NextFin News - India's technology industry is climbing the value chain even as artificial intelligence hollows out the entry-level jobs that built it, a shift that has pushed software services exports to roughly 5.2% of gross domestic product, up from 3.3% before the pandemic, according to ING economist Deepali Bhargava. In a note published Wednesday, Bhargava wrote that business services - finance, accounting, engineering, research and analytics - have more than doubled their share of GDP to 3.3% over the same period.
The two trends sit in tension: the work India sells to the world is getting more valuable, but the army of junior engineers that delivered the old model is no longer needed in the same numbers. The question for investors, policymakers and the 6 million people the sector employs is whether India can move upmarket fast enough to outrun the job losses AI is creating at the bottom.
The Numbers: Growth Without the Hiring
The industry body NASSCOM projects India's technology sector will reach $315 billion in revenue in fiscal 2026, up 6.1% from a revised $297 billion in fiscal 2025, with exports of about $246 billion. That growth is arriving with strikingly little headcount expansion: NASSCOM expects a net addition of roughly 135,000 jobs in fiscal 2026, lifting total employment to about 5.95 million - a hiring rate of just 2.3% year on year. Revenue is growing nearly three times faster than payrolls.
The composition of that employment is changing just as fast. The NASSCOM-Zinnov India GCC Landscape Report 2026 counts 2,117 global capability centers - captive offshore units owned directly by multinationals rather than third-party vendors - operating across 3,728 units, employing 2.36 million professionals and generating an estimated $98.4 billion in revenue. The ecosystem has grown 32% since fiscal 2021, with 506 Forbes Global 2000 companies now running a GCC in India. On NASSCOM's own headcount estimate, nearly 40% of the sector's workforce now sits inside these in-house hubs rather than at the outsourcing vendors that dominate the stock indices.
The equity market has drawn the obvious conclusion. The Nifty IT index has fallen about a fifth this year, with its 10 constituents losing a combined $73 billion in market value. The five largest listed vendors - Tata Consultancy Services, Infosys, Wipro, HCLTech and Tech Mahindra - have seen their combined market capitalization drop more than 46% to 18.15 trillion rupees as of July 2026, from a record 33.71 trillion rupees in August 2024. In early June, shares of TCS fell nearly 8% in a single session and Infosys slipped more than 3%, dragging the technology gauge down close to 5%.
Why the Pyramid Model Is Broken
For three decades, Indian outsourcing ran on the pyramid: hire tens of thousands of fresh engineering graduates every year, put the cheap juniors at the base to do routine coding and testing, and stack a thin layer of senior architects on top. Margins came from labor arbitrage and billable hours. AI attacks both.
"The pyramid model is gone. With coding agents, we no longer need basic coding."
That is V. Balakrishnan, the former chief financial officer of Infosys, and the statement captures the mechanism, not just the symptom: when an AI agent can write, test and debug routine code, the bottom third of the pyramid - the fresher cohort that once absorbed a share of India's roughly 1.5 million annual engineering graduates - stops being an asset and becomes a cost line with no work to fill it.
The evidence is already visible in hiring data. Xpheno, a specialist staffing firm, reported in March 2026 that India's technology sector had 119,000 active job openings - up 9% from February but still far below the 262,000 recorded four years earlier. Within that total, IT services - historically the largest absorber of engineering graduates - accounted for 46,000 openings, down 16% year on year. Kamal Karanth, co-founder of Xpheno, has said active demand for entry-level talent across all sectors is under 50,000, with only about one-third coming from the technology sector.
What replaces the pyramid is a diamond: fewer juniors, more mid-level engineers who can orchestrate AI agents, and a premium on domain expertise. That is a different business - and a different kind of hiring pipeline. It also means the sector's traditional role as a mass absorber of graduate labor is winding down, even as the industry's revenue keeps climbing. Wipro cut its fiscal 2026 fresher intake target to 7,500-8,000 from an earlier 10,000-12,000, and has said it has no fixed fresher-hiring target for fiscal 2027. TCS, the country's largest private-sector employer, cut more than 12,000 jobs last year - the only major Indian IT services provider to announce mass layoffs in the AI era.
The Second Shock: Clients Want More for Less
AI has not only reduced the work available; it has flipped the pricing power in contracts.
"It's a desperate market for the service providers. The odds are very much in favour of clients."
That is Jimit Arora, chief executive of research and advisory firm Everest Group, and the shift is visible in how contracts are now written. Outsourcing giants are increasingly tying fees to performance outcomes instead of hours worked - precisely because clients no longer accept paying for time that AI compresses. TCS chief executive K. Krithivasan said about 80% of the company's contracts in its finance, human resources and other business-services segment are now based on outcome performance measures, a figure that has doubled since AI went mainstream in late 2023. Cognizant struck an AI and automation deal with Daimler Truck in February in which AI-related cost savings are split between vendor and client. HCLTech structured a multiyear cloud-management deal with German utility E.ON so that it is not paid in the first year, with subsequent payments tied to efficiency gains.
The pressure on price is explicit. Sandeep Kalra, chief executive of Persistent Systems, said clients were demanding the same work for 25% to 30% less while expecting faster delivery and higher productivity. Tech Mahindra chief executive Mohit Joshi warned on an analysts' call that some rivals are factoring productivity gains of 70% to 80% over five to seven years into their pricing and guaranteeing rates despite rising chip costs.
