NextFin News - Artificial intelligence is set to cut the value of traditional IT services work by as much as 25% over the next four years, the chief executive of Indian technology-services firm Hexaware Technologies said at the company's investor day in Chennai. The warning, delivered by CEO R. Srikrishna at Hexaware's AI Day 2026 on August 21, lays bare the central paradox of the industry's AI moment: the same technology that is opening hundreds of billions of dollars in new opportunity is also deflating the revenue base that pays for it. Management said some of that deflation is already visible in its current revenue, and it is betting the company can grow into the gap by selling AI-led services that replace the very work that is disappearing.
The 20-25% figure is the most precise public estimate yet of AI's revenue impact on the Indian IT services model, and it comes from a company that has moved faster than most peers to embrace the technology. More than half of Hexaware's revenue is now AI-infused, its second-quarter revenue grew 6.1% year over year to USD 405.4 million, and its shares rose as much as 8% in the days after the event. If a company this far along the AI curve is still calling out a quarter of value erosion in its core business, the deflation is not a distant risk. It is arriving inside the current revenue base.
The Deflation Number and What It Covers
Hexaware's management expects 20-25% gross deflation in traditional IT services over the next four years, according to the company's AI Day presentation and subsequent executive interviews. The distinction matters: this is not a forecast that revenue will fall by a quarter. It is a statement about the unit economics of the work itself. When AI tools write code, test it, run operations and close support tickets, the labor content of each contract falls, and the price a client is willing to pay for a given scope of work falls with it. In an industry that has billed by the hour for three decades, lower effort per unit of output translates directly into lower revenue per unit of output.
The company has been unusually explicit about which parts of its own business are exposed. Under its "Zero Friction Enterprise" framework, launched August 17, Hexaware groups its offerings into six pillars: Zero Vulnerability, Zero Tech Debt, Zero Backlog, Zero Defects, Zero Tickets and Zero License. Srikrishna said Zero Tickets and Zero Defects are "deflationary" - reducing revenue from existing services - while Zero Backlog is partly deflationary because AI reduces the work required even as it may unlock higher volumes. The remaining pillars represent new addressable markets. The framework is delivered through an AI layer the company calls Zerovity, and it is underpinned by a data-readiness, security and governance foundation called Infinite Trust.
That candor is unusual for an industry that has spent the past two years framing AI almost exclusively as a growth story. Most large Indian exporters have told investors that AI will expand the total addressable market faster than it erodes legacy work. Hexaware is saying something different: the erosion is real, it is measurable, and it starts now. Management told investors it has identified more than USD 300 billion in net new AI addressable market across six opportunity areas - reducing technical debt, cutting vulnerabilities, lowering SaaS license costs, and AI work around data, governance and cybersecurity - and that it expects to offset the deflation by scaling into those areas.
Why the IT Services Model Is Structurally Exposed
The root of the problem is the business model, and it is worth stating plainly because the whole edifice rests on it. For three decades, Indian IT services grew by moving human labor from high-wage countries to lower-cost delivery centers, billing largely on time-and-materials or headcount-linked contracts. Revenue scaled with effort. A project that took 100 engineers for 12 months produced more revenue than one that took 50 engineers for six months. AI inverts that relationship: the more efficient the delivery, the less effort - and therefore the less billable work - a contract requires.
This is why the deflation is structural rather than cyclical. A cyclical downturn in client technology spending reverses when budgets recover; orders deferred this quarter arrive next year. What Hexaware is describing does not reverse. Once a support ticket is resolved by an autonomous agent, or a backlog of code is generated rather than hand-written, the human hours do not come back. The productivity gain is permanent, and in a competitive services market it gets passed through to clients as lower prices. The company's own history supports the structural reading: CLSA noted that Hexaware has delivered a 12.5% dollar revenue CAGR over the past 12 years under the current CEO despite deflationary pressure already visible in its core infrastructure-management services business. In other words, the deflation is not new; AI is accelerating a pattern that was already in place.
There is a second structural pressure underneath the first, and it targets a different revenue stream. Software licensing has long been an anchor for services firms that resell, implement and customize third-party applications. Hexaware is now arguing that AI has made it cheap and fast to build custom software that replaces licensed enterprise applications.
