Indian IT built a $315-billion empire on headcount and billable hours — now the very engine of that growth is running in reverse, and H1FY27 is where the bill comes due.
It was, on paper, a perfectly ordinary guidance tweak.
On June 18, 2026, Accenture — the world's largest IT consulting firm — trimmed its full-year revenue growth forecast by a single percentage point, from 3–5% to 3–4%. On a company doing roughly $70 billion a year in revenue, that is a rounding error. The company had just reported quarterly revenue of $18.7 billion — up 6% year-on-year — and beaten its earnings-per-share estimates.
One percentage point of guidance, and the market knocked off nearly a fifth of the company's value in a single day.
Accenture's shares plunged ~18% — its steepest single-day fall on record — despite beating earnings estimates and expanding margins. By the next morning in India, the damage had arrived. Infosys led the Indian fall with a drop of more than 8%; TCS, Tech Mahindra, LTIMindtree, HCLTech, and Persistent Systems recorded losses of around 5–6% each.
In a single trading session, Indian IT stocks wiped out over ₹1.35 lakh crore in market value.
The market was not reacting to a bad quarter. It was reacting to a question — one that has been quietly building through 2026 — that nobody in the industry wants to answer aloud: Is the core business model of Indian IT structurally broken?
To understand the fear, you first need to understand the machine.
The Indian IT sector contributes 7.3% of the country's GDP, accounts for approximately 43–45% of total services exports, and employs around 5.8 million tech professionals. India's technology sector reached $315 billion in revenue in FY26, with exports exceeding $246 billion.
The engine behind all of this is elegantly simple. A client in New York or London needs software built, maintained, tested, or modernised. Indian IT companies hire engineers — by the thousands, then by the lakhs — and bill those engineers' time to the client. The more engineers, the more billing. The more billing, the more revenue.
This is what analysts call the FTE-led model — FTE standing for Full-Time Employee. It is the reason TCS, Infosys, HCL Technologies, and Wipro collectively employed over 1.5 million people at their peak. Adding headcount was the growth strategy.
AI is now attacking that model at its foundation.
Here is the mechanism that matters — and it is not the one most people imagine.
AI tools like GitHub Copilot, Anthropic's Claude, and OpenAI's coding models do not eliminate an IT project. They compress the effort required to execute it. A task that once needed a team of 50 engineers over three months might now take 20 engineers in six weeks. The output is the same. The billable hours are not.
As generative tooling automates code, testing, and support work, the same scope of a client engagement requires fewer billable hours — and clients expect to pay accordingly. This produces a paradox the sector has not faced before: a vendor can deliver more output, win the same logos, and still watch revenue per contract shrink. For firms whose growth was historically a function of adding engineers, the lever that once drove the top line has become the thing eroding it.
Wall Street is increasingly concerned that the traditional headcount-based, billable-hour model is being structurally cannibalized by rapid enterprise adoption of generative AI, which compresses consulting hours. Industry commentary suggests this could translate into roughly 2–3% annual deflation in traditional services revenues as automation improves delivery efficiency and squeezes effort-based pricing.
The segments under pressure are telling. Coding, ADM (application development and maintenance — the ongoing work of keeping enterprise software running), and testing are where the pressure is most visible. These are not peripheral activities. They are the bread-and-butter of Indian IT outsourcing — the work that built Bengaluru and Hyderabad into global tech hubs. Analysts estimate that 30–40% of IT services revenue sits in exactly these zones, concentrated precisely in the work most exposed to AI-led deflation.
HCL Technologies CEO C. Vijayakumar has put a blunt number on it: 40% of the industry runs the risk of being disrupted by AI and could shrink at a 3–5% CAGR for a few years — which, mapped to HCL's own portfolio, works out to 2–3% annual revenue deflation.
Accenture's results matter to Indian IT for a specific structural reason: Accenture is the industry's earliest and most reliable indicator of enterprise technology spending 6–18 months ahead. Its bookings and guidance tell you what CIOs are actually committing to, not what they are saying in surveys.
