Indian IT: A few will thrive, while others will bleed

Summary: AI is reshaping Indian IT, but not in the way most investors fear. The real story lies beneath the headlines, where shifts in business models are quietly redefining the sector. Understanding this change may be key to navigating what comes next.

In February 2026, a Substack essay and a software tool release wiped Rs 5.6 lakh crore off Indian IT stocks in a single month. The Nifty IT index fell 19.5 per cent, its worst month since September 2008. Foreign investors pulled out $1.85 billion from the sector in February alone.

The panic was real. The question raised was wrong.

The market asked: Will AI kill Indian IT? It won't. The right question is: Which IT companies are already adapting, and which are defending a dying billing model? Your returns over the next five years depend entirely on your answer.

The doom chorus

The sell-off had four triggers, each amplifying the next.

A research firm called Citrini published a fictional memo from 2028, imagining how AI dismantled the outsourcing industry. It was a scenario, not a forecast. But markets didn't care. Nifty IT fell over 5 per cent the next day, its worst single session since March 2020.

Days later, Jefferies downgraded six major IT stocks and slashed price targets by up to 33 per cent. Its core argument: AI would shrink managed services, which account for 22 per cent to 45 per cent of revenues at large IT firms. In a worst case, valuations could fall a further 30 per cent to 65 per cent.

At the India AI Impact Summit, venture capitalist Vinod Khosla declared that IT services and BPO would "almost completely disappear within five years." Anthropic released AI tools targeting the exact work Indian IT bills consultants charge Rs 12,000 an hour to perform.

The narrative was overwhelming. But narratives have beneficiaries. As Mitchell Green of Lead Edge Capital observed, all these venture capitalists must portray software as dying because they need to justify how much they are spending on AI companies. The loudest doom voices have billions deployed in AI ventures. Their warnings are driven more by vested interests.

What the bears get right

Intellectual honesty demands admitting what is real.

The man-day billing model is dying. TCS grew revenue while its headcount barely moved. BPO hiring collapsed from 1.3 lakh in FY23 to under 17,000 annually. Wipro itself guided a 15 per cent to 20 per cent revenue reduction from entry-level IT services. These are not projections from outsiders. They are management's own numbers.

Margin compression is underway. HSBC estimates AI will cause 14 per cent to 16 per cent gross revenue deflation across the sector as a result of a reduction in deal values. Applied to $220 billion in annual exports, that is Rs 2.5 to 2.9 lakh crore of value that simply evaporates.

And the productivity paradox is real. If AI enables one engineer to do the work of five, the client will not pay for five. Revenue per project falls even as capability rises. Indian IT has never faced a technology that simultaneously makes it better at delivery and worse at billing.

Indian IT will not crash with headlines. It will erode silently, same revenue but lower hiring, same clients but tighter contracts, same guidance but weaker pricing power. No panic. Just gradual compression that feels stable until it doesn't.

Blockbuster had customers. Nokia had distribution. They did not collapse overnight. They became irrelevant slowly.

What the bears get wrong

The doom case rests on a premise that is roughly 15 years out of date: that Indian IT's entire value is built on cheap labour. It is not.

Enterprise IT is not a coding problem. It is a trust and accountability problem. AI can write code. It cannot take accountability when a system fails at 2 a.m., and a regulator demands answers. As Green put it, the competitive advantage of a software company was never about R&D. It is about distribution, customer relationships, and institutional knowledge.

Workday retains 98-99 per cent of its annual revenues. Its enterprise clients spent three to five years implementing the software. If they are not going to rip that out, why would they rip out TCS or Infosys from equally embedded, mission-critical systems?

The partnerships tell the story. Anthropic, the company whose tool release triggered the sell-off, partnered with Infosys to deploy its AI models into enterprises. OpenAI partnered with TCS. AMD partnered with TCS to build AI infrastructure. Anthropic opened its first India office in Bengaluru, noting India is its second-largest market after the US. If Indian IT were about to be destroyed, the supposed destroyers would not be signing up as partners. They need services firms to reach enterprise clients. The AI labs build models. Indian IT deploys them into messy, complex, regulated environments.

And there is a precedent that the doomsayers ignore. In 1999-2000, every analyst said big-box retail was finished. The top e-commerce companies in America today, after Amazon? Walmart, Home Depot, Lowe's, Target. The incumbents adopted the technology and used their distribution advantage to dominate. The base rate for new technology destroying well-capitalised, actively adapting incumbents is actually very low.

Nasscom's Srikanth Velamakanni reframed the COBOL threat as an opportunity at the 2026 Leadership Forum: 800 billion lines of legacy code at even Rs 170 a line translates into a Rs 1.3 crore services opportunity. AI does not eliminate this work. It makes previously unviable modernisation projects commercially feasible. More work, not less.

The split

The honest assessment is this: the sector will not go to zero. But it will not reach Rs 25 lakh crore on the old model either.

What is actually happening is a value migration within the sector. From labour arbitrage to AI-embedded, outcome-based delivery. The total addressable market probably does expand. But the companies that capture the expanded opportunity will be fewer, leaner, and more differentiated.

Two or three Indian IT companies will be dramatically more valuable in five years. The rest will slowly bleed. The Nifty IT index averages the winners with the losers, which makes the sector-level trade a trap. This is a stock-picking problem, not a sector-allocation problem.

How to tell which side a company is on

Three tests separate the future winners from the slow bleeders.

First, is revenue decoupling from headcount? If headcount stops being the primary driver of revenue, then it could mean that the model is transforming. The traditional Indian IT model scales with headcount and depends on it purely. If this dependency continues, it should be a warning.

Second, does the valuation offer a margin of safety? The Nifty IT index trades at 20.6 times earnings, well below its five-year average of 29.2 and its cheapest since July 2020. The weighted return on equity is 26 per cent. Free cash flow yield is 5 per cent. Dividend yield is 3.5 per cent. These are value-grade numbers for an industry that has never posted an annual decline in 24 years. But being cheap alone is not enough. The company must be cheap and transforming.

Third, is the company already billing for outcomes? Is AI-led revenue a measurable share? Are AI-native service lines (where AI itself is the product) in production rather than just news flashes? Is AI more than just a cream layer in their core product and has a significant impact? Is the company winning contracts structured around delivered results rather than deployed headcount? The answer to this question separates genuine adaptation from marketing.

The opportunity

The last time Indian IT was this cheap was July 2020. Investors who bought the sector then earned over 80 per cent in the following 18 months. Today, the pessimism is deeper, the valuations lower, and the fundamental case for the adapting companies arguably stronger.

As Parth Shah of DSP Mutual Fund observed, lower valuations, higher return on equity, higher free cash flow yield and higher dividend yield are all in favour of IT companies. What is not in favour is the narrative. And good businesses are not available at these valuations unless there is deep pessimism against them. The courage to go against the narrative is rare.

At Value Research Stock Advisor, we have applied this three-part framework to every major Indian IT company. We have found one, a mid-cap, not a mega-cap stock, that clears all three tests. It is already billing for outcomes, not hours. Its revenue is decoupling from headcount. And it trades at a valuation that prices in the pain but not the opportunity ahead.

We will reveal it in the upcoming Stock Advisor Live session on Saturday, April 11, 2026, at 12:30 pm.

Explore Stock Advisor today