Everything About AI
NextFin News - India's information-technology services industry is showing its first credible signs of a turning point, as artificial-intelligence revenue begins to move from boardroom pilots into reported earnings — but the recovery is uneven, and the sector's old growth model is not coming back with it.
HCL Technologies, India's third-largest software exporter, reported fiscal 2026 revenue growth of 3.9% and disclosed annualized Advanced AI revenue of $620 million, while Tata Consultancy Services posted its first annual revenue decline in more than two decades and Infosys cut its forward guidance. The divergence inside a single earnings season is the story: AI is no longer a hypothetical threat to the offshore outsourcing model, it is a line item on the income statement — and it is arriving at different speeds for different firms.
The Numbers: AI Revenue Arrives, But Guidance Is Still Shrinking
HCLTech's fiscal 2026 results, released in April, gave the clearest read yet of AI monetization among the large Indian exporters. Services revenue grew 4.8% year over year in constant-currency terms, led by Engineering and Research & Development services, which expanded 9.8%. New deal wins totaled $9.3 billion in total contract value, and the company guided fiscal 2027 revenue growth to a range of 1.0% to 4.0% in constant currency — ahead of the 0% to 2% range analysts had penciled in after Infosys's warning.
The contrast with the sector's largest player is stark. TCS, the industry bellwether with a market value of roughly 8.95 trillion rupees, recorded its first full-year revenue decline in more than 20 years. Infosys, the number two exporter, crossed $20 billion in annual revenue for fiscal 2026 — reported revenue of $20.158 billion, up 4.6% year over year, or 3.1% in constant currency — but then lowered its fiscal 2027 growth forecast to 1% to 2%, below the 2.03% consensus estimate. Wipro, which reports quarterly rather than annual guidance, projected a first-quarter revenue decline of 2% to 0% in constant currency after its IT services revenue fell 1.6% for the full fiscal year to $10.5 billion.
Read together, the numbers describe a sector that has stopped falling but has not yet learned to run. AI revenue is real and growing; overall growth is still being throttled by clients who are delaying discretionary spending while they reorganize their technology budgets around AI. The industry body has valued India's broader technology sector at close to $300 billion, and the four large exporters account for a dominant share of its services exports to the United States and Europe.
Why This Recovery Is Not the Last Three Recoveries
The first question any veteran of Indian IT should ask is whether this is just another cyclical upturn dressed in new vocabulary. It is not. The last three recoveries — after the dot-com bust, the 2008 financial crisis, and the 2020 pandemic — all shared the same transmission mechanism: clients cut budgets, then released pent-up demand into the same labor-arbitrage delivery model, and Indian exporters captured it with more engineers on more hours.
This cycle runs through a different channel. Clients are not simply deferring spend; they are changing what they buy. Generative AI lets enterprises automate the very routine coding, testing, and support work that made the offshore model profitable. That means the recovery in total contract value does not translate one-for-one into revenue the way it used to. HCLTech's $9.3 billion in new deals and Infosys's $14.9 billion in large-deal wins are being won on AI transformation mandates that are priced differently — more outcome-based, more platform-anchored, and slower to ramp.
The evidence that this is structural rather than cyclical sits in the composition of growth. HCLTech's Engineering and R&D services grew 9.8%, nearly double the 3.7% growth in core IT and business services. Engineering services embed software into products and platforms — the part of the stack where AI tooling is being built, not the part it is replacing. Similarly, the technology and services vertical grew 15% in constant currency, while financial services grew 7.5%. The growth is concentrating in the work that builds AI systems, not the work AI displaces.
That composition shift has a hard implication: the industry's revenue per employee, the metric that powered two decades of margin expansion, is under pressure even as deal flow improves. A cyclical recovery would lift all segments together. What is happening instead is a rotation inside the revenue base — the structural leg is AI engineering; the cyclical leg is everything else, and it is still mean-reverting downward. This is why the turning point is real and fragile at the same time.
The Tariff Overhang: The Cyclical Drag That AI Cannot Fix
There is a second reason this is not a clean cyclical recovery, and it has nothing to do with technology. The United States remains the largest market for Indian IT exporters, and Washington's tariff policy has introduced a tax on the very trade model the sector was built on. Companies like Infosys and Wipro have warned that retail and manufacturing clients — the most tariff-exposed segments of their customer base — will be slower to release discretionary budgets.
This matters because it separates the AI story from the macro story. A client can believe in generative AI and still delay a decision if it is waiting to see how tariffs affect its own margins. HCLTech's chief executive, C Vijayakumar, framed the dynamic on the earnings call: the company expects "large opportunities" as clients use generative AI and other technologies to reduce costs in the tariff-driven environment. Cost-cutting demand is real, but it arrives in small, fast deals rather than the multiyear transformations that move revenue needles. The tariff overhang is the reason the sector can be right about AI and still guide to low single-digit growth.
The second-order effect is worth spelling out. Tariffs raise the cost of on-site delivery and client-facing travel, which should, in theory, push more work offshore — a tailwind for the Indian model. But they also compress client budgets, which pushes clients to do less overall. Which effect dominates depends on whether the tariffs are temporary negotiating leverage or a permanent feature of trade policy. The industry is currently pricing the second case, and that is the drag sitting underneath every guidance number.
