIMindtree managed 8.3% in the same quarter.
HFS Research attributes this to agility. Mid-tier firms are less weighed down by legacy transformation programmes and large, slow-moving deals.
The pyramid is collapsing
The change that will outlast the share prices is in hiring. Xpheno data shows fresher hiring in Indian IT fell from about 600,000 in FY22 to roughly 120,000 in FY25, an 80% decline.
EY estimates entry-level IT roles in India have fallen 20% to 25% because of automation. The tasks that used to be a junior engineer’s apprenticeship — testing, basic coding, documentation — are the tasks AI does well.
The headcount numbers follow. India’s five largest IT firms cut a net 6,981 jobs in FY26, after adding 12,718 the year before. Freshers now account for around 15% of hiring, down from nearly 28%.
What it means for the next cohort
The industry has stopped manufacturing the people it now says it cannot find. That is the arithmetic behind the skills gap.
McKinsey Global Institute estimates 30% of job hours in India could be automated by 2030. Quess Corp puts the AI and data skills gap across India’s GCC ecosystem at 38% to 42%. TeamLease Digital data shows mid-to-senior talent rose from 60% of GCC hiring in 2023 to more than 77% in FY26.
Business Standard reports that a technology programme which once took 50 engineers a year can now be delivered by four to ten people in about four weeks with AI. The entry rung that trained a generation is closing behind them.
What to watch
Whether AI revenue gets a definition that survives audit, because without one the market will keep discounting the disclosures. Whether the mid-tier’s growth holds as its acquisitions bed in and AI deflation reaches its contracts. And whether the industry rebuilds a deliberate early-career path, because the senior AI engineer of 2031 is the junior hire of 2026.
The firms that re-rate will be the ones that change what they sell. The ones that only change what they report will keep getting the same treatment.
Editor’s Note
This analysis draws on NASSCOM‘s Strategic Review 2026 and its quarterly industry review, Reuters reporting on the sector’s shift to outcome-based contracts, brokerage and industry commentary from HFS Research, Everest Group and Greyhound Research, company results from Accenture, Coforge, Persistent Systems, LTIMindtree, TCS and Cognizant, and hiring data from Xpheno, EY, TeamLease Digital, Quess Corp and the McKinsey Global Institute. Where a figure is an estimate or a vendor’s own, we say so.