Opinion | Age Of The Pink Slip: India's Brutal IT Layoffs Are No Longer About Cost-Cutting

On the morning of July 29, a lot of Visa staff in India opened their laptops to an HR email telling them not to come in. Their jobs were gone. Visa is cutting about 2,600 roles worldwide, about 7% of its people, and India, its biggest tech base outside America, took a real hit. By early August, the axe had climbed the org chart to vice-presidents, senior directors, and engineering leads. A US filing on July 31 logged 320 losses at its California headquarters alone. One India team of 28, employees told reporters, lost ten.

Visa's official line is that this is about efficiency and reinvesting for growth, and that AI is only one reason among several. Worth noting the company is making plenty of money and isn't leaving India. This isn't a struggling business trimming fat. It's a profitable one, deciding it needs fewer humans.

Visa is not an outlier either. TCS, the country's largest private employer, said this summer it would let go of about 12,200 managers under a "future-ready transformation". Look at the full year, and the picture is uglier: net headcount down 23,460, ending March at 584,519. The four biggest IT firms, TCS, Infosys, HCLTech, and Wipro, removed a net 9,100 people in the April-June quarter this year. A year earlier, in the same quarter, they had added 22,622. Oracle has cut roughly 30,000 globally, many of them in India. Microsoft has gone back for more. Freshworks dropped 11% of its staff and named AI as the reason. Beneath the announced figures sits what nobody files paperwork for: freshers at Infosys, Cognizant and Accenture whose joining dates keep slipping, benches quietly emptied.

A Very Different Storm

This round of layoffs feels different from 2023. Back then, firms were just unwinding pandemic over-hiring. This time, they are shrinking while the money still rolls in. NASSCOM puts the sector's FY26 revenue near $315 billion, up around 6%, yet headcount rose only 2.3%. Revenue grew three times faster than the number of people earning it. For thirty years, Indian IT ran on one sum: revenue equals people times hours times billing rate. Win a client, hire a hundred engineers. That sum is what's breaking. The firms are shifting to charge for outcomes instead of bodies, and the old edge - paying an Indian company to throw more cheap people at a problem - stops meaning much once one AI agent does the work of ten analysts. Some reckon half a million jobs could go in the next two or three years.

I don't buy the apocalypse, though, and nor should you. The Indian tech workforce actually grew last year, to roughly six million. The pain is landing on particular work: repetitive, process-heavy tasks, and the mid-career engineer keeping some decade-old system alive that a coding assistant can now half-run by itself. Meanwhile, the hiring is real in AI, cloud, cybersecurity, and platform work, and the global capability centres, the in-house tech arms multinationals run out of India, are still adding people.

Look At The Fine Print

TCS was careful to call its cuts a skills mismatch rather than an AI dividend, and that wording matters more than it looks. The trouble isn't machines wiping out work wholesale. It's that the work keeps mutating faster than people can retrain for it. The number that should worry everyone: India may need over a million AI professionals, yet by the government's own estimate, only about 16% of the current IT workforce is genuinely AI-skilled. The bottleneck was never the supply of engineers. It's how few of them are ready for the job that exists today.

Which is where the response has to go, and to be fair, some of it already is. A relook and reassessment is being undertaken at many levels, and two areas that need attention are below.

The first task is retraining people and actually moving them into the roles that are hiring. FutureSkills PRIME, run by the IT ministry with NASSCOM, now claims over 25 lakh registered learners across 3,000-odd courses in AI and cloud. The IndiaAI Mission is pushing AI teaching from undergraduate to PhD level, and the Skill India Digital Hub has pulled training onto one platform. Companies are doing their bit, with close to half of tech workers now saying they get AI training on the job and Infosys running its own programmes. But a certificate is not a career. The real test is whether a laid-off maintenance engineer lands in a GCC or product team, or just collects another badge that changes nothing on payday.

The Fixes Lie Outside The Workplace

The second task is fixing the pipeline before it reaches the workplace, so graduates turn up already fluent instead of needing a retrofit. In July, the government and NASSCOM confirmed they are rewriting the AI syllabus for undergraduate courses, a job expected to take about six months. Get it right, and the gulf between what colleges teach and what employers want starts to close. Get it wrong, and we keep minting graduates trained on last decade's tools for jobs that have already vanished.

None of this makes the layoffs painless. Behind every "restructuring" is a person locked out of their own inbox, a fresher watching a start date drift into the fog. But India is not a passenger here. It has the youngest big workforce on the planet - two in three Indians are under 35 - and the deepest bench of AI skills of any country. The real contest is one of speed: how fast AI swallows the routine jobs against how fast the country can push its people up into the harder ones. It can't slow the first. It can win the second, if it stops treating retraining as a press release.

(Subimal Bhattacharjee is a policy adviser on digital tech issues and the author of 'The Digital Decades: Thirty years of the Internet in India')

Disclaimer: These are the personal opinions of the author