For more than three decades, India has built one of the most successful engines of middle-class expansion in modern economic history through large-scale knowledge industrialization. Millions of engineers maintain systems, process tickets, test code, and keep the world’s business software running. If China is the world’s factory, India has become its back office.
This arrangement worked amazingly well because the global economy rewarded implementation. Follow the process. Reduce contrast. Provide predictability at scale.
Collapse of the employment pyramid
Entire ecosystems have sprung up around this account: engineering colleges along highways, apartment economies in Bengaluru, training centers promising jobs. Parents who once dreamed of government jobs now dream of their children taking up IT specialization. Stability has gained a new definition: entering a large organization, rising through the hierarchy, avoiding unnecessary risks, and building a life around predictable increments. It was rational advice.
Read also | Building a governance framework for artificial intelligence
Artificial intelligence is now weakening the economic logic behind this social contract. This is because AI is already pressing the need for scalable cognition itself. Indian IT became globally dominant because organizations needed armies of engineers to customize and support sprawling software architectures. Generative AI specifically attacks this layer.
An experienced engineer working alongside AI tools can increasingly perform work that previously required teams under his leadership. Testing, documentation, and code migration started to get stressful. When one engineer plus AI can do the work of five, the hiring pyramid collapses. The “campus situation” for beginners – the passport to the middle dream – is what is stifled first.
However, everyone in the ecosystem believes that this will happen to others, but not to us. Think of an engineer I know in Bengaluru. He’s paid a bomb to preserve ancient technology because people with his specific expertise are on the verge of extinction. His organization will do anything to keep him. It is believed that it is indispensable. What he doesn’t know is that his future is already set elsewhere.
He will not be sacked because he lacks brilliance; But because he is exceptionally good at what he does. His problem is that his technological world is shrinking. Every additional year spent deepening this ancient discipline reduces the number of hours it is exposed to the contemporary and adaptive technologies that are replacing it. His brilliance is used to consume his time, leaving him no bandwidth to reinvent himself.
What is tragic is his refusal to look at the evidence. Across the ecosystem, there is an active denial of reality, even as major Indian tech companies quietly shed thousands of roles and tech forums are abuzz with panic over impending restructurings at global giants like Oracle. But for those inside the bubble, the noise is muted.
Shift in center of gravity
I’ve seen this course earlier. Years ago, companies paid huge amounts of money to programmers who understood COBOL systems because global banks still relied on them. For a while, they seemed indispensable and basked in the spotlight. Then the economic center of gravity shifted. The world simply stopped building its future around COBOL and took some time to migrate. The tragedy was not that these engineers lacked talent. The tragedy was that temporary scarcity created the illusion of permanent importance.
Read also | Raghuram Rajan warns that the AI craze may ignore major risks
Recently, Bengaluru-based startup mentor at Google, Shrinath V, made an observation that gets to the heart of this stagnation. He said that India has always chosen the easier path. While the West built deep production capabilities, India mastered services. “Products force customers to adapt to your worldview,” Shrinath said. “Services adapt to what customers want.”
Indian IT has grown by absorbing the complexity created elsewhere. This always meant hiring more people to provide more reliability. So, big contracts mean big teams. AI threatens this arbitrage because the personalization itself has become dramatically cheaper.
Not only that. Shrinath suggests we look at our technology culture: follow instructions, reduce uncertainty, and judge a manager by the number of employees he controls. This is deterministic thinking. But AI systems are probabilistic. It rewards exploration, synthesis, and adaptive thinking, the same “exploration muscle” that our process-based culture has atrophied. This transformation is more than just a technological transformation; It is a shift in cognitive culture itself.
But we are about to see a change in this hierarchy of value. The premium will be on systems thinking and original abstraction. But millions of highly competent people are not trained in this. Instead, they know what disciplined compliance is at the moment when history wants them to be adaptive.
What is striking is that the old questions are still intact. Parents still wonder which engineering path is “safe”. They see “AI/ML” in the college brochure and treat it like the new Java — a software certification that must be memorized over four years to secure a placement certificate. Meanwhile, mid-level managers still believe that AI will eliminate jobs elsewhere, and technology companies discuss AI through the language of productivity rather than confronting what happens when the economics of headcount themselves begin to collapse.
The machine has already been embraced. The question is whether a society built on the old model fully understands what is about to become obsolete.