The Talent Question: Universities and the Skills Reality

Numbers have a way of performing confidence. India produces approximately 2.5 million STEM graduates annually — more than any other country in the world. It has 1,168 universities. Its median age remains under 30, at a moment when Europe ages, Japan contracts and China confronts the downstream consequences of its one-child policy. Every competitiveness narrative about India begins with these figures, and they are all accurate. None of them is sufficient. Graduate volume without graduate readiness is a demographic paradox rather than a demographic dividend. A twenty-five-year-old with an engineering degree who cannot perform the tasks an employer needs is not an asset in the labour market; he is a person whose potential has been partially realised and whose trajectory will depend on whether the system around him corrects the gap or compounds it.

This article is about that gap: where it is widest, why it persists, what is being done to close it and whether the pace of correction is adequate to the ambitions of the India 2.0 industrial programme. The honest answer is that progress is real, the direction is correct, and the distance remaining is substantial. The semiconductor fabrication readiness figure of below one per cent is not a critique of individual students. It is a measurement of the distance between what a mass education system that expanded rapidly has delivered and what a frontier industrial programme requires. Closing that distance is the central human capital challenge of the next decade.

The AISHE 2022–23 data records 1,168 universities across India: 59 central, 486 state public, 186 state private and 128 deemed universities. Total student enrolment is 4.33 crore. The gross enrolment ratio of 28.4 per cent reflects genuine expansion from a much lower base, but it remains well short of the 50 per cent target set for 2035, and it is lower than China’s current enrolment ratio of approximately 60 per cent. The IIT system, which carries the greatest international recognition, produces roughly 17,500 undergraduate and approximately 25,000 total graduates including postgraduate and doctoral students annually. These are globally competitive institutions producing globally mobile talent. They are also a narrow apex in a system whose base is far less consistent.

The quality distribution is the critical variable. Approximately 35 per cent of eligible universities and 28 per cent of eligible colleges hold NAAC accreditation, with roughly 20 per cent across the accredited population achieving A++ to A grades. Faculty-to-student ratios average 1:26 against the AICTE and UGC recommended standard of 1:20. A ratio 30 per cent worse than the recommended norm, sustained across a system of this scale, compounds progressively: larger classes mean less individualised instruction, laboratories are shared across more students, and the applied practical experience that converts theoretical knowledge into industrial readiness is diluted. The expansion of higher education over the past two decades created access. It did not, with equal systematicity, create depth.

The aggregate employability figures conceal a more granular and more instructive pattern. The SHL National Employability Report 2024 and Wheebox India Skills Report 2024 both assess overall engineering graduate readiness at approximately 47 to 51 per cent. But that average is a composite of sectors with very different profiles. Banking, financial services and insurance reports 78 per cent employability among relevant graduates, according to CII and Wheebox data for 2025 — a sector that has historically invested heavily in graduate assessment and structured its hiring around standardised test performance. IT services and software reports 71.5 per cent. These are sectors that have spent twenty years building hiring pipelines, campus recruitment infrastructure and structured induction training that can absorb and convert graduate potential into productive output.

Move from established services into frontier industrial sectors and the picture changes sharply. NASSCOM’s Strategic Review 2026 places VLSI design readiness at approximately 25 per cent — among the 125,000 to 150,000 engineers who already work in the sector globally and represent 20 per cent of the world’s VLSI design workforce. That 25 per cent figure applies to people already in the field. For semiconductor fabrication — the physical manufacturing of chips in a cleanroom environment — HRGenix data for 2025 estimates readiness at below one per cent of engineering graduates, corresponding to roughly 6,000 individuals nationwide. India proposes to run multiple semiconductor fabrication facilities, beginning with the Tata–PSMC 28-nanometre fab at Dholera. The workforce it requires does not yet exist at the scale those facilities need. Building it is not a curriculum adjustment. It is a fundamental reorientation of what a significant portion of India’s technical education system produces.

The Chips to Startup programme, managed by MeitY with a sanctioned budget of ₹1,200 crore, is the most structurally ambitious curriculum intervention in India’s technical education in a generation. It works across 113 institutions — IITs, NITs and selected private universities — organised into three tiers. Twenty-five Tier 1 research institutions are mandated to develop System-on-Chip and complete VLSI systems with a target of five or more tape-outs per year. Thirty-five Tier 2 design institutions focus on application-specific integrated circuits and FPGA-based prototyping. Fifty-three Tier 3 skill hubs deliver large-scale Level 1 VLSI design training, each targeting a minimum of 1,000 industry-ready graduates annually. Across all tiers, participating ECE departments are shifting from a predominantly software and coding curriculum to a system with 40 per cent hardware and VLSI credit weighting.

