Harvard economist Gita Gopinath delivered Davos’s most uncomfortable statistic: since the 1980s, only 30 per cent of India’s growth has come from labour. The rest? Capital.
For a country that spent decades celebrating its demographic dividend, this represents fundamental failure. India hasn’t chosen to be capital-intensive—regulatory friction forced it. Companies opted for machines rather than navigate labour laws.
This isn’t India-specific. It’s pattern recognition playing out globally: capital replacing labour faster than new jobs emerge, skills gaps widening faster than training adapts, regulatory complexity making buying machines easier than hiring people.
Five themes from Davos make clear this is accelerating.
Entry-level jobs are vanishing
Since January 2024, entry-level job postings have fallen nearly 29 per cent globally. Youth unemployment in the US hit 10.8 per cent—double the overall rate. India: 17 per cent. China: 16.5 per cent. Morocco: 36 per cent. In the UK, 1.2 million graduates competed for 17,000 entry-level positions.
This isn’t temporary. The World Economic Forum’s Future of Jobs Report 2025 projects 78 million new roles by 2030 whilst 22 per cent of current jobs undergo structural change. The timeline doesn’t work for people graduating today.
Corporate response reveals the contradiction: 41 per cent plan workforce reductions in AI-exposed roles. Seventy per cent plan to hire for AI skills. Neither helps graduates who can’t get first jobs because entry-level roles are automated whilst AI positions require experience they can’t gain.
Gen Z’s pragmatic response: 37 per cent now pursue blue-collar work. When white-collar entry paths close, you find alternatives.
AI is redesigning work (faster than HR is redesigning roles)
Eighty-two per cent of executives plan to adopt AI agents within three years. These aren’t productivity tools—they’re autonomous systems making decisions, processing information, interacting with customers.
Sunil Bharti Mittal described how AI permeates Airtel: spam detection processing billions of messages, network management, fraud protection. “Agentic AI is pervasive all over.” Some call centre jobs are affected. “Different jobs will need to be found for those.”
That sentence does heavy lifting. “Different jobs will need to be found” assumes someone is finding them. Evidence suggests otherwise.
Union Minister Ashwini Vaishnaw emphasised India’s progress, noting “95 per cent of the work can be done with a small 50 billion parameter model. You don’t require a trillion parameter model.” This makes AI accessible beyond tech giants. But it doesn’t resolve employment—it just makes displacement affordable.
Business process outsourcing faces disruption. Call centres, coding jobs, IT services—all under pressure. Gopinath’s assessment: the pivot “will take directed conscious effort, it’s not going to happen very quickly.”
Companies are automating tasks without rethinking roles, deploying tools without redesigning workflows, replacing people without developing survivors. Gunter Beitinger from Siemens: “The decisive advantage will not come from automation alone, but from redesigning end-to-end workflows around human-AI collaboration.” Most organisations aren’t doing this.
The most uncomfortable Davos truth: productivity is rising while pathways into work are collapsing
The skills paradox is compounding
Ninety per cent of leaders report 20 per cent workforce overcapacity in legacy roles. Simultaneously, 94 per cent face AI-critical skill shortages. One in three report gaps of 40 per cent or more.
Companies have too many people doing obsolete work whilst desperately needing skills they don’t have. The obvious solution—reskill existing workers—runs into implementation reality.
Gopinath identified the deeper problem: “There is a mismatch between what jobs can be created and the skill of the labour force.” This isn’t training gap. It’s fundamental misalignment between how education develops capability and what economies need.
Pearson’s Sulaekha Kolloru called this the “learning gap”—the distance between what AI tools can do and how well workforces use them. “Sustained productivity benefits will come through people’s ability to harness the technology effectively.”
Most organisations claim they’re addressing this. Eighty-five per cent offer upskilling. Seventy-seven per cent provide AI training. These numbers are suspiciously high given 94 per cent report critical shortages. Either training isn’t working, or organisations measure activity rather than capability.
Workers aren’t waiting. LinkedIn data shows AI skills adoption has more than doubled across sectors since 2016. People control their own development because they’ve learnt employers won’t. Women particularly—currently 28 per cent of STEM workforce, 22 per cent of AI professionals—face both challenge and opportunity in this reskilling moment.
