AI Revenue Is Real. So Is the Problem It's Hiding.

For three years, Indian IT's AI story ran on adjectives, not arithmetic. Every earnings call brought a new platform, a new agent, a new slogan about "AI-led transformation" — and almost no hard numbers to back it up. That changed this quarter. TCS reported a $2.6 billion annualised revenue run rate from AI. Infosys said AI now makes up 8.2% of total revenue. HCLTech's Advanced AI business grew 62% year-on-year to $171 million.

For anyone who has spent the last three years asking Indian IT to show its work, this should feel like vindication. The numbers are finally on the table.

Except the same earnings season also brought weaker guidance, sluggish revenue growth, and visibly more cautious client spending across the board. The AI story got sharper at exactly the moment the growth story got blurrier. That is not a coincidence, and it is not a footnote — it is the actual story CXOs should be paying attention to.

The measurement problem was never the real problem

The old complaint about Indian IT's AI narrative was that it was unfalsifiable — impossible to verify, easy to inflate. That complaint is now, largely, resolved. Companies are disclosing run rates, revenue share, and segment growth with a precision that would have been unthinkable eighteen months ago.

But solving the measurement problem doesn't solve the business problem. It just makes the business problem visible. And what it reveals is a scale mismatch: AI revenue lines are growing fast off a small base, while the core service business — the multi-billion-dollar engine that actually funds these companies — is growing slowly, or not at all. Percentage growth on a small number is easy to produce. It is not the same thing as enterprise-level momentum.

The question CXOs should actually be asking

Boards and analysts have spent the last two years asking, "How much AI revenue do you have?" That question has now been answered, and the honest answer is: not enough to matter yet, on its own.

The more useful question is a harder one: is AI revenue additive to the business, or is it substituting for revenue that used to come from something else? When AI-driven efficiency lets a services firm deliver the same client outcome with fewer billable hours, that shows up as a positive in the AI segment and a drag in the traditional segment — often in the same client relationship, in the same quarter. A headline AI growth number can be real and still not tell you whether the company, in aggregate, is growing.

This is precisely the dynamic we flagged in our earlier analysis of the GCC model — the shift from "how much work did we do" to "how much value did we create" is not confined to captive centres. It is now showing up in the P&L of the service providers themselves, in public earnings calls, in real time.

What this means operationally

For enterprise buyers, this is a signal to change how AI-related deals get evaluated. A vendor's AI revenue growth rate is not a proxy for their ability to deliver your outcome faster or cheaper — ask instead what portion of that AI revenue is net-new work versus repriced existing work, and how pricing models are shifting deal-by-deal, not just in aggregate disclosures.

For GCC and IT services leadership, it's a warning against a specific kind of complacency: treating a strong AI revenue slide as evidence the transformation is on track. The metric that matters is blended growth — AI revenue plus core revenue, together — not AI revenue in isolation. A business can hit every AI KPI on the dashboard and still be shrinking.

For boards, it's a reason to push past the headline number in every AI update. Ask what the AI run rate looks like net of cannibalised legacy revenue. Ask what guidance would look like if the AI segment were excluded. If the answer is uncomfortable, that discomfort is the actual insight — more useful than the number that prompted it.

The real shift

Indian IT spent years being criticised for not disclosing AI revenue. Now that it has, the harder truth is emerging: transparency was never going to be the fix. The fix is building an operating model where AI revenue growth and enterprise growth move in the same direction — not one propping up a narrative while the other quietly slows.

That is a structural challenge, not a disclosure challenge. It's the same pivot we help GCCs, technology services firms, and AI-native enterprises navigate — from reporting activity to owning outcomes.

Neovay works with technology services and GCC leadership teams on exactly this shift. To talk through what it means for your organisation, visit www.neovayglobal.com.