India's IT industry is going through a quiet but significant transformation. The six largest services companies—Tata Consultancy Services, Infosys, HCL Technologies, Wipro, Tech Mahindra, and LTIMindtree—collectively recorded about 92,000 exits over the past year. That's in line with the industry's normal attrition rate of 11–15%. But here's the thing: they only refilled about 68,000 positions. The result? A net loss of 24,422 employees and a combined headcount that fell nearly 3% to 797,722 in October 2026 from 822,144 a year earlier.
This isn't a collapse. Revenue isn't shrinking. But the math has changed. Roughly one in four departing seats is being left vacant or absorbed. Companies are using attrition as a lever to reshape their workforce without resorting to layoffs. The exits are outpacing new hires in IT and engineering functions in five of the last six months.
The squeeze is happening in specific areas. Staffing firm TeamLease Digital points out that replacement hiring is currently lower than the number of people leaving, particularly in the 5–12 years' experience band. These are roles in application maintenance, manual testing, infrastructure support, and L1/L2 services—exactly where artificial intelligence is reducing the need for routine work.
None of the six companies quantify how many routine roles AI has displaced. But the management commentary points in the same direction. Tata Consultancy Services says AI is creating deflation through productivity gains that aren't yet fully offset by new AI opportunities. AI revenue grew 19% while traditional non-AI revenue declined. HCL Technologies describes AI-driven deflation in core services, offset by market share gains and Advanced AI offerings. Management says the link between revenue and headcount is becoming "non-linear". Transcript +3
Tech Mahindra expects its AI toolchain to bring "lesser headcount requirements" on existing fixed-price projects, with released people feeding growth into new work. Infosys CEO said the same work may need fewer people, though overall work volume is growing. LTIMindtree cites examples where AI agents cut development effort by 38% and manual effort by 50%. Transcript +3
There's another piece to this puzzle. In a weak-growth environment, companies are trimming staff not assigned to any client project—the bench. Pareekh Jain, CEO of market research firm EIIRTrend, says last year was mostly about bench optimization. Companies reduced bench in a low-growth environment, which led to a fall in headcount without impacting revenue.
The traditional pyramid model, which relied on a large base of freshers and a substantial bench to handle large deals, is shifting. Industry reports suggest utilisation targets are being pushed from the standard 80–85% to 85–90% or higher. Bench strength, once 15–20% of the workforce, is expected to shrink to around 10%. Companies like Wipro and Tata Consultancy Services have moved towards contract-based hiring for just-in-time needs, bringing bench capacity to a bare minimum.
Here's the paradox.
AI, cloud, cybersecurity, and specialized engineering roles are growing at around 16% year-on-year. This growth occurs against a backdrop of reduced technology spending from clients amid a challenging macroeconomic environment.
The demand is centred on experienced, specialized professionals. Naukri's CEO notes that the disparity between AI hiring and overall IT recruitment reflects where companies are continuing to invest. AI is becoming a core capability. But there's a catch: people losing jobs in traditional roles may not have the skills required for these new opportunities.
The six companies are pouring resources into upskilling. Tata Consultancy Services reports 367,000+ associates with higher proficiency in AI/ML, with 31.8 million learning hours year to date. Infosys has over 80,000 employees using coding tools like Claude Code and Codex, and plans to build 6,000 "Frontier Engineers" through a mix of external hiring and internal training. Tech Mahindra has about 48,000 AI-certified employees, with 24,000+ upskilled on AI productivity tools and 4,500+ on agent platforms. Investor Presentation +2
HCL Technologies is setting up a centralised Forward Deployed Engineer academy aligned to OpenAI, Anthropic, AWS, Google, and NVIDIA. LTIMindtree runs an AI 1,000 program with nearly 1,000 employees in mentor-led journeys, and over 2,000 certifications across Claude, Gemini, Copilot, and Azure AI. Xpheno notes that the combination of upskilling employees to be AI-ready and mainstream AI deployment in operations "drives net drop" in headcount. Transcript +1
The September quarter showed how differently each company is positioned. Tata Consultancy Services added 4,258 employees, HCL Technologies added 6,450, and Tech Mahindra increased headcount by about 5,400. Meanwhile, Wipro cut 6,716 employees.
The divergence reflects client market exposure and project pipelines. Tata Consultancy Services is hiring to demand, especially for next-generation skills, with an order book TCV of $9.6bn. HCL Technologies reported bookings of $3.3bn, up 35% quarter-on-quarter, and added 6,136 freshers. Tech Mahindra saw headcount rise partly due to an employee transfer tied to the Orange deal, with Q2 TCV of $1,023M. Transcript +5
Wipro, on the other hand, faced a delayed mega-deal ramp-up, weak US healthcare, and flat revenue. The CHRO said headcount is roughly flat year-on-year, campus hiring was skipped in the first half, and AI productivity is contributing. LTIMindtree added 949 employees, with BFSI returning to year-on-year growth and Consumer growing 8.6%. Transcript +3
The client environment is cautious but not collapsing. Clients are squeezing price on traditional work while moving spend towards AI and data. Infosys cut its FY27 revenue guidance to 1.5–3%, citing softer volumes, productivity asks, and pricing competition. LTIMindtree management said traditional services will continue to be challenged, with large traditional deals fought at unsustainable price points. AI rate cards, however, are priced higher. Transcript +3
Wipro expects the pricing deflation cycle to reach an inflection in a few quarters as most contracts have been renewed. Tech Mahindra argues the deflation debate is overstated, noting that rising cloud and hardware costs are a bigger budget factor than AI tokens. Transcript +2
Jain calls the headcount decline a temporary phase driven by weak growth. Recovery in hiring would depend on when growth returns for the industry. The September quarter additions at Tata Consultancy Services, HCL Technologies, and Tech Mahindra suggest some service providers have already started recruiting and reporting headcount additions.
But the structure of the workforce is changing permanently. The shift from a pyramid to a "diamond" structure, prioritising mid-level hiring over entry-level recruitment, is underway. Companies are building capacity ahead of demand for next-generation skills while leaving routine seats unfilled. The math is simple: fewer people doing traditional work, more people doing AI work. The transition is painful for those in the middle, but for the industry, it's a necessary evolution.
