In January 2025, the World Economic Forum published its Future of Jobs report: 170 million jobs created by 2030 versus 92 million eliminated, a positive net balance of 78 million jobs worldwide. That same week, three French media outlets ran headlines about “the inevitable end of developers.” The gulf between the data and its media coverage has never been more dizzying.
This disconnect creates a practical problem for decision-makers. How do you make recruitment, training, and technology investment decisions when media noise drowns out reliable signals? What the numbers show—and what this article examines sector by sector—is both less catastrophic and more nuanced than the headlines.
TL;DR — AI is redistributing tech employment rather than destroying it. Data from the WEF, Gartner, and Coface points toward a positive net balance by 2030. But this redistribution affects people unevenly: unspecialized junior professionals struggle, while senior developers who master AI see their value soar. The key is understanding which jobs are shifting, which are emerging, and how to position yourself, rather than fearing automation.
The Media Narrative Versus Raw Data: Two Irreconcilable Stories
Headlines That Fuel Panic
In early 2026, a Coface-OEM study announcing that “five million jobs are threatened by AI in France by 2030” triggered a wave of alarmist articles. The general press repeated the figure without nuance. Franceinfo, Le Monde, BFM: each chose an “existential threat” angle.
What these articles systematically omit is that the Coface study does not say five million jobs will be eliminated. It says 16.3% of French employment will be “affected”—meaning certain tasks within those jobs will be automatable. The gap between “affected” and “eliminated” is enormous.
What the Authoritative Reports Actually Say
Three data sources carry authority on this topic. They tell a different story from the anxiety-inducing headlines.
| Source | Scope | Jobs created | Jobs displaced | Net balance | Horizon |
|---|---|---|---|---|---|
| WEF Future of Jobs 2025 | Global | 170 million | 92 million | +78 million | 2030 |
| Gartner (Oct. 2025) | Global IT | Not quantified | Neutral impact by 2026 | +500 million | 2036 |
| Coface-OEM | France | Not quantified | 3.8% of employment at risk today | Not specified | 2030 |
The WEF is unequivocal: software and application developers are among the fastest-growing occupations by volume. Gartner goes further: by 2036, AI will generate more than 500 million net new jobs supporting AI initiatives themselves.
Coface's Methodology Deserves to Be Read in Full
The Coface-OEM study stands out for its rigor: instead of reasoning about entire occupations, it breaks each into individual tasks and assigns an automation score to each. A senior developer spending 40% of their time writing “standard” code will be affected in that portion, but not in the remaining 60% devoted to architecture, code reviews, team management, and business discussions. The job is affected, not necessarily destroyed.
Tech Layoffs in 2024–2026: Anatomy of a Misunderstood Phenomenon
The Actual Job-Loss Figures
Tech layoffs are real and substantial: more than 150,000 jobs cut across 542 companies in 2024, more than 80,000 in 2025, and already 45,363 since the start of 2026. Intel announced 21,000 job cuts, 20% of its workforce. Amazon alone accounts for 30,000 cuts.
These figures are indisputable. Their cause is much less clear.
AI as a Convenient Scapegoat
Of the 80,000 tech layoffs in 2025, only 9,238—approximately 20%—are officially attributed to AI integration and automation. Analysts warn that most of these reductions involve “AI washing.” Companies invoke AI to justify restructuring that actually responds to excessive post-Covid hiring, normalized interest rates, and shareholder pressure on margins.
Tech giants hired aggressively between 2020 and 2022: Meta doubled its workforce in two years. Today's layoffs are primarily a correction of overhiring, rather than evidence of runaway automation.
The French Paradox: Hiring Falls While AI Vacancies Rise
In France, hiring for professional and managerial IT roles fell 18% in 2024 according to APEC, with junior recruitment down 19%. Yet France simultaneously published more than 166,000 AI-related job postings in 2024, leading Europe ahead of Germany's 147,000 and the United Kingdom's 125,000, according to PwC's barometer.
The explanation for this paradox is that the market is being reshaped rather than contracting. Traditional development roles decline while hybrid development-plus-AI roles surge. According to PwC, “AI-augmented” jobs grew 252% between 2019 and 2024.
Sector by Sector: Who Gains, Who Loses, Who Changes
Software Development: Augmentation Rather than Substitution
Software development is where panic is strongest—and the data most reassuring. GitHub Copilot, the most widely used coding assistant, now has 20 million users and is used by 90% of Fortune 100 companies.
Productivity figures are substantial: developers equipped with Copilot finish tasks 55.8% faster, and pull request processing time has fallen from 9.6 days to 2.4 days, a 75% reduction. AI now writes 46% of code, reaching 61% among Java developers.
But these productivity gains do not translate into job cuts. Why? Because demand for software is virtually unlimited. Every productivity gain helps tackle a backlog that is overflowing in most companies. Developers are delivering more, rather than coding less.
Administration and Support Functions: The Real Point of Impact
The Coface-OEM study identifies the most exposed functions: administration, accounting, legal work, and support. These occupations involve substantial document processing and structured repetitive tasks—precisely what generative AI automates best.