What Accenture's June quarter showed was not just a small guidance cut. New bookings, at $19.3 billion, fell 2% from the same quarter a year ago and about 13% from the record set in the prior quarter. That last number matters. New bookings are Accenture's pipeline. They typically convert to recognised revenue over a 6–18 month window. When bookings fall, revenue follows.
Management lowered full-year local-currency revenue growth guidance to 3–4% from the prior 3–5% range, citing US federal spending cuts and Middle East conflict disruptions. Management attributed roughly 1 percentage point of that headwind to weakness in US federal government spending; excluding federal contracts, the company expects revenue growth of 4–5%.
The translation for Indian IT is direct and painful. ICICI Securities expects Q1FY27 revenue growth for the top-six IT services firms to be in the range of minus 0.9% to plus 1.1% quarter-on-quarter in constant-currency terms — essentially flat to mildly negative. This is the quarter that TCS kicks off on July 9.
JP Morgan's forecasts for Q2FY27 are even more sobering: Infosys at 1–2.5% sequential growth, HCL Tech at 1–3.5%, and Wipro at negative 2% to flat.
Source: ICICI Securities estimates, July 2026; JP Morgan estimates
The seasonal context makes this worse. H1 of the Indian financial year — April to September — is normally the stronger half. Clients finalise annual budgets, projects ramp up, and deal conversions accelerate. Analysts at Motilal Oswal expect demand commentary to stay soft in Q1FY27, as macro, AI, and geopolitical overhangs weigh on discretionary spending, with the soft start likely extending into Q2FY27. The H1 tailwind has become a headwind.
Here is the number that should give anyone watching the sector genuine pause: Total Contract Value (TCV) of deals — the headline figure companies announce when they win large contracts — has actually been healthy. Infosys closed FY26 with large-deal TCV of $14.9 billion, with 55% net new. These are not bad numbers.
And yet revenue growth for the full year was a modest 3.1%.
This is the TCV-revenue disconnect that analysts have flagged as a structural warning sign. The gap between deal wins and revenue realisation implies something specific: the delays in converting TCV to revenue stem from incremental macro weakness and higher AI deflation — with margins pressure mounting as companies front-load AI-productivity benefits and invest in AI capabilities to win the next deal.
In plain English: companies are winning deals, but the deals are delivering less revenue per rupee of TCV because AI has shrunk the effort component. Clients are renegotiating. Pricing per hour is falling. The deal size on paper doesn't match the invoice that gets sent.
When AI reduces effort, revenue linked to manpower hours comes under pressure before new AI-led work becomes large enough to offset the decline. Efficiency does not automatically flow to the top line — not yet.
The employment data tells this story more honestly than any analyst note.
In FY22, India's IT sector added 600,000 net jobs. In FY26, it added only 140,000 — an 86% decline in a single cycle. For the first time in the industry's history, revenue growth and employment have structurally decoupled.
TCS, which laid off 12,000 staff in mid-2025, plans to hire just 25,000 fresh graduates this year — compared to an average of around 40,000 per year over the previous three years. Gross hiring across IT firms, which averaged around 230,000 annually for five years, came in at roughly 170,000 in FY26.
Revenue per employee has risen sharply. Infosys now leads at around $63,000 per employee, followed by TCS at $52,000, HCL Tech at $48,000, and Wipro at $45,000 — a clear departure from the historical $40,000–$55,000 range. Fewer people are generating more output.
This might look like a productivity miracle. In one sense, it is. But it also means the sector's traditional route to revenue growth — hire engineers, bill hours — has been quietly retired. The problem is that no comparably scaled replacement has arrived yet.
The correction has not been uniform, and that matters.
Mid-tier IT companies — Coforge, Mphasis, and Persistent Systems — have meaningfully outperformed their large-cap peers, with average revenue growth of around 20% year-on-year in recent quarters, compared to flat to 1.5% for the top-five firms. The reason is structural, not accidental. Midcap IT carries less exposure to legacy managed-services contracts — the segment most directly in AI's crosshairs — and is more agile in pivoting to focused verticals like BFSI (banking, financial services, and insurance), healthcare, and hi-tech.