AI is not the problem. The failure to redesign work around people is
Leadership development can’t keep pace
IKEA’s Juvencio Maeztu Herrera, describing India entry: “Try to understand from inside and not from outside. You cannot come with your prejudice, your biases and pretend.” His approach: meeting customers on shop floors and in homes, engaging officials personally, understanding context rather than imposing templates.
The uncomfortable parallel for HR: how many CHROs have spent meaningful time on factory floors understanding what skills gaps actually look like rather than what dashboards report?
The leadership bottleneck isn’t about programmes. It’s about developing judgement at scale and speed. Organisations need more leaders making good decisions faster, not more people completing modules. Current approaches optimise for yesterday’s problems: managing predictable complexity, not navigating genuine uncertainty.
Entry-level jobs are disappearing just as organisations claim they can’t find talent
Reform without implementation
Vaishnaw cited impressive progress: 1,600 antiquated laws removed, 35,000 compliances eliminated, telecom permits reduced from 270 days to seven. Mittal, whose career began in 1976, confirmed: “I’ve seen this plethora of books, handbooks, manuals… Much of that is gone.”
But Gopinath’s scepticism: “Is making another committee to get rid of bureaucratic red tape the best way forward?” India remains challenging for business. Land acquisition is messy. Judicial reforms show little movement. Labour market flexibility remains constrained.
Recent reforms raised regulatory thresholds from 100 to 300 workers. But as Gopinath noted, “if you want to be part of big global supply chains, you’re going to have to be way more than 300.”
Reform is happening. Implementation gaps remain. The fundamental challenge—making hiring people easier than buying machines—persists. Until that changes, the 30 per cent problem won’t resolve through regulatory tweaks alone.
Human capacity is being treated as infinite while its costs remain invisible. Davos showed why that assumption is no longer sustainable.
The brain economy makes costs visible
The newest theme: cognitive, emotional, and social capacity are economic assets being systematically underinvested in.
Gopinath made this concrete: India’s pollution costs to GDP exceed any tariffs imposed. Some 1.7 million lives lost yearly—18 per cent of deaths. Cognitive function declines with air quality. Decision-making degrades. Health costs compound.
This extends to sleep, mental health, stress, cognitive overload. Organisations treating human capacity as infinite resource discover costs compound invisibly until they become crisis. Yet most HR metrics track inputs—training hours, engagement scores—rather than outputs like judgement quality and decision speed.
What gets measured gets managed. What doesn’t get measured gets depleted.
From disappearing entry-level jobs to capital-heavy growth, Davos exposed why today’s talent systems no longer work
What this means
The five themes converge on uncomfortable reality: HR’s operating model is misaligned with what organisations need.
Entry-level elimination whilst claiming talent shortages works until experienced workers age out and pipelines are empty. AI adoption without workforce development creates productivity gains that don’t translate to sustainable growth. Skills paradoxes should trigger reskilling but instead trigger layoffs and rehiring—expensive, disruptive, demoralising.
Leadership development can’t scale because systems assume stable contexts. Brain economy underinvestment creates productivity costs that don’t show in quarterly reports but compound into structural weakness.
Davos set the agenda, but the real decisions will be made inside organisations that must choose between redesigning work or replacing people.
The question nobody answered
India will become the world’s third-largest economy. Whether that growth translates into widespread prosperity or scales inequality at higher GDP remains open.
The same question faces every organisation: redesign work to augment human capability, or replace humans with capital and wonder why productivity gains don’t translate to sustainable growth?
When only 30 per cent of growth comes from labour, you’re building an economy that doesn’t need workers. When that happens at scale, social costs eventually exceed economic gains.
The conversations in Davos set agendas. Decisions in workplaces determine outcomes. For HR leaders, the signal is clear: the systems we have won’t solve the problems we face.
Entry-level jobs vanishing, AI adoption outpacing reskilling, skills paradoxes unsolved, leadership bottlenecks unaddressed, human capacity depleted. These aren’t problems requiring better execution of existing playbooks. They require fundamentally different approaches to how organisations develop and deploy human capability.
Whether that happens before windows close and costs compound—that’s the question Davos raised but couldn’t answer. The answer gets written in organisations where HR leaders choose between managing systems designed for yesterday or building capabilities required for tomorrow.
The clock, as they say, is ticking.