Paradoxically, the highest income deciles are most exposed: 22.1% of tasks performed by France's top-paid 10% are automatable, compared with only 6.5% for the lowest-paid 10%. AI threatens the top of the ladder rather than low-skilled employment.
Finance and Consulting: A Profound Reshaping
In finance, AI automates data analysis, report writing, and regulatory compliance. At the same time, it creates demand for specialized fintech engineers, AI scoring-system architects, and algorithmic compliance specialists. The WEF ranks fintech engineers among the fastest-growing occupations.
Consulting is undergoing a similar transformation: data gathering and synthesis tasks formerly assigned to junior consultants are gradually being automated. Firms now recruit people able to manage AI analysis tools instead of performing the work manually.
Manufacturing and Construction: Less Impact than Predicted
Contrary to alarmist predictions, manual and industrial occupations remain largely untouched. The WEF even forecasts that the largest absolute employment growth will occur in frontline occupations: agricultural workers, delivery drivers, construction workers, and salespeople. AI transforms design and planning processes, but the last mile remains human.
Emerging Roles: The 2026 Opportunity Map
The Hybrid Developer–AI Profile: The Most Sought-After Skill Set
The most coveted profile in 2026 is a developer capable of designing systems that integrate AI models: autonomous agents, RAG pipelines, and LLM orchestration. This is no longer a niche: market data shows that 85% of companies want to incorporate generative AI into their projects.
This hybrid profile combines software architecture skills with language model expertise. Salaries reflect its scarcity:
| Emerging role | Gross annual salary (France) | Job posting growth 2025–2026 |
|---|---|---|
| Lead ML Engineer | €70,000–€105,000 | +30% |
| AI Architect | €58,000–€90,000 | +35% |
| Prompt Engineer | €55,000–€80,000 | +32% |
| AI Data Engineer | €45,000–€65,000 | +28% |
| Agentic AI Developer | €60,000–€95,000 | +40% |
The AI Salary Premium: A Measurable Advantage
PwC's AI Jobs Barometer quantifies an unmistakable salary advantage: employees with AI skills earn an average of 56% more than peers without them. The gap has widened considerably compared with previous years, signaling growing pressure on this talent pool.
Roles That Did Not Exist Three Years Ago
The AI ecosystem is creating entirely new functions:
- AI Ethics Officer: responsible for AI systems' ethical and regulatory compliance, driven by the European AI Act
- AI Trainer: specializes in fine-tuning and evaluating models for business use cases
- Agent Orchestrator: designs multi-agent architectures and autonomous AI agent workflows
- AI Product Manager: manages products incorporating AI components, at the intersection of business and technology
Gartner predicts that 40% of enterprise applications will include specialized AI agents by the end of 2026, compared with less than 5% in 2025. This surge creates structural demand for developers able to design, deploy, and maintain those agents.
The Real Divide: Juniors Versus Seniors, Generalists Versus Specialists
The Squeeze on Junior Developers
US data is stark: the number of software developers aged 22–25 in the workforce has fallen 20% since late 2022. The explanation is that tasks traditionally assigned to juniors—CRUD code, simple unit tests, documentation—are now handled by AI tools, rather than those jobs simply disappearing.
A survey conducted before GDC 2025 found that 11% of developers had been laid off during the previous 12 months, four percentage points more than a year earlier. Generalists without a clear specialization were most affected.
Augmented Seniors: More Productive, More Essential
Conversely, senior developers who master AI tools see their productivity and market value soar. A senior using Copilot, Cursor, or Claude Code reduces development time on coding tasks by more than 50%, freeing time for architecture, code review, and discussions with business teams.
Gartner is explicit: by 2027, 80% of software engineers will need to upskill. The message is clear: this is the end of the developer who refuses to adapt, rather than the end of the developer.
What This Means for Recruiters and CIOs
Practical guide: Five signals for assessing a developer in 2026
- Daily AI tool use: 90% of developers report using AI in 2025. A candidate who does not use it is behind.
- Ability to evaluate generated code: AI writes 46% of code, but 29.1% of generated Python code contains vulnerabilities. Human judgment remains critical.
- LLM integration experience: API calls, RAG, context management—these distinguish an augmented developer from a conventional one.
- Architectural vision: AI accelerates execution rather than design. A developer who thinks about architecture before implementation is more valuable than ever.
- Documented adaptability: a career showing stack changes, skill development, and active technical curiosity.
The Quality Question: AI Produces More, but Does It Produce Better Work?
The Paradox of Apparent Productivity
GitHub Copilot's productivity figures impress: 55% gains and a 75% shorter pull request cycle. But a 2025 GitClear study reveals a downside: the volume of cloned code—duplicate or nearly identical code—has quadrupled since AI assistants became widely adopted.
Copilot suggestion acceptance rates hover around 30%, and 88% of accepted code remains in the final version. That means 70% of suggestions are rejected, while the accepted 30% still require rigorous human validation.
Security: The Achilles' Heel of Generated Code
Security data deserves attention: 29.1% of AI-generated Python code contains vulnerabilities, with a 6.4% rate of leaked secrets such as API keys and credentials. This supports keeping competent senior developers in the validation loop, rather than rejecting AI.
