Infosys: The AI Revenue Is Real, but the Growth Guidance Is Not
A $5.2 billion deal quarter and 29% sequential AI growth could not prevent a cut to 1–2%—because AI is now both the new revenue and the deflation.
INFY · Earnings Call · 2026-10-23
The paradox in one quarter
Infosys reported Q2 FY27 on October 23 with revenue of $5.145 billion, up 1.5% year over year and 1.2% sequentially in constant currency, operating margin stable at 21%, and free cash flow of $928 million. On the surface, the quarter had plenty to advertise: large deal TCV of $5.2 billion, 62% net new, including a $2 billion-plus seven-year net-new mega deal and a $1 billion renewal. Yet management cut its full-year growth guidance to 1–2% from the prior 1.5–3.5% band. The culprit was not a collapse in demand but a narrower mix of price, volume, and vertical softness—precisely the kind of slow-burn problem that a strong AI narrative can mask for a quarter or two.
The standout was AI revenue. AI-first work now accounts for 10.5% of total revenue, up sharply from the 5.5% benchmark Infosys first published at its AI Day for Q3 of last year. “AI-first revenue now account for 10.5% of our revenues.” — Jayesh Sanghrajka, Chief Financial Officer · 2026-10-23 And the sequential momentum is unusual for a services business of this size: “Our AI revenue was standout of our total revenue, growing 29% sequentially.” — Salil Parekh, Chief Executive Officer and Managing Director · 2026-10-23 The problem is that 10.5% is still a minority of the base, while the legacy 89.5% is where pricing pressure and productivity pass-backs are doing damage.
The disclosure shift
What changed most versus prior calls is not the existence of AI revenue—it is the willingness to quantify it. In the April 2026 call, Salil Parekh repeatedly declined to give a number. “The growth in AI services is very strong, but we have not disclosed that revenue number externally yet.” — Salil Parekh, Chief Executive Officer · 2026-04-23 When pressed on whether AI had reached double digits, he was even more direct: “Is it 10% or 15%, but we're not sharing the number.” — Salil Parekh, Chief Executive Officer · 2026-04-23 Now the company is sharing 10.5%, and the CFO explicitly anchored the prior 5.5% data point. That is a meaningful disclosure pivot—management wants the market to see the AI ramp as new revenue, not merely a relabeling of existing work.
The revenue of 10.5% AI we are sharing is new revenue.
The claim is plausible: Infosys says AI is embedded in almost every large transaction, and the use cases range from invoice processing to software modernization. But the market will rightly ask whether that new revenue is additive or cannibalistic. The answer appears to be both. AI is creating work that did not exist before, while simultaneously giving clients the leverage to demand productivity concessions on the work that did.
Where the guidance went
The revenue miss was not one line item. CFO Jayesh Sanghrajka laid out three pressures: pricing expansion that did not materialize, softer volume, and slower discretionary spend in retail and high tech. The pricing comment was blunt. “the pricing pressure on the back of competitive intensity, which has increased in last couple of quarters.” — Jayesh Sanghrajka, Chief Financial Officer · 2026-10-23 This is the competitive intensity theme that now sits near the top of Infosys's own keyword board—a phrase that was present before but has moved from background risk to guidance driver.
High Tech was especially weak, with management attributing much of the pressure to one large U.S. client. Retail was hit by lower third-party revenue and softer consumer demand, leading to ramp-downs. Manufacturing faces a large-client issue expected to hit mostly in Q4; the CFO said “the large portion of the impact will come in Q4.” — Jayesh Sanghrajka, Chief Financial Officer · 2026-10-23 That is not entirely new—prior calls flagged European auto weakness—but the timing has been pushed later, which makes FY27 back-half risk more visible.
There is also a self-inflicted component. Infosys chose not to pursue a large European manufacturing deal because the economics did not meet its return threshold. Management estimates that decision, plus lower offshoring revenue, accounts for roughly 1.8–2% of the full-year growth impact. That is disciplined capital allocation, but it does not make the near-term growth optics better.
The deflation flywheel
The most important tension on the call is that AI is now a deflationary force inside the legacy book. Clients are asking for AI-driven productivity benefits to be passed back in pricing, especially on large multi-year programs. Salil Parekh said the productivity expectation is broad-based, and in some cases clients are asking for more than the economic return Infosys is willing to accept. The pricing pressure is therefore not a cyclical air pocket; it is a structural negotiation over who captures AI's efficiency gains.
This has been building for several quarters. In the July 2025 call, Salil described the dynamic plainly: “typically, there are productivity gains, and those are shared between clients and ourselves.” — Salil Satish Parekh, Chief Executive Officer and Managing Director · 2025-07-23 What is new is the quantification. The CFO now says that at the low end of guidance, pricing is assumed to deteriorate slightly: “At the lower end of the guidance, we have assumed similar to slightly deterioration on the pricing.” — Jayesh Sanghrajka, Chief Financial Officer · 2026-10-23
Management's defense is that the legacy book is finite. Contracts below $50 million signed recently already embed AI productivity benefits; larger multi-year deals depend on where they sit in the renewal cycle. Salil argued the deflation is a finite period, while AI growth is secular. That may be true, but Infosys will not quantify how much of the legacy book has already repriced, which leaves investors to trust the shape of the curve rather than the level.
What would make this a turning point
The bull case rests on conversion. Infosys signed $8.8 billion of large deals in H1, up 28% year over year, with 62% net new. The $2 billion-plus mega deal begins transition in the second half of Q4 and is expected to contribute more than 1% to FY28 revenue growth. The Large deal pipeline is described as healthy, with additional mega deals under discussion. If those convert on schedule, FY28 should show an inflection.
But the current guide implies a weaker H2: compensation hikes effective October 26 and January 27, acquisition amortization, seasonally softer revenue, and continued AI productivity sharing. Project Maximus and currency are partial offsets. The stock had already been lifted earlier in October by Accenture's strong quarter, which raised the bar for Infosys's own print. Instead, Infosys delivered a classic services paradox: better AI disclosure, a strong deal book, and a lower growth number. “We feel good that the organic be in a good shape in next year.” — Salil Parekh, Chief Executive Officer and Managing Director · 2026-10-23 The market will need more than feel-good—it will need proof that AI revenue is outrunning the deflation it creates.