When Akanksha checked her email on the morning of March 31, she found a message from Oracle saying there would be an important update. Within an hour, her Slack access was cut. She was 24, three years out of IIT Roorkee, the only developer working on a client-critical feature. She had assumed indispensability would be protection. It was not. She was one of an estimated thirty thousand Oracle employees cut globally in a single day, part of a restructuring linked to funding AI data centre investments.
India's IT sector has absorbed difficult cycles before: the dot-com crash, the 2008 downturn, the post-pandemic correction. Workers and analysts have learned to distinguish cyclical pressure from structural shift. The current moment is generating genuine disagreement about which kind of event this is, but the evidence for a structural element is accumulating in specific ways that are worth looking at clearly.
Start with the numbers. India's top five IT companies, the foundations of a three-hundred-billion-dollar outsourcing industry, recorded net hiring of just seventeen employees in the first nine months of FY26. The year before, the same figure was nearly eighteen thousand. TCS cut 2 per cent of its workforce in 2025. Wipro entered this year with no fresher hiring target. Infosys terminated hundreds of trainee contracts. The staffing firm Quess Corp estimates that fresh graduate hiring at the top ten IT services companies has roughly halved in two years, from around two hundred and sixty thousand per year to one hundred and twenty-five to one hundred and fifty thousand.
The most telling pattern is where the cuts concentrate. Research from King's College London, examining UK firms with high AI exposure, found junior positions fell 5.8 per cent while total employment fell 4.5 per cent: displacement is hitting the bottom of the career ladder hardest. Stanford's Digital Economy Lab found US employment for young software developers down nearly 20 per cent between late 2022 and September 2025. The explanation from practitioners is concrete: an experienced engineer with an AI coding assistant can be five to ten times more productive than before, which means companies can meet the same output with far fewer entry-level hires.
This matters for more than the immediate employment numbers. Entry-level roles in software are not just jobs. They are how technical judgment gets built, how engineers learn to read other people's code, navigate large unfamiliar systems, and develop instincts about when something is likely to fail. An industry that stops bringing in large numbers of junior developers is quietly dismantling the training pathway that produces its future senior engineers. The effects of this will not show up on a balance sheet for several years.
The anxiety this creates is not only economic. Senior engineer Kailash Raj, speaking to The News Minute, described receiving an AI coding licence that turned days of test-writing into minutes. The productivity gain is real. So is what follows: "It makes me feel less important, less mission-critical, and kind of useless." He upskills constantly and still describes backend developer roles as likely AI-replaceable within four to five years. That is not the voice of someone facing a manageable transition; it is the voice of someone who has done the maths.
Japan offers a view of where the adjustment process leads. A Tokyo Shoko Research survey of over six thousand Japanese companies found that nearly half of large firms expect to shift personnel or restrain hiring as a result of AI adoption, up significantly from a similar survey just eight months earlier. Kazuhiko Toyama of Japan Platform of Industrial Transformation identifies two groups likely to retain value: managers who can use AI effectively to make high-quality decisions, and workers in roles requiring physical presence and human interaction. The large middle layer of rules-based knowledge workers faces the most pressure. That middle layer, in India's IT context, is exactly where much of the sector's employment has historically sat.
The honest answer to whether current disruption is permanent is that some of it clearly is, and distinguishing it from the cyclical component in real time is genuinely difficult. The macro conditions and pandemic over-hiring explain a portion of the current picture. But the entry-level hollowing, the specific pattern of AI tools enabling fewer workers to produce the same output, is not a correction. It is a change in the production function. Whether new roles emerge fast enough, and for enough people, to absorb the displacement is an open question. The answer will depend on political and policy choices that are still being made slowly, while the technology moves quickly.