The large caps are not standing still. TCS has announced HyperVault — a plan to build 1 gigawatt of data-centre capacity to power AI workloads, positioning itself as a "full-stack AI services player from Infrastructure to Intelligence." For context: total Indian data-centre capacity in early 2026 is roughly 1.4 GW. TCS, a services company, is signalling it intends to build infrastructure at near-national scale and monetise AI demand through it.
One useful industry framework splits IT work into three buckets: AI-disrupted work (shrinks at 3–5% CAGR), AI-amplified work (grows in line with the underlying business), and AI-native work (compounds at 25–30%). The whole question for FY27 is whether that third bucket gets big enough, fast enough, to outrun the drag from the first.
From its peak of 46,089 on December 13, 2024, the Nifty IT index is now down 43%, as multiple headwinds converge on earnings expectations and valuation comfort. In calendar year 2026 alone, the index has fallen roughly 32%, against a decline of only ~8% in the broader Nifty 50. With AI-linked automation threatening project duration, pricing, and billable effort, the market is increasingly treating this shift as structural rather than cyclical. The combined weight of the five major IT companies in the Nifty 50 has shrunk to below 7.6%, compared to an era when the group held more than 20%.
The valuation reset has been dramatic. TCS is trading at approximately 18x PE — a significant discount to its historical average of 25–28x. More broadly, the Nifty IT index is trading near its lowest valuation since mid-2020. Even after Accenture's ~18% decline, the top-five Indian IT companies still trade at roughly a 70% premium to Accenture, leaving room for further valuation compression.
This is either a screaming value opportunity or a value trap, depending on which side of the structural argument you believe. Peak impact from AI on IT revenues is expected during FY26–FY28, with recovery projected only from FY28-end or FY29 onward.
Not everyone believes this is structural collapse. There are legitimate reasons for caution about the bear case.
First, some of the macro headwinds are cyclical. Geopolitical disruptions — the Iran–Israel conflict, US federal spending cuts — have knocked enterprise confidence independently of AI. Accenture now expects full-year revenue growth of 3–4% in local currency; excluding the estimated 1% impact from its US federal business, it forecasts 4–5% growth for the year. Indian IT firms have minimal federal exposure.
Second, deflation is not uniform. New AI capabilities command a 30–40% billing premium for AI-skilled engineers. Companies that build and deploy this talent effectively are looking at margin expansion, not just revenue risk.
Third, the TCV pipeline remains healthy. Large deals are still being signed. The conversion problem is real but may be partly a lag — a slower ramp, not a structural closure.
And fourth, if the AI disruption thesis proves overstated, or if Indian IT companies successfully transition to AI-led delivery models, current valuations could look very attractive in hindsight.
Here is the irony that will define the next two years of Indian IT: the sector that made its fortune by being humanity's most efficient provider of technology labour is now confronting a technology that makes labour itself cheaper to replace.
It would be a mistake to file this under emerging-market wobble. The core concern is that AI tools can automate work historically delivered through large offshore teams, putting pressure on pricing and billable hours — directly challenging the labour-arbitrage advantage that underpinned decades of growth for India's software exporters. The same forces are pressing on Accenture, Cognizant, and Capgemini. The Indian majors simply carry the largest, most labour-intensive delivery engines — which makes them the clearest place to watch the deflation play out in real time.
The Q1FY27 results season that starts with TCS on July 9 will not resolve this debate. The numbers will be muted — the forecasts make that clear. What matters are the signals within the numbers: Is AI-native revenue growing fast enough to show up materially in the mix? Are deal structures shifting from time-and-materials billing to outcome-based pricing? Are margins holding even as headcount falls?
Those answers will take quarters, possibly years, to arrive in full. FY27 is the year the math becomes legible — when the new revenue mix is large enough to either validate the bet or expose the gap.
Indian IT spent two decades building an empire on the reliable compounding of human effort. The question it must now answer is whether it can rebuild one on the compounding of intelligence — and whether it can do it fast enough before the old model runs out of runway.