TCS Cut 23,460 Jobs and Says AI Isn't Why. Infosys Added Headcount at the Same Time.

Tata Consultancy Services ended FY26 with 584,519 employees, down 23,460 from a year earlier — the single largest headcount contraction any major Indian IT services company has posted in years, and the number that has anchored practically every "AI is coming for Indian IT jobs" story written this year. It's a real number and a real story. It just isn't quite the story most of that coverage tells.

What TCS itself says happened

TCS's own leadership has been unusually blunt about pushing back on the AI framing. COO Aarthi Subramanian has said directly that the restructuring was not driven by AI, describing it instead as a deliberate move to build a "future-ready workforce" — tighter bench policies, voluntary exits concentrated in mid-to-senior roles, and a company-wide shape correction after a long stretch of hiring ahead of demand. The company has also said the cycle is now complete: it has already rolled out 25,000 campus offers and remains on track for around 40,000 fresher hires this year, which is not the hiring posture of a company that believes AI has structurally shrunk the amount of work it needs people for.

At the same time, TCS is not quiet about AI's growing role in its business. Annualised AI-linked revenue rose to $2.3 billion by the end of FY26, up from $1.8 billion in the December quarter alone — real, fast growth in AI-driven work. Both things are true at once: AI revenue scaling quickly, and the company's own account of why headcount fell being something other than AI displacement. That combination is easy to read as spin. It's also exactly what you'd expect if AI is changing which skills a services company needs rather than how many people it needs in total — a much less dramatic, much harder-to-headline story than "robots took the jobs."

Infosys ran the same playbook and grew

If AI were mechanically cutting headcount across Indian IT the way the TCS number gets used to suggest, Infosys should show the same pattern — it competes for the same clients, sells the same category of AI-augmented services, and has talked about AI adoption just as publicly as TCS has. Instead, Infosys closed FY26 with 328,594 employees, up from 323,578 a year earlier: a net addition of roughly 5,000, alongside more than 20,000 fresher hires and voluntary attrition falling to 12.6% from 14.1% a year prior. There was a dip within the year — headcount fell by over 8,400 in the final quarter alone — but the full-year number still landed positive. Two companies selling into the same AI-disrupted market, in the same fiscal year, produced opposite headcount outcomes. That gap is the actual story, and it's a company-specific one about workforce shape and prior hiring pace, not a sector-wide verdict on what AI does to IT employment.

The rest of the sector doesn't agree on a direction either

Zoom out past these two companies and the picture gets messier, not clearer. Across the six largest listed Indian IT services firms, headcount data for the year shows a small net decline overall — but it's an average of divergent company-level stories, not a uniform trend line. Some firms added people through the year, some cut, and the ones cutting weren't all citing the same reason TCS gave. That inconsistency is itself informative: if AI adoption were mechanically forcing headcount down across the industry on a predictable schedule, you'd expect the six majors to move roughly together. They haven't. What they do agree on, separately, is where they are in AI adoption — and it's earlier than the "AI is already reshaping jobs" framing implies. Industry surveys this year have put roughly 60% of organisations still at the pilot or exploration stage of enterprise AI adoption, despite two-plus years of investment. Most of the sector is still figuring out what AI-augmented delivery looks like at scale, which cuts against the idea that a mature, sector-wide displacement wave already explains a given company's FY26 headcount number.

The gap that actually matters isn't headcount — it's skills

A joint Deloitte-Nasscom report, Advancing India's AI Skills, frames the real pressure point differently, and more precisely: demand for AI-skilled talent in India is projected to roughly double, from the 600,000–650,000 range up to more than 1.25 million between 2022 and 2027, while the underlying AI market itself grows at 25–35% a year. That is not a story about fewer jobs. It's a story about a supply of AI-ready talent that is growing more slowly than the demand for it — which means the practical risk for an individual IT employee isn't "AI eliminates my role," it's "the market increasingly wants a skill set I don't have yet, and the company doing the hiring or firing gets to decide which employees clear that bar." TCS trimming a broad base of mid-to-senior generalists while ramping AI-linked revenue, and Infosys simultaneously hiring, are both consistent with that read: it isn't that AI needs fewer people, it's that it needs different people, and which company you work for determines whether that shows up on your resume as a layoff or an internal reassignment.

That reframe matters for how anyone in or around the sector should actually respond to headlines like the TCS number. Treating a 23,460 headcount drop as proof that AI is shrinking Indian IT overall misreads a company-specific restructuring as an industry verdict, and it also misses the more useful, less comfortable finding underneath it: the gap the data actually supports isn't between "AI jobs" and "no jobs" — it's between roughly 600,000 AI-skilled people today and a market that will want double that within a year of 2027, in a workforce of several million people who are not yet counted as AI-skilled at all. Whether that gap closes through retraining, through fresher hiring pipelines like the ones both TCS and Infosys are still running, or through outright attrition of the wrong skill sets is the story worth actually tracking — not whether one company's single-year headcount number went up or down.