The programme’s most practical features are the ones that address the specific bottlenecks that have historically prevented Indian universities from producing fabrication-relevant graduates. Centralised cloud-based access to Electronic Design Automation tools from Synopsys, Cadence and Siemens addresses the cost barrier that has historically put professional-grade design tools out of reach for most institutions. Mandatory coursework on RISC-V, the SHAKTI and VEGA open-source processors grounds students in architectures they will encounter in industry. The tape-out fund — which allows students to send their designs to the Semi-Conductor Laboratory in Mohali or to global foundries for physical fabrication — bridges the gap between design as a theoretical exercise and design as an industrial process. Against a five-year target of 85,000 to 1,00,000 specialised learners, 42,000 had registered across the C2S portal as of January 2026. That is 42 to 49 per cent of target in roughly the first half of the programme’s life. The trajectory is plausible. Whether 113 institutions is sufficient to meet the workforce requirements of multiple operational semiconductor fabrication facilities — which need not just designers but process engineers, materials scientists, cleanroom technicians and supply-chain specialists — is a question the programme does not yet answer.

India’s gross R&D expenditure stands at 0.64 per cent of GDP, according to DST and UNESCO data for 2022–23. This figure was examined in detail in the Industrial Depth article earlier in this series, where it was compared with China at 2.55 per cent, South Korea at 4.64 per cent and Taiwan at 3.49 per cent. The comparison is equally relevant here, because research intensity and human capital depth are not separable variables: doctoral ecosystems produce researchers, and researchers produce the industrial knowledge that advanced manufacturing requires. India produces 30,831 PhD graduates per year. That is a substantial absolute number. It translates to 255 researchers per million population. China has 2,400 per million. South Korea has 8,500.

The structure of India’s R&D investment compounds the quantity gap. Government institutions account for 60 per cent of gross R&D expenditure; the private sector contributes 40 per cent. In leading industrial economies, private sector R&D typically exceeds government R&D, driven by technology companies and manufacturing firms that invest in applied research directly connected to production. India’s private R&D share of 40 per cent is growing but reflects an economy where the corporate sector has historically treated R&D as a cost rather than a core competitive investment. WIPO ranks India sixth globally in resident patent filings, which indicates that the innovation infrastructure produces output at scale. The question is whether that output is generating the applied, industry-connected research depth that semiconductor fabrication, defence systems and rare earth processing require — or whether it remains concentrated in software patents and information technology applications where India’s existing strengths naturally cluster.

India’s labour market is structurally informal. The Periodic Labour Force Survey 2022–23 places the informal employment share at 90 per cent, defined to include the self-employed and those without written employment contracts or access to social security. This figure is not primarily a measurement of the unregistered economy; it is a description of the contractual and institutional character of the overwhelming majority of Indian employment. The formal sector, defined by EPFO payroll membership and comparable metrics, added 1.62 crore net workers in FY2024–25 — a genuine and meaningful expansion of organised employment. It is also roughly 16 million people in an economy where the working-age population grows by an estimated eight to ten million annually, suggesting that the pace of formalisation, while positive, is not yet rapidly transforming the structural character of the labour market.

The graduate unemployment rate of 13.4 per cent, from PLFS 2023–24, is the figure that most directly describes the tension at the heart of India’s human capital challenge. It is not the case that India lacks jobs; it is the case that a meaningful share of degree-level graduates cannot find formal employment commensurate with their qualifications. This is a structural mismatch problem, not primarily a demand problem. The RBI Bulletin’s employment elasticity data quantifies the broader dynamic: an aggregate elasticity of 0.2 means that a one per cent increase in economic output produces a 0.2 per cent increase in employment. Manufacturing elasticity is 0.19 — marginally below the aggregate. Services elasticity is 0.3. These are not low numbers in isolation; they describe an economy where growth generates formal employment, but at a pace that the combination of labour force growth and graduate output consistently outstrips.

India’s female labour force participation rate was 37 per cent in the PLFS 2023–24 survey — the highest recorded in recent data and an improvement from a low that had concerned economists and policymakers for over a decade. The number is worth examining in international context. China’s female LFPR is 61 per cent. Vietnam’s is 69 per cent. Bangladesh — a country with a significantly lower per-capita income than India — reports 38 per cent, roughly at parity. Brazil reports 54 per cent. The gap between India and China alone implies that a very large number of educated, capable women are either not participating in the formal labour market or are engaged in forms of work that PLFS methodology captures incompletely, particularly agricultural and domestic work.

The constraints are real and compound each other. Labour market informality reduces the availability of formal employment with predictable hours, safety guarantees and legal protections that make participation viable for women managing household responsibilities. Urban safety concerns affect commuting and late-shift work. The occupational segmentation of the Indian labour market means that women with particular qualifications often find that the jobs available to them cluster in roles that are underpaid relative to equivalent male-dominated roles. The productive consequence is direct: at 37 per cent LFPR against a potential comparator rate of 61 per cent, India is not deploying approximately a quarter of its educated female workforce in the formal economy. In the context of a human capital challenge defined by depth rather than volume, that is the single largest untapped productivity reserve in the system.

