If you’re a software developer, tech recruiter, or technology leader, you’ve likely noticed something disturbing in the post-pandemic job market: internships and entry-level positions have declined rapidly.
In this article I wanted to analyze the causes, compare the numbers, but also the response strategies, from four major policy makers: US, EU, India and China.
When the Entry-Level Vanishes
Across the EU, junior tech positions fell 73% in 2025, as observed by Ravio. In the United States, junior hiring in AI exposed fields declined by 13% since late 2022, compared to stable or rising employment for older workers in the same roles (Stanford)*. In India, only 70,000 engineering freshers found employment from a pool of 1.5 million graduates in FY24, representing a twenty-year low. (Business Standard). From China there’s no official YOY figure for junior IT roles available, but multiple reports show that tech jobs for junior and generalist development positions are in decline, while 2024 saw a massive, 450,000 graduates output (CTOL Digital Solutions)**.
Reporting gaps:
*US junior tech hiring data from 2024–2025 is still being assembled. The most credible studies (such as Stanford) are working with data that has lags or covers only subsets of the market. **China’s government data is suppressed and estimates from independent analysis are limited.
This is happening while companies simultaneously post millions of unfilled senior IT vacancies and report acute talent shortages. This paradox could use an explanation, so let’s dive in.
A Structural Imbalance
This is not a supply-demand problem in the traditional sense. The supply exists (1.5 million Indian engineering graduates, 450,000 Chinese computer sciences graduates, 120,000+ U.S. CS degrees annually). The demand exists (1.1 million U.S. tech job postings, 11.5 million projected Indian data roles, 4 million Chinese AI specialist shortage by 2030). The imbalance comes from the mismatch: companies seek the output (experienced talent) without investing in the input (junior training). No market mechanism solves this without intervention, because every individual company benefits from hiring experienced talent trained elsewhere, but the industry collectively suffers when no one trains the next cohort.
Is This Reality-Grounded Market Evolution?
Up to a point, this transformation is inevitable and even necessary. AI tools have genuinely automated tasks that once defined junior roles: routine code generation, software testing, help desk triage, basic debugging. A senior engineer augmented by GitHub Copilot, Claude Code, and automated testing frameworks can now produce output previously requiring two to three junior developers. From this perspective, maintaining more junior positions than needed seems to be economically irrational.
The counterargument carries equal weight: every senior engineer was once a junior. If companies collectively abandon entry-level hiring while simultaneously competing to poach the same pool of senior talent, the industry is engineering its own collapse – a tragedy of the commons, Silicon Valley edition.
The rhetoric for AI displacement doesn’t “bypass” other professions. But, when Elon Musk calls medical schools “pointless”, hospitals don’t rush to gut their residency programs. Medicine shows that licensing requirements, regulatory frameworks, and even institutional inertia can (still) shield the training pipeline from disruption – a protection that the software industry, chronically unregulated, has never had.
Are Companies Following AI Hype Over Reality?
Indeed, some AI adoption has outpaced genuine capability. The 2023-2025 tech layoffs affected over 400,000 workers, with junior staff disproportionately cut under claims of AI replacement.
The compute spending pattern reinforces this interpretation. U.S. hyperscalers committed $350-450 billion in AI infrastructure capex for 2025, much of it speculative rather than tied to demonstrated ROI. Simultaneously, these companies eliminated tens of thousands of positions to “focus resources on AI.”
But this justification is rarely confirmed by lived reality. Only 4-5% of tech layoffs explicitly cite AI as the driver, far below restructuring or cost-cutting causes. The pattern that becomes visible is that AI hype provides cover for cost-cutting that would be politically difficult to justify on efficiency grounds alone. Junior positions are eliminated due to AI, whether AI actually replaces their output or not.
Is This Nearsighted Investor Pressure?
The most cynical interpretation is that tech positions (whether they’re junior or not) are being eliminated not because AI replaces them, but because cutting them improves short-term financial metrics.
Junior employees have the highest training-cost-to-immediate-productivity ratio. But the math doesn’t hold on in the long run. According to this article, choosing training over immediate productivity becomes economically advantageous in about 3 years’ time.
The problem is when executives operate on a 3-year overall timeframe. By the time the junior evolves to mid-level, current leadership will have moved on, vested their options, and collected their bonuses.
The Toxic Combination
The most likely explanation is that genuine structural change, AI hype, and short-term financial pressure operate as mutually reinforcing vectors. AI does genuinely automate some entry-level tasks (structural). This creates a plausible narrative for mass junior cuts (hype). Which improves quarterly financials and satisfies investor pressure (greed).