"It's no longer expensive or time consuming to build custom software to replace SaaS," Srikrishna said. "We do it in weeks or months."
That is a direct attack on the SaaS subscription model - and on the implementation revenue that rides on top of it. The company is targeting a slice of the roughly USD 900 billion global SaaS market with AI-enabled alternatives built for specific clients rather than sold as horizontal subscriptions.
The timing is not accidental. Management said client spending on AI has moved beyond pilots and proofs of concept to roadmaps with defined returns. The company added that AI tokens are expected to become the fourth-largest line item in client IT budgets by 2027, after labor, software licensing and cloud. That ordering is itself a signal: token spend is rising fast enough to displace established budget categories, and every dollar that moves into tokens is a dollar that clients will scrutinize for productivity gains. Hexaware has responded by attaching token-linked pricing options to new proposals - fixed-cost and gain-share structures that let the company capture part of the savings its AI tools create rather than passing all of it through as lower prices.
The scale of the AI push inside Hexaware is larger than the deflation warning alone suggests. Brokerage analysis of the AI Day presentation put the company's AI adoption at more than 900 agents in production, with over 70% of delivery AI-assisted, more than 46% of the client base already delivering AI, and over 50% of active engagements AI or AI-infused. Motilal Oswal Financial Services noted that Hexaware is testing eight commercial models built around token costs, a sign that pricing - not just technology - is where the battle for the productivity surplus will be fought.
The Counter-Thesis: Cheaper Code Means More Code
The strongest argument against the deflation thesis is the Jevons paradox of software: when the cost of producing something falls, demand for it rises, often by more than the efficiency gain. If AI makes coding nearly free, enterprises will build far more software, integrate far more systems, and maintain far more workflows than they do today. The labor content per unit falls, but the number of units explodes. On this view, IT services revenue does not shrink - it migrates from routine maintenance to higher-value architecture, integration and governance work, and the total pool grows.
There is real evidence on this side. Hexaware's own numbers show AI-infused work already accounting for more than half of revenue while the top line still grew 6.1% year over year in the second quarter, and EBIT margin expanded 68 basis points quarter over quarter to 13.6%. The company added one customer in the USD 20 million-plus category, taking the total to 16, and top-10 customer concentration held at 35.7%. That is not the profile of a business in immediate shrinkage. Across the sector, large deals continue to be signed, and the industry's largest players have repeatedly said they see AI expanding demand faster than it compresses legacy work.
But the counter-thesis has a timing problem, and that timing problem is the whole story. New AI-led revenue streams take years to build and scale; the deflation of existing contracts happens deal by deal, quarter by quarter, as each contract comes up for renewal or renegotiation. The gap between the two - the "valley" in which revenue from the old model falls faster than revenue from the new model rises - is exactly what Hexaware is flagging with its 20-25% figure. Whether the valley is shallow or deep depends on one variable: pricing power. If Hexaware can capture part of the productivity surplus through outcome-based pricing, the deflation is survivable and may even be margin-accretive. If the market forces the savings through as lower prices, margins compress even as productivity soars.
"In the past, every technology change - whether cloud or ERP - meant you retained your people, formed the right partnerships and moved ahead," Srikrishna said. "That playbook will not work anymore because everyone in the industry will do that."
That is an acknowledgment that the usual adaptation strategy - retrain the workforce, sign the platform partnerships - is now table stakes, not differentiation. When every vendor can do the same thing, the benefit accrues to the client, not the vendor.
What the Market Is Pricing In
Investors initially responded to the strategy. Shares rose as much as 8% to an intraday high of Rs 574.85 on August 24 before closing up about 5% near Rs 559, with volume running well above recent averages. But the rally sits on top of a weak long-term trend: the stock is down roughly a third over the past year, with a 52-week range of Rs 400.20 to Rs 829.40, reflecting the broader pressure on mid-tier IT services names.
