There is a clear sense of recovery taking shape: TCS' Krithivasan

TCS CEO K Krithivasan talks about his vision of AI-first TCS and shift in its strategy towards acquisition

K Krithivasan, chief executive officer (CEO) and managing director (MD) of Tata Consultancy Services (TCS), believes that calendar year 2026 would be better for the firm on the back of a strong pipeline of deals and artificial intelligence (AI)-led demand. In an interview with Shivani Shinde, he talks about his vision of AI-first TCS and shift in its strategy towards acquisition. Edited excerpts:

 

On the analyst call, you said growth momentum in the second quarter (Q2) has carried into Q3. Do you expect that momentum to sustain in Q4, especially given that growth in international markets still appears subdued?

 

If you look closely at the numbers, the picture is not as subdued as it may appear at first glance. In Q2, international growth was around 0.6 per cent, while it came in at about 0.3 per cent in Q3.

 

However, if you look only at international services revenue, growth was closer to 0.4 per cent in Q3, despite the impact of furloughs during the quarter. Our international revenue growth in Q3 is actually stronger than in Q2.

 

So I would not characterise this as a subdued environment. When you also factor in the nature of deal conversations we are having with clients, there is a clear sense of a recovery taking shape. It is not a sharp or steep recovery, but what we are seeing is a moderate, steady recovery over the medium term.

 

Some would argue that TCS has been a bit late in explicitly calling out its AI-first focus, especially since several competitors have been talking about and working on being AI-first for nearly two years. How do you respond, and what has been happening internally?

 

I would disagree with the view that we have been late. In fact, a recent Everest report places TCS near the top in terms of talent readiness and what we are doing internally on AI. Each organisation also has its own communication style. At TCS, we communicate externally only when there is something material to share and when we are confident about both the direction and the execution.

 

We were among the first to introduce platforms such as WisdomNext as a pilot and bring it to market. We also laid out an early architecture on how organisations can adopt AI in a structured manner.

 

Since the time generative AI gained traction, we have consistently engaged with customers at every summit, discussing how we can help them adopt AI effectively. From that standpoint, we genuinely believe we have been ahead of the curve compared to many organisations, even if we chose to articulate that focus publicly at a later stage.

 

What does it mean to be an AI-first company? 

It has multiple dimensions. The most important one for me is this: in everything we do, AI is the first lens we apply. Whenever we commit to a customer outcome, the starting question is whether we are leveraging AI to drive better productivity, improved customer experience, or faster delivery.

 

The same principle applies to employee skilling, capability building, and internal systems.

 

That is why we articulated our five-pillar strategy — internal transformation, service delivery, workforce, customer value, and ecosystem impact. Across all these areas, the question we ask is whether we are leveraging AI and whether it is the default option. Everything we do, including client engagement, is approached with an AI-first mindset.

 

TCS’ AI focus has also meant being more acquisitive. Your comment? 

We have put out our vision to be the largest AI services company. We also need to look at how we scale it up at speed. When the question of scale and speed arises in building competency, we realise that acquisition is also an important lever through which we can achieve that speed and scale.

 

The headcount drop suggests restructuring impacted more than 2 per cent of the workforce. What’s the update? 

As announced earlier, this exercise is expected to continue until the end of the year. During this quarter, about 1,800 associates were released. We have continued to assess where there is a fit for the future-ready talent pool. Based on that assessment, decisions are being taken, and we expect to close this exercise by the fourth quarter.

 

I do not want to get into detailed numbers again, but to be clear, around 1,800 exits this quarter were directly linked to the restructuring. Beyond that, overall headcount is influenced by several other factors, including the skills we want to hire going forward, natural attrition, and the efficiencies we are driving internally.

 

We have been able to redeploy people more productively, offer them opportunities in new growth areas, and improve utilisation across the organisation. We have continued to hire people. We have added over 16,000 employees in Q3.

 

What has been the progress on HyperVault? 

Following the announcement of the deal with TPG, we have been in very intense discussions with multiple stakeholders. Some of these are prospective partners, while others are hyperscalers, and we are evaluating various options.

 

We are also in advanced conversations with several state governments to identify suitable land parcels for the data centres. In parallel, discussions with the required original equipment manufacturers (OEMs) are underway. Once a deal is finalised and announced, we expect the data centre to go live in about six quarters.

 

For the first gigawatt that we have planned, we will not be looking at additional equity partners. The equity component is already in place, and the remaining funding will be structured through weighted leverage.

 

Can you elaborate on discretionary spending and early budget discussions you are hearing from clients, especially given the ongoing global uncertainty? 

Overall, the demand environment has been improving. Within discretionary spending, we are seeing early signs of movement, though it remains selective rather than broad-based. Projects with a clear business case and a short payback period, particularly those where value is visible quickly, are moving through approvals much faster than before.

 

On the other hand, large discretionary programmes with longer gestation periods or extended payback timelines continue to face deeper scrutiny and longer approval cycles. So while discretionary spending as a whole has not fully opened up, there are definite pockets where activity is improving.

 

On upcoming budget cycles, I have been saying this for the past two years that clients have largely moved away from rigid annual budgeting. While budgets are still being set, they are no longer sacrosanct and are continuously recalibrated based on market sentiment and evolving macro conditions.

 

Decisions to increase or cut spending are now taken mid-quarter or mid-year, rather than waiting for annual reviews. As a result, we do not place too much emphasis on a client’s stated annual budget.

 

What matters more is the underlying bias — whether clients are inclined to invest or remain cautious. That directional intent is far more important than the budget number itself.

 

You also said CY26 growth will be better. What gives you that confidence? 

You have to look at overall performance this year. If you consider just the first three quarters, we have already announced close to $29 billion in total contract value (TCV). With one more quarter to go, assuming another $9-10 billion, we are likely to end the year with a TCV of around $38-39 billion.

 

The second indicator is what we are seeing in client conversations, where revenue visibility continues to improve at a healthy pace. Third, our deal pipeline also looks strong. When you put all these factors together — TCV momentum, improving revenue signals, and a solid pipeline — it gives us confidence that growth in CY26 will be better.

 

Considering AI-led growth is faster, have deal sizes changed? We are yet to see large-scale AI deployments. Your comment. 

Pure-play AI deals may be small, but deals that meaningfully leverage AI can be quite large. When clients look at consolidation opportunities, AI typically cuts across multiple components — improving productivity, enabling coding assistance, automating ticket resolution, or transforming processes such as claims settlement.

 

On large-scale implementation, my view is that we should not become impatient. Technologies like this take time for customers to internalise and clearly understand where the real benefits lie. Most customers are still in the implementation phase — running pilots, experimenting, and gradually moving select use cases into production.

 

The full benefits of AI emerge only when it is deployed end-to-end, rather than in isolated pockets. We need to give this technology another year or two to become more mainstream before judging its full delivery potential.