Each element provides cover for the others. Executives can claim “AI transformation” while pointing to legitimate productivity tools to justify cost cuts that boost earnings. Investors reward the stock performance, creating positive feedback. Meanwhile, the long-term consequence – talent pipeline depletion – remains invisible for years.
How Different Regions Are Responding
United States: Market-Driven with Minimal Intervention
The U.S. response is characterized by minimal government intervention and reliance on market forces. The U.S. Tech Force (1,000 federal positions) addresses less than 1% of the problem. Skills-based hiring reforms remove degree requirements but don’t create positions. The implicit assumption is that market corrections will eventually force companies to restart junior hiring when mid-level shortages materialize. This approach prioritizes “winning the AI race” over workforce stability, treating employment displacement as an acceptable cost of technological leadership. No policies incentivize private sector junior hiring, no apprenticeship mandates exist, and no tax mechanisms reward long-term talent development. The U.S. effectively bets that future labor shortages will self-correct through wage inflation.
European Union: Solid Framework, Soft Enforcement
The EU’s Union of Skills coordinates €150 billion in existing funds under a unified strategic framework. The European Alliance for Apprenticeships targets 700 pledges by 2030, representing 3 million training opportunities. Yet education remains a national competency, as the framework only offers guidance, not mandate. Member states implement at varying speeds with different priorities. Germany maintains strong apprenticeship traditions but struggles to extend them to tech. Southern European nations face higher youth unemployment but weaker vocational infrastructure. The approach is comprehensive in design but fragmented in execution. Without binding enforcement, the Union of Skills risks becoming more aspirational than transformative.
India: Massive Scale, Execution Gaps
India’s IndiaAI Mission and PM Internship Scheme (10 million placements) represent the most ambitious government response globally in absolute scale. Subsidized GPU access (~0.7USD/hour), 27 AI training labs in tier 2 and 3 cities, and employment-linked incentive schemes targeting millions demonstrate commitment.
Yet these measures address less than 1% of the 1.5 million annual engineering graduate pool. A government taskforce has proposed AI Transition Funds, wage insurance, and Human Impact Assessments before automation-driven cuts, but proposals are not yet implemented. The gap between ambition and execution is substantial. Plus, India trains talent for global markets rather than domestic absorption. Brain drain continues as top graduates emigrate for higher salaries abroad.
China: Interventionist Management, Uncertain Effectiveness
China represents the most interventionist approach. The AI+ initiative explicitly calls for steps to “reduce employment impacts”. Officials have proposed an “AI + Employment” framework with tax incentives, wage subsidies, and potential limits on job displacement. China previously regulated delivery platform algorithms to protect gig workers, demonstrating willingness to slow AI adoption for social stability.
Critically, China’s technical approach also differs, investing only 15-20% of U.S. AI infrastructure spending through efficiency (DeepSeek’s 95% cost reduction) rather than brute compute. If AI deployment costs drop dramatically, the economics of “junior + AI” combinations may become viable where U.S. costs make them prohibitive. However, China seems to be one step ahead in the data centers experiment, while youth unemployment remains elevated, graduate placement rates are low, and government data is partially suppressed, so there’s no way to tell how successful the interventions have been.
What Are the Key Takeaways?
The response from governments, companies, and educators has been fragmented and insufficient. Without deliberate, large-scale intervention, the tech industry risks losing a generation of talent and the pipeline that sustains its future. Here’s what you can do, today:
For Juniors: Supplement degrees with AI, cloud, or cybersecurity certifications. Seek apprenticeships, government programs, or roles in Global Capability Centers (India) or public sector tech (US/EU).
For Companies: Treat junior hiring as R&D. Firms that abandon training today will pay a premium for seniors tomorrow. Structured graduate programs and internal bootcamps are gaining.
For Policymakers: Create more binding incentives: tax credits, apprenticeship quotas, or public procurement conditions tied to junior hiring. The current responses are too small to match the crisis.
For Universities: Update curricula to integrate AI tools, cloud platforms, and cybersecurity. Partner with employers to ensure graduates enter the market with relevant skills.
The collapse of junior IT hiring is not inevitable. It’s a choice, and one with consequences that extend beyond any single company’s responsibility. It still takes a whole village to raise a child – and a whole industry to raise a junior engineer. And if we abdicate that responsibility entirely, we shouldn’t be surprised when the next generation of specialists is lost due to the broken worldview of the people who were left in the room – people like Sam Altman.