"Clearly, there is a ton of competition out there, and our competition at times is doing irrational things."
Infosys has told analysts it walked away from contracts that were no longer economically viable.
The irony is that the same force crushing the giants is opening the door for smaller rivals. Firms that can deploy senior leaders quickly and offer flexible pricing are winning mandates that once went to the biggest names on reputation alone.
"Many Tier 2 firms have been more agile and hungry in this phase."
That is Phil Fersht, chief analyst at HFS Research. Persistent Systems and Coforge have both grown revenue in dollar terms by double digits for at least eight consecutive quarters; in the April-June period, Persistent's revenue surged 16% while Coforge's sales jumped by a third. The big four - TCS, Infosys, Wipro and HCLTech - managed only 1% to 3% growth in the same stretch. Scale, once the sector's moat, is now a weight.
Cyclical Pain or Structural Reset?
Is this a downturn that will pass, or a regime change that will not revert on its own? The answer splits cleanly across two time horizons.
The near-term pain is cyclical and will ease. Global technology spending is recovering, clients are moving past the initial "wait and see" freeze on discretionary projects, and the worst of the inventory correction in software budgets is behind us. A rate-cutting cycle in the United States - the destination for the bulk of India's IT exports - typically lifts enterprise technology budgets with a lag of two to three quarters. If that plays out, the revenue growth NASSCOM is forecasting becomes achievable, and the stock multiples that have compressed could stabilize.
But the underlying driver is structural, and it will not mean-revert. Three pieces of evidence support that call. First, the technology itself is a permanent step-change: coding agents and agentic workflows do not get "used up" - they improve, which means the deflation they impose on routine work is durable, not a one-off shock. Second, the business model has already shifted at the contract level: outcome-based pricing and risk-sharing constructs are now the norm in large deals, and clients do not unilaterally return to paying billable hours once they have learned they can pay for results. Third, the employment structure has changed irreversibly - the pyramid that absorbed fresh graduates at scale cannot be rebuilt because the work at its base no longer exists in the same volume.
The critical distinction: revenue can grow structurally while employment grows cyclically - or not at all. That is the jobless-growth risk at the heart of this transition. India's technology sector is proving it can sell more valuable work; it has not yet proved it can employ the same share of the country's graduates while doing it.
The Counter-Case: Upmarket May Not Be Enough
The strongest argument against the "India wins by moving upmarket" narrative is arithmetic. The jobs being created at the top - AI engineers, data scientists, product managers, cybersecurity analysts - are far fewer than the jobs being destroyed at the bottom, and they require skills that a fresh engineering graduate does not have. Reskilling a 22-year-old trained in routine coding into an AI-solutions architect is not a matter of a six-week bootcamp; it is a multi-year proposition, and the window for that transition is exactly the period in which hiring is frozen.
There is also a competitive threat from the upmarket move itself. As India's work becomes more sophisticated, it starts competing less with low-cost alternatives and more with talent in Eastern Europe, Latin America and the Philippines - and, increasingly, with in-house teams that AI has made viable. The global capability center boom is itself evidence of this cannibalization: multinationals are bringing work in-house because AI has made it manageable to run complex operations from their own captive hubs. Every GCC seat is a seat that does not go to an Indian third-party vendor.
Nor is the pricing pressure one-sided in India's favor. If clients are demanding 25% to 30% cost reductions and vendors are guaranteeing 70% to 80% productivity gains over five to seven years, then revenue deflation could outrun the creation of new value-added work. In that scenario, India's IT sector grows its GDP share - as ING's data already shows - while its profit pools compress and its employment elasticity falls toward zero.
The signal that would falsify the structural-upgrade thesis is specific and observable: if NASSCOM's fiscal 2027 data show sector revenue growth above 8% while headcount growth remains below 2% and entry-level hiring stays under 50,000 openings, the jobless-growth reading is confirmed and the "upmarket saves the model" story fails. Conversely, if campus hiring for engineering graduates returns to pre-2024 volumes while revenue keeps growing, the structural-break call is wrong and the current pain was largely cyclical.
What to Watch
Short term, the sector's direction turns on the U.S. technology-spending cycle and the pace of Federal Reserve rate cuts, which flow through to enterprise budgets with a lag. Medium term, the key data points are the NASSCOM annual review's headcount and fresher-hiring figures, the share of outcome-based contracts at the large vendors, and whether the mid-tier firms can sustain double-digit growth without margin erosion. Long term, the question is whether India's education pipeline can pivot from producing routine coders to producing AI-literate problem-solvers fast enough to fill the diamond-shaped workforce the new model requires.
Base case: India's technology sector grows into its upmarket shift, with revenue crossing $315 billion in fiscal 2026 and GCCs continuing to absorb high-value work, while employment growth stays muted at 2% to 3% and entry-level hiring remains a fraction of its historical norm. Upside case: AI creates new service categories - agent orchestration, AI governance, legacy modernization - faster than it destroys old ones, and the sector re-accelerates hiring by fiscal 2027. Downside case: price deflation outpaces value creation, the large vendors' margins compress, and India's technology employment contracts in absolute terms for the first time in decades.
The upmarket shift is real, and the GDP data prove it. But a country does not get rich on GDP shares alone - it gets rich on the jobs those shares support. India's technology sector is learning to sell smarter work. The test ahead is whether it can still hire the people who used to make that work possible.
Explore more exclusive insights at nextfin.ai.
