What South Africa's Tech Sector Is Calling Efficiency, History Will Call a Mistake

The entry-level work AI is automating was never just about output. It was about building the people who would run the teams.

Video summary: https://youtu.be/YvEdSaJ5TFY

Sign up for my Substack daily AI newsletter here.

See my AI Training course portfolio for corporate Business Leaders here.

Article link: https://open.substack.com/pub/johanosteyn/p/what-south-africas-tech-sector-is?r=73gqa&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/

Every senior developer in South Africa was once the junior developer their employer was not entirely sure was worth the investment. They wrote the routine code. They did the basic testing. They administered the systems. They made the beginner mistakes, learned from them, and over several years of unglamorous, repetitive, foundational work, became the people their organisations could not function without. That process — slow, expensive, imperfect, and entirely necessary — is being systematically eliminated across the global tech sector right now. And the companies doing the eliminating are calling it efficiency.

CONTEXT AND BACKGROUND

The evidence of what is happening to junior developers in South Africa is now accumulating from multiple independent sources, and it tells a consistent and troubling story. Writing in TechCentral on 21 April 2026, journalist Nkosinathi Ndlovu reported on OfferZen’s State of South Africa’s Developer Nation 2026 Salary and Benefits Report, based on a survey of more than 2,200 software, data and tech professionals. The headline finding was stark: 62 % of junior developers feel underpaid — the highest rate of any seniority band — and average entry-level fintech salaries fell from R37,748 per month in 2025 to R27,777 in 2026, a 26% collapse in a single year. The full OfferZen report confirms that 97% of South African tech teams now use AI, which has in the report’s own words, almost eliminated the need for very junior developers to perform mundane tasks. One anonymous survey respondent captured the structural shift plainly: AI has removed the need for juniors, but seniors are now in higher demand to oversee AI.

The OfferZen data is not an outlier. The Harambee Youth Employment Accelerator, in collaboration with The Bridgespan Group, published research in February 2026, finding that South Africa’s AI market grew 31 per cent year-on-year between 2023 and 2024 — and that over 40% of tasks in IT-enabled services sectors are technically automatable, with changes disproportionately affecting junior and entry-level positions. The Harambee research offers a more optimistic framing: that AI is reshaping rather than eliminating jobs, creating hybrid roles that require human judgment and emotional intelligence. That framing is important. But it does not resolve the central tension: hybrid roles that require judgment developed over years of experience are not entry points for young people who have not yet had the chance to develop that experience. They are the destination. The pathway to them is precisely what is disappearing.

INSIGHT AND ANALYSIS

The problem with describing the elimination of junior work as efficiency is that it misunderstands what the work was actually for. The routine tasks that AI is now absorbing — the basic testing, the repetitive coding, the system administration — were never primarily about output. They were about formation. They were the mechanism through which junior developers built the technical intuition, the debugging instincts, the architectural awareness, and the professional judgment that cannot be acquired any other way. You cannot read your way to understanding why a system failed at 2am. That knowledge is built through the doing — slowly, over years, through exactly the kind of work that tech teams are now automating away.

Calwyn Baldwin, automation team lead at Obsidian Systems, identified this dynamic precisely in a February 2026 interview with TechCentral on the impact of AI-assisted coding on South Africa's developer pipeline. The most significant impact of AI coding tools, he argued, is that junior programmers will not have the opportunity to develop the strong, high-level architectural skills that their senior counterparts possess — in part because AI removes the cognitive load that was previously the training ground for those skills. If developers simply instruct AI, test whether the output works, and never engage with the underlying code, the long-term consequence is a generation of developers who do not know the codebase and cannot explain what it does. The antidote, Baldwin argued, is ensuring that juniors and seniors engage more frequently so that skills are transferred — a condition that becomes impossible to meet when companies have stopped hiring juniors altogether.

The global data support this prediction precisely. Research from BEON.tech on the 2026 software talent market identifies the same polarisation now visible in South Africa — an oversupply of junior and generalist developers competing for fewer entry-level positions, alongside a genuine and growing scarcity of senior engineers who can operate complex systems in production. Its conclusion is unambiguous: these senior skills only develop through years of real production experience. No bootcamp or two-year career window gets you there. The senior talent scarcity that 48% of South African tech leaders are already reporting is not despite the decision to stop hiring juniors. It is, in part, a direct consequence of it.

History confirms this pattern with uncomfortable precision. As analysis of tech hiring cycles documents, after the financial crisis of 2008, hiring freezes created a gap in the experience curve. By 2012, companies struggled to find engineers with three to five years of experience because few had been hired as juniors during the recession. The industry eventually recovered — but only after years of painful competition for a cohort of mid-level talent that had simply not been produced. The same dynamic is forming today across the global tech sector, with AI providing the rationale rather than economic pressure. The rationale is different. The structural consequence is identical.

IMPLICATIONS

For South African tech leaders, the practical question that OfferZen’s Jason Tame posed in the TechCentral report deserves a direct answer: who trains the next generation of seniors if nobody is willing to hire juniors now? His own framing of what AI cannot replace is precise — product taste, architectural judgment, and the critical thinking to know when AI got it wrong. Those capabilities are not taught in classrooms or acquired through bootcamps. They are built through the foundational years of a developer’s career, in exactly the work environments that South African companies are currently dismantling.

The Harambee and Bridgespan research offers a path forward that deserves serious engagement from industry and government. It identifies two possible AI scenarios for South Africa: one in which AI augments human work, supporting young people into hybrid roles with higher value and clearer career progression, and one in which AI is used purely to cut costs by taking humans out of the equation entirely. The first scenario requires what Harambee describes as intentional, coordinated action: targeted investment in AI policy frameworks, digital infrastructure, inclusive reskilling, and public-private collaboration. Without that intentionality, the second scenario — the one in which entry-level roles are automated away and higher-value roles become increasingly inaccessible — is simply the path of least resistance.

For South Africa specifically, the stakes of choosing the wrong path extend far beyond the tech sector. A 26% collapse in entry-level fintech salaries in a single year, in a country with youth unemployment among the highest in the world, is not a market correction. It is the erosion of one of the few credible pathways through which young South Africans from under-resourced backgrounds could build technical careers without access to capital, connections, or postgraduate credentials. The tech sector’s decision to skip the junior rung is not a neutral efficiency calculation. In South Africa, it is a decision about who gets access to economic opportunity — and it is being made without that consequence ever appearing in the efficiency metrics that justify it.

CLOSING TAKEAWAY

The history of technology hiring is littered with decisions that looked rational in the quarter they were made and catastrophic in the decade that followed. The systematic elimination of junior developer roles across South Africa’s tech sector is the current iteration of that pattern — a short-term efficiency gain that is simultaneously destroying the formation process through which the next generation of senior talent would have been built. The routine work AI is absorbing was never just about getting things done. It was about building the people who would eventually run the teams — and the products, and the organisations, and the economy that depends on them. The senior developers of 2035 are the junior developers of 2026. And right now, nobody is hiring them.

Author Bio: Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net