Brokerage reaction has been mixed, which is itself informative. CLSA rates the stock outperform with a target of Rs 730, saying it is excited by the AI strategy and sees a path to 10% year-over-year growth by the end of fiscal 2027 with stable margins. HSBC rates it hold with a target of Rs 575, acknowledging the USD 300 billion identified TAM but citing near-term business weakness. Jefferies rates it underperform with a target of Rs 500. JM Financial reiterated a buy rating with a target of Rs 720, noting that Hexaware has developed 238 AI use cases within healthcare and life sciences alone, is targeting a 30% win ratio, and is aggressively pursuing deals above USD 100 million - while cautioning that "conversion into material revenue and bookings remains a key aspect to monitor." Across the analyst community, views span from about 6% downside to as much as 37% upside from recent levels.
The dispersion is a market saying it does not yet know how to value a services firm in transition. The old multiples were built on steady, labor-linked revenue growth. The new model - outcome pricing, token-linked fees, AI-native delivery - has no clean comparable, and the transition period will show up as noise in the numbers before it shows up as clarity.
Who Else Is Exposed
Hexaware is the first to put a number on the deflation, but it is not the only company exposed. The mid-tier Indian IT services firms - those with heavier concentrations of application maintenance, infrastructure management and testing work - carry the most direct risk, because those are precisely the workloads that AI automates most readily. The larger exporters face the same physics but have more room to absorb the shock: deeper client relationships, larger deal pipelines, and more ability to reprice work on an outcome basis.
The exposure is not limited to India. Global IT services and consulting firms that built revenue pools around manual testing, help-desk operations, legacy application maintenance and custom integration work face the same unit-economics inversion. What differs is the speed of recognition. A company that bills on time-and-materials feels the deflation immediately; a company that bills on fixed-price outcome contracts can capture the productivity gain as margin - at least until the contract comes up for renewal and the client demands the benefit.
There is also a second-order exposure that few are pricing: the SaaS vendors themselves. If AI makes custom-built software cheap enough to replace horizontal subscriptions, the companies selling those subscriptions face a different kind of competition - not from another vendor, but from their own customers. Hexaware's bet on the USD 900 billion SaaS market is a bet that enterprises would rather own a tailored system than rent a generic one, provided the cost and time of building it collapses. That is a threat to software revenue, not just services revenue, and it is why the deflation Hexaware is describing may be the first visible ripple of a much wider repricing.
What to Watch Next
The deflation call is testable, and it does not require four years to validate. Over the next four quarters, three signals will show whether Hexaware is navigating the valley or falling into it. First, revenue growth in the traditional services lines - particularly infrastructure management and application maintenance - should decelerate or turn negative in constant currency if the 20-25% deflation is real; management has already said part of the impact is in the current base. Second, the mix of AI-infused revenue should keep climbing above the current "more than half" while EBIT margin holds near the 13-14% range; if mix rises but margin falls, the company is giving the productivity savings away. Third, large-deal conversion - total contract value wins and the speed at which pilots become roadmaps - will show whether the USD 300 billion opportunity is converting into signed work.
The base case is that Hexaware grows through the transition but at a lower margin structure than the industry's historical norm, with the deflation offset gradually rather than all at once. The upside case is that outcome-based and token-linked pricing let the company capture a meaningful share of the productivity surplus, turning deflation into a margin story. The downside case is that competitive pressure forces the savings through to clients faster than new AI revenue scales, compressing margins for several years and leaving the stock re-rated lower.
There is one falsifying signal worth naming explicitly. If, over the next four quarters, Hexaware's constant-currency revenue growth holds above 8% while EBIT margin expands beyond 15% - growth and margin rising together - the deflation thesis as stated is wrong, because it would mean the new AI-led revenue is scaling faster than the old work is eroding, with pricing power intact. That outcome would suggest the Jevons paradox is winning, not the deflationists.
For the wider IT services industry, Hexaware's candor is a warning that the AI transition will show up in the revenue line before it shows up in the profit line, and that the companies which survive it will be the ones that learn to charge for outcomes rather than hours. Productivity is the promise of AI; deflation is the bill. The question for every services CEO is not whether the bill arrives - it is whether they have built a business that can pay it.
Explore more exclusive insights at nextfin.ai.