AI does not eliminate the need for human expertise. It shifts that need toward review, architecture, and security, and away from mechanical code writing. The developer of 2026 is more of an architect and auditor than an implementer.
What McKinsey Observes in Practice
According to McKinsey, 32% of companies expect AI to reduce headcount by at least 3% over the coming year. Simultaneously, 92% plan to increase AI investment over the next three years, with the highest performers allocating more than 20% of their digital budget to AI. Massive AI investment creates equally substantial demand for deployment expertise.
How to Position Yourself: A Practical Guide for Decision-Makers and Tech Teams
For Leaders of SMEs and Intermediate-Sized Businesses
The priority is to redirect tech teams rather than cut them. Companies using AI to replace developers are taking a risky bet: they lose the human expertise needed to supervise, correct, and evolve the AI systems themselves.
The effective approach is to invest in the existing team's AI skills. According to Gartner, organizations combining “AI readiness” and “human readiness” capture and sustain value more durably than those betting solely on automation.
For CIOs and CTOs
Syntec Numérique's Numeric'Emploi program offers one route: 80% of people retrained for digital careers find employment. At national scale, this model could train 60,000 jobseekers a year for tech occupations.
For CIOs, the recommendation has two parts:
- Audit the current team's AI skills: how many developers use AI assistants daily?
- Plan upskilling over 12–18 months: Gartner predicts that 80% of engineers will need to upskill by 2027.
For Individual Developers
Practical guide: An AI upskilling plan for developers
Timeframe Action Expected impact 0–3 months Master an AI assistant: Copilot, Cursor, Claude Code Productivity +50% 3–6 months Learn LLM API integration: OpenAI, Anthropic, Mistral Access to hybrid roles 6–12 months Specialize in AI agents, RAG, or fine-tuning Salary premium +56% 12–18 months Develop AI architecture expertise Access to lead/architect roles
What Will Really Happen by 2030
The Scenario Major Studies Converge On
Reports from Gartner, the WEF, McKinsey, and PwC converge on a central scenario:
- Short term, 2026–2027: neutral impact on total tech employment, but a major shift in demanded profiles. Routine tasks are automated and hybrid roles surge.
- Medium term, 2028–2030: a positive net balance. AI creates more jobs than it displaces, primarily in designing, deploying, and maintaining AI systems.
- Long term, 2030–2036: according to Gartner, more than 500 million net new jobs linked to AI initiatives. The developer of 2030 will resemble the developer of 2020 as little as the latter resembled a COBOL programmer of the 1990s.
The Three Skills That Will Make the Difference
Beyond changing tools and frameworks, three structural skills separate resilient professionals from vulnerable ones:
- Architectural abstraction: designing systems rather than just code. AI speeds execution rather than providing the overall vision.
- Technical judgment: assessing generated code for suitability, security, and maintainability. With vulnerabilities in 29% of AI code, this skill is vital.
- Business communication: translating business needs into technical specifications and vice versa. AI cannot yet understand a project's political and organizational context.
FAQ
Will AI Replace Developers by 2030? No. The World Economic Forum ranks software developers among the fastest-growing occupations. Gartner forecasts a positive net balance of 500 million AI-related jobs by 2036. Developers who master AI will be more in demand, not less.
How Many Tech Jobs Have Actually Been Eliminated Because of AI? Of the 80,000 tech layoffs in 2025, only 20% are attributed to AI. Most reflect post-Covid corrections and conventional restructuring. Analysts use “AI washing” to describe the misuse of AI as a justification.
Which Tech Professionals Are Most Threatened by AI Automation? Junior generalist developers without a clear specialization are most exposed. In the United States, the number of developers aged 22–25 has fallen 20% since 2022. Senior and specialized professionals, conversely, are seeing their value rise.
What Salary Impact Do AI Skills Have for a Developer? According to PwC's barometer, tech professionals proficient in AI earn an average of 56% more than peers without AI skills. In France, a Lead ML Engineer earns €70,000–€105,000 gross annually.
Does Generative AI Produce Code Good Enough to Replace a Developer? No. Although 88% of code accepted through Copilot remains in production, 29.1% of generated Python code contains security vulnerabilities. Human judgment remains essential for review, architecture, and security.
How Should a CIO Adapt Tech Recruitment Strategy to AI? Gartner recommends planning AI upskilling for 80% of engineering teams by 2027. The priority is training the existing team on AI tools rather than replacing positions. Hybrid developer–AI profiles are the most strategic recruits.
AI Coder Squad: Senior Developers Who Master AI Instead of Being Overtaken by It
The data in this article converges: the future belongs to teams combining senior technical expertise with mastery of AI tools. This is exactly the model AI Coder Squad applies daily: developers with more than ten years of experience, augmented by AI to deliver faster without sacrificing rigor.
AI Coder Squad designs custom applications and AI agents for companies that want to move quickly without sacrificing quality—with senior developers and an AI-powered approach.
→ Start your project and discover how AI Coder Squad can accelerate your next delivery.