New Delhi, Nov. 2 -- degree, job, salary and expenditure. AI has found a chink in the chain.
Consider two numbers from Nasscom's latest Strategic Review. In FY26, India's tech industry revenue grew 6.1 per cent to about $315 billion. Its workforce grew 2.3 per cent, with a net addition of 1.35 lakh people, taking it to 59.5 lakh. Revenue has begun to outrun headcount, and now AI is trying to fill this critical gap.
The big five tell a sharper story. Combined headcount at TCS, Infosys, HCLTech, Wipro and Tech Mahindra fell by 7,389 in FY26, after a net addition of 12,718 the year before. TCS announced about 12,000 job cuts. HCLTech's fresher hiring fell 45 per cent in a single quarter. Year after year, these firms had been the destination for India's engineering graduates. This may no longer be the case.
Why should anyone outside IT care? Because of the identity every economics student learns: aggregate demand equals consumption plus investment plus government spending plus net exports. In FY26, private consumption was 61.5 per cent of GDP, the highest since 2011-12, according to the Economic Survey. Consumption is wages spent. If AI increases what firms can produce without increasing what households can earn, supply grows into a demand shortfall. Productivity that never reaches a pay packet is not prosperity. It is inventory.
Though AI may not be the sole perpetrator, early signs ring alarm bells. The Periodic Labour Force Survey puts unemployment among 15-29-year-olds at 15.9 per cent in April-June 2026, the highest in the revamped series, against 5.4 per cent for all adults. Among young women, it is 19.6 per cent. Skill gaps and weak global demand may also be contributing. Yet these factors shape the market a fresh graduate now enters.
Focus on the other parameters in the equation: investment may skyrocket into data centres, power and chips, yet capital still requires customers. Revenue is the base of government expenditure, but India's income-tax base is narrow, as evident from recent Lok Sabha data highlighting that around 66.5 per cent of income-tax filers had zero tax liability in FY25. Net exports lean on services: software exports rose 8.2 per cent to $221.4 billion in 2025-26, with 54.1 per cent going to the US. This depends on foreign customers looking for Indian talent rather than machine output.
Over the next three years, the damage will probably be felt by freshers in routine coding, testing, support and back-office analysis, translating into slower campus hiring rather than dramatic layoffs. Over three to ten years, entire job families will be reshaped, rewarding those who direct and verify AI. Beyond that, intelligence may become a utility like electricity, and the question will turn to who owns it and what everyone else does with their time.
One of the first places to act would be education. India has over 12 lakh AICTE-approved engineering seats, and engineering enrolment reached 62 lakh in 2023-24. Gross enrolment in higher education is 30 per cent, against the NEP's goal of 50 per cent by 2035. A model built for scarce information has been scaled up; what happens if information becomes free? If an infinitely patient tutor sits on every phone, a college that merely transmits content loses its reason to exist. The purpose of education must shift to character, capability and judgement: working in teams, building things, framing questions, checking claims and owning outcomes. This means a different syllabus: AI fluency, systems thinking, data reasoning, ethics and communication in every stream, not just engineering, with less memorisation. It also means turning colleges into centres of practice, with labs, legal clinics, newsrooms and apprenticeships. It requires teachers to act as mentors, which brings us to an uncomfortable fact. Of 18,951 sanctioned teaching posts in central universities, 4,889, or 25.4 per cent, were vacant as of July 2025. Empty chairs do not provide mentorship. Pricing must follow suit; outcome-linked fees, income-share agreements and modular credentials should be demanded of institutions charging premium fees on implied promises of jobs.
Young engineers should aim to be problem owners, not code producers: pick a domain such as agriculture, health, energy or finance, learn it deeply and orchestrate AI within it. For lawyers, journalists, accountants and arts graduates, the edge is trust. When synthetic content is infinite, someone who can be believed and held accountable becomes scarce. New work is already visible: AI auditors, workflow designers, data stewards, robotics technicians and a vast care economy for an ageing society.
Industry must answer the question it prefers to avoid. If everyone's job is automated, who buys? A firm that automates away its customers saws off its own branch. Companies should fund apprenticeships, co-design curricula and share productivity gains through profit-sharing and shorter workweeks. Demand is a shared asset.
The state has tools. Aadhaar, bank accounts and mobile infrastructure allow India to move money faster than most countries. The Economic Survey of 2016-17 devoted a chapter to a quasi-universal basic income covering 75 per cent of the population but stopped short of recommending it. Economist Jean Dreze noted that this variant would cost 4 to 5 per cent of GDP. The same Survey floated the idea of a citizens' dividend from a sovereign wealth fund. That debate deserves reopening, not as a surrender but as a floor bundled with learning credits, health cover and public-service options: a universal basic opportunity. A levy on automation or computing capacity, or a shift of the tax base from labour towards capital and data, should be debated carefully so as not to strangle innovation. Competition rules should prevent a handful of firms from capturing the entire dividend.
The real danger is idleness. Work produces not only income but also self-confidence, identity and belonging. A country this young cannot afford a generation with a bank transfer and no purpose. Channels are needed for purposeful engagement: community care, climate restoration, local infrastructure, sport, the arts, civic service and micro-capital for small enterprises. There should also be a debate about the concentration of AI power, deepfakes that corrode trust, skills that atrophy through over-reliance, and a world split between countries that own computing infrastructure and those that rent it.
The ministries of education and skills should fund curriculum redesign and faculty hiring now. Universities should pilot outcome-linked degrees. Industry should commit to apprenticeships. Parliament should open the debate on income floors and automation before the shock, not after.
India's demographic dividend was always conditional. It will translate into Viksit Bharat only if young people are equipped for the economy that actually arrives.
Views expressed are personal. The writer is a public policy expert based in New Delhi
Published by HT Digital Content Services with permission from Millennium Post.