The Market Already Voted: From AI Victim to AI Beneficiary
Investors spent the first half of 2026 pricing Indian IT as an AI casualty. The Nifty IT index fell roughly 31% between January and June — its worst first-half performance in 18 years. Then the narrative flipped. In July the index rallied nearly 19%, its biggest monthly gain in six years, after several brokerages upgraded the sector and clients began describing AI as a demand driver rather than a demand destroyer.
The mechanics of that repricing are worth separating from the sentiment. The first-order effect is straightforward: if AI increases enterprise technology spending, Indian exporters win share because they deliver it cheaper. The second-order effect is subtler and more important. Hyperscalers and software vendors are the ones absorbing the capital expenditure on AI infrastructure; Indian services firms are the ones monetizing the integration layer on top of it. That is a lower-risk, lower-capital position — but it also caps the upside. The market's rotation from semiconductors back into Indian IT services in July was, in part, a flight from the balance-sheet risk of the AI buildout to the fee income of the AI rollout.
Analyst estimates put SaaS-related work at 10% to 25% of revenue for some Indian IT firms — a range wide enough to show that the market is still guessing at the actual exposure. That uncertainty is why the rally has been led by guidance beats rather than valuation expansion. HCLTech's stock jumped 8% on its results day, its best single session since September 2020, while the Nifty IT index rose 4.4% and the benchmark Nifty 50 gained just 0.2%.
"While putting the fears of a washout FY26 to rest, it implies HCLTech would outperform both TCS and Infosys at the upper of its guidance range," analysts said after the HCLTech results.
As of October 27, the dispersion remains visible in the tape. TCS trades at 2,473.70 rupees, down 2.42% on the day, still well below its 52-week high of 3,350 rupees. Infosys is at 1,141.80 rupees, down 2.08%, against a 52-week high of 1,728. Wipro sits at 186.12 rupees, down 1.63%, and HCLTech at 1,397.70 rupees, down 1.36%. All four are down sharply from their peaks — the market is rewarding AI execution selectively, not re-rating the entire sector.
The Counter-Case: What If AI Shrinks the Pie Before It Grows It?
The strongest argument against the turning-point thesis does not come from the laggards; it comes from the arithmetic of the model itself. Indian IT services grew for 30 years on labor arbitrage: billing clients for large teams of engineers doing work that could be done more cheaply in Bangalore than in Boston. Generative AI attacks exactly that margin. If a client needs 30% fewer developers to build and maintain the same software, then even a successful AI transformation reduces the revenue base the exporter can bill against.
This is not a fringe view. It is the logic behind TCS's first annual revenue decline in more than two decades and behind Infosys's decision to cut guidance even after crossing $20 billion. The counter-thesis holds that AI bookings and AI revenue — HCLTech's $620 million annualized run rate — remain too small to offset the deflation hitting the much larger maintenance and application-management book. Until AI revenue scales to a meaningful share of total revenue, the sector can grow its AI line and still shrink overall.
The falsifying signal is quantifiable. If AI-related revenue across the large exporters stays below roughly 5% of total sector revenue through fiscal 2027 while operating margins compress by more than 100 basis points, the turning-point thesis is wrong: AI would be a revenue substitute, not a revenue additive. HCLTech's $620 million annualized against a services base that runs to tens of billions of dollars is the starting point for that math. Conversely, if AI revenue doubles from current disclosed levels while operating margins hold above 20% — where Infosys closed fiscal 2026 at 20.3% — the structural-shift view wins.
What Comes Next: Three Time Horizons
Short term (next two quarters): volatility dominates. Clients are still reassessing discretionary spending in response to U.S. tariff policy, and the largest market for Indian IT remains exposed to Washington's trade stance. Wipro's guidance for a first-quarter revenue decline of up to 2% is the clearest signal that the demand recovery has not yet cleared the tariff hurdle. Cost-cutting demand is real, but it is the kind of demand that arrives in small, fast deals rather than the multiyear transformations that move revenue needles.
Medium term (fiscal 2027): the divergence between winners and laggards widens. The base case is modest sector growth of 1% to 4%, with HCLTech at the top of the range and TCS and Wipro struggling to return to positive growth. The upside case requires large-deal conversion: Infosys's $14.9 billion and HCLTech's $9.3 billion in new wins must ramp into revenue rather than stall in pilots. The downside case is a tariff shock that pushes U.S. clients to delay decisions into fiscal 2028, extending the guidance reset into a second year.
Long term (structural): the industry that emerges is smaller in headcount, richer in engineering content, and more concentrated in the firms that can sell AI outcomes rather than AI hours. The winners will be the exporters whose AI revenue is embedded in product engineering and platform integration — the segments already growing in the high single digits — not the ones selling generative-AI wrappers on legacy maintenance contracts. This is a different business, not a recovered version of the old one.
The data to watch is simple: quarterly AI revenue disclosures from each exporter, the ratio of AI revenue to total revenue, and operating margins. Deal wins are marketing; revenue recognition is truth. The turning point is not a date on a calendar — it is the quarter in which AI revenue stops being a footnote and starts carrying the growth number.
India's IT exporters are not being disrupted out of existence — they are being re-priced into a different business. The turn has begun, but the companies that make it are the ones that stop selling AI as a service and start selling what AI makes possible.
Explore more exclusive insights at nextfin.ai.
