IT Outsourcing on the Brink: Why AI Won't Kill Indian Giants (Yet)

For decades, the business model of the largest IT corporations was built on one simple principle: find thousands of smart people, train them, and sell their time to Western clients. This scheme worked perfectly until large language models learned to write code faster and cheaper than an average junior developer after three cups of coffee. When a person of C.

P. Gurnani's caliber, who led the multibillion-dollar Tech Mahindra for many years, goes into free sailing to create an AI-native platform, it sounds like an official obituary for the old way of doing business. In an interview with Bloomberg, Gurnani directly speaks of the industry facing the most serious challenge in its history, but this does not mean that tomorrow all Indian programmers will be left without work.

Context here is more important than the headlines themselves. The Indian IT sector worth 250 billion dollars is an enormous inertial machine. The problem is not that neural networks write code, but that old workflows have become too cumbersome. Gurnani emphasizes that modern corporate systems resemble ancient cathedrals: they are magnificent, but they cannot be quickly rebuilt to meet the needs of the AI era. This is where the opportunity for transformation lies. Instead of simply adding a chatbot to an existing interface, companies need to rebuild processes from scratch, making AI a foundation, not a decorative finish. This requires colossal infrastructure investments that old players can afford better than bold but poor startups.

Gurnani pays special attention to data security and sovereignty. You cannot simply connect ChatGPT to your core banking system and hope everything will be fine. Real money now lies in creating secure "pipelines" that will allow corporate AI to work with sensitive information without violating laws or allowing leaks. The same Indian giants have a unique advantage—they have spent decades studying how the internal mechanisms of global corporations work. This accumulated experience and understanding of business specifics is their main moat, which neural networks cannot yet overcome through simple token enumeration.

Of course, the transition period will be painful. In the short term, margins will fall as companies will have to spend huge sums on GPU purchases and retraining hundreds of thousands of employees. However, those who survive this digital mutation will stop selling their employees' time. They will start selling results. This is a fundamental shift from a "pay for process" model to a "pay for value" model. Gurnani by his example shows that industry veterans are not going to retire, they are simply switching to faster machines while the old ones haven't stalled yet.

The key point: Indian IT giants will survive only if they stop being staffing agencies and become technology laboratories. Will they be able to change their DNA faster than AI learns to understand business logic without intermediaries?