The brain drain from India’s elite technical institutions has been a persistent feature of the country’s human capital landscape for five decades. Stanford and NBER research published in 2024 estimates that 30 to 34 per cent of IIT graduates have historically emigrated, primarily to the United States, primarily in STEM fields. That historical rate appears to be declining: the same research places emigration for the 2021–23 cohorts at approximately 22 per cent — a meaningful reduction, though not a reversal. The drivers of the decline include the expansion of high-paying technical roles within India, the growth of the Indian start-up ecosystem, and the increasing difficulty of the US immigration pathway for Indian nationals given the extreme backlogs in the employment-based green card system.

The reverse flow is less reported but economically significant. NASSCOM and MEA data published in 2025 estimates 35,000 to 40,000 STEM professionals returning to India annually, bringing with them capital, networks, technical experience from frontier institutions and companies, and a willingness to take entrepreneurial risk that is partly a function of having operated in high-risk environments abroad. The Start-up India ecosystem, as of the DPIIT Status Report 2025, counts 1.17 lakh registered start-ups, 115 unicorns and 12 lakh direct jobs generated. The diaspora’s contribution to this ecosystem — as founders, as angel investors, as mentors and as early hires — is substantial and systematically undercounted in the human capital accounting of India 2.0.

The global context for India’s talent challenge is defined by two competitors who represent very different strategic profiles. Taiwan produces approximately 85,000 STEM graduates per year — roughly one-thirtieth of India’s volume. More than 60 per cent of those graduates are considered fab-ready for physical semiconductor manufacturing, a product of a university system that has been co-designed with TSMC, UMC and the broader foundry ecosystem over four decades. Taiwan has roughly 18 to 20 per cent of the global VLSI design workforce, comparable to India’s share, but its advantage is in fabrication depth rather than design volume. The English Proficiency EF EPI ranks Taiwan 30th globally. Researchers number approximately 7,500 per million. Entry-level STEM salaries of US$28,000 to US$32,000 — four to five times India’s comparable range of US$6,000 to US$8,000 — reflect both a higher cost base and a higher productivity floor.

Vietnam occupies the opposite end of the complexity spectrum. With approximately 100,000 STEM graduates annually and 5 to 8 per cent fab-readiness, it is winning the low-to-mid complexity manufacturing competition through labour cost parity with India, aggressive free trade agreements and the operational simplicity of assembly and OSAT operations that do not require deep technical specialisation. Vietnam ranks 44th on the Global Innovation Index 2025 against India’s 38th. Its researchers number roughly 700 per million. Its EF EPI rank of 58th reflects a lower English proficiency level than India’s 52nd. India’s competitive advantage relative to Vietnam is not cost — wages are comparable — but design depth. While Vietnam assembles electronics, India’s 125,000 to 150,000 VLSI engineers increasingly design the components inside them. The C2S programme is India’s mechanism to extend that design advantage into fabrication before Vietnam and other regional competitors develop the institutional infrastructure to close the gap.

The IMF’s 2024 analysis of generative AI and the future of work estimates that 60 per cent of India’s workforce is exposed to AI-driven change, with 24 per cent at high risk of displacement in their current roles. The distinction between exposure and displacement matters: exposure includes roles where AI will augment rather than replace, shifting the skill mix required rather than eliminating the position. The 24 per cent at high risk are concentrated in routine cognitive tasks — data entry, basic analysis, standardised document processing — that are precisely the entry-level white-collar roles that have historically absorbed a significant share of India’s degree-level graduates in the IT services, business process management and administrative sectors. India has 4.2 lakh AI and ML professionals, according to NASSCOM’s 2025 AI/ML Skill Talent Report — a workforce that is growing rapidly but remains small relative to the scale of the economy and the pace at which AI is being integrated into industrial and services processes.

India ranks 35th on the WEF Future of Jobs Readiness Index 2024, which reflects a composite assessment of education system adaptability, reskilling infrastructure, regulatory environment and technology adoption. The ranking is neither alarming nor reassuring: it places India in the upper-middle tier globally but well below the leading positions occupied by Northern European economies and East Asian technology hubs. The skill polarisation that automation typically produces — where demand grows at the high end for roles requiring creativity, systems thinking and technical depth, and at the lower end for roles requiring physical presence, while the middle compresses — interacts specifically badly with India’s graduate unemployment challenge. If the entry-level white-collar roles that have historically provided the stepping stone from degree to career are automated before the high-skill roles that India is trying to build become large enough to absorb graduate output, the structural mismatch problem worsens before it improves.