Is it obsolete or is it operational?

Over the past couple of years, and especially in the recent past, Indian IT is on the back foot, with its back firmly against the wall. Valuations have dropped, margins squeezed, and even toplines rattled. Look at TCS, the most prominent firm. In 2025-26, it reported a more than two per cent decline in annual revenues, the first time since it was listed more than two decades ago. The stock price is down by a third compared to the high in 2025. Since February this year, it shed almost INR 750, or more than 20 per cent. Yet, TCS seems optimistic.

In a media interview, the firm rejected the predictions that the IT sector may become obsolete by 2030 due to artificial intelligence (AI). Its CEO, K Krithivasan, instead claimed that some experts wrote several obituaries against the industry over the past decade. Yet, Indian IT proved to be “resilient,” and thrived rather than survived. This is because it has moved from a mindset of mere cost arbitrage to one where it has acquired a “depth of skill sets” to move on with the times. In the age of AI, clients will need it, either more, or less.

According to a senior TCS executive, there is invariably a gap that exists between tech innovation and adoption, which is true for AI. This presents huge opportunities for Indian IT to help clients to both incorporate and integrate AI into their existing systems. Of course, the latter can do it on their own, given the enormous potential of AI. But this is where past skills and costs will matter. Indian firms will do this faster, more efficiently, and cheaper. If they can incorporate AI in their processes and systems, they will be able to help the clients better. This window, or maybe a door, will continue to exist for years.

Post-adoption, there are questions related to scaling up, extracting value, and keeping up with the constant tech changes and disruptions to decide what is the next best bet or move. This is where Indian IT can help the foreign clients. Just the deployment of AI is not enough. It does not logically and automatically lead to results and outcomes. One needs skills to extricate value out of AI across processes, workloads, operations, and firms’ hierarchies. Different managers and departments need different approaches. This is where external, insightful, and professional advice will matter.

As a media report indicates, TCS views the developments in advanced systems like Claude Cowork and Claude Mythos as “opportunities rather than threats… as generative AI is enabling legacy modernisation, particularly for older systems such as mainframes, allowing companies to reduce technology debt.” In simple terms, users will aim to reduce the huge capital costs, even if they are one-time expenses, to shift to new tech. They will opt for changes to the existing tech to make it modern, and capable of better performances with the right inputs, patches, and band aids to manage the change.

Thus, there will emerge, at least over the next several years, two separate layers of AI adoption. There will be a set of firms, or some departments within a firm, that will co-opt new AI tools immediately in an urgent fashion. Costs will not matter, only changes will. But there will be another set of firms, or departments within, that will do patchwork to modernise and upgrade, rather than shift. The huge costs of new tech will dissuade them. Thus, the gains will matter. Unless they are in the range of 50-100 per cent, they may be side-stepped.

Despite the annual revenue decline, TCS is confident about the future, due to its highest-ever total contract value of nearly $41 billion in 2025-26, and $12 billion in January-March 2026 quarter. “Customers are actually becoming more confident in investing in their projects, and the decision-making cycle is improving,” Krithivasan told a newspaper. Despite the geopolitical conflicts, including the war with Iran, there were no delays in deal closures in the recent past. Large deals were signed in March, despite the war, which showed the clients’ appetite for AI-led projects, which they cannot undertake on their own.

Changes in Indian IT work processes, and internal systems are visible in the hirings and layoffs. TCS claimed that it hired 44,000 freshers, or trainees, in 2025-26, and issued 25,000 offers for the next hiring cycle. Yet, its employee count, at the end of January-March 2026 quarter, was more than 23,000 lower than a year ago. The firm is shedding huge numbers, and hiring maybe lower numbers, but a different set of people. This reflects a churn in the skill sets that may be required in the future. Since outcomes and results will matter, not people deployed in a project, one needs people who can work faster, longer, and have AI skills.

As far as hirings and layoffs go, different IT firms may adopt different strategies. Some may focus on senior and middle-level experience and multi-level management skills to use and deploy AI efficiently. Others may look at freshers, and younger people, who are generally better equipped to handle new tech, especially AI. Some may focus at the cutting and innovative edge, and others at fixing legacy tech. The so-called uniformity in Indian IT, or the seeming-similarities, are likely to vanish within 24 months, if not 12 months. Firms may even experiment with different strategies and visions to see what works for them.

Hence, there will be a churn in projects, clients, and nature of work. Gains in productivity may lead to “deflation” in contracts. This is the most obvious scenario, which will chip at the toplines, and undercut the margins. But benefits at the clients’ end can easily be adopted by Indian IT if it embraces AI in a big way. If the clients understand that AI-led changes can be outsourced, as tech work was, there will be only a changeover pause for Indian IT. This is what happened during the process changes from onshoring, offshoring, to outsourcing.

Of course, knowing what Indian IT does best, it will need to give a sharp, precise, crisp, and attractive branding to it. Outsourcing clicked at the end of the last millennium and the tech crash. Maybe it is merely a case of AI-In, AI-Out IT. AI-In is more expensive, and will need time to re-skill the existing employees at the clients’ end. AI-Out will be cheaper, faster, and more efficient and effective. Or will it?