African Edtech Is Dying: We're Training for Jobs That Don't Exist

Edtech is training millions of Africans for skills the market doesn’t want.

This is my opinion. You might completely disagree. Don’t crucify me for it, share your thoughts in the comments instead. My thoughts may be very theoretical, but hear me out first.

TechCabal, in its SOTIA 2025 Year in Review, reported that the number of African edtech deals dropped 54.7% year-over-year in 2025. Investors who were writing checks for “democratising education” two years ago now ghost your emails. The standard explanation? Market correction, macro headwinds, the usual suspects.

I think the real reason is simpler and more uncomfortable: edtech is training millions of Africans for skills the market doesn’t want.

The Jobs Question

Over the last couple of years, I’ve sat through investment conversations with edtech founders doing impressive numbers - strong scale, solid completion rates, enthusiastic user testimonials. One question consistently breaks the camel’s back:

How many of your graduates actually got jobs?

The answer, after some thought and careful framing, is usually around 15%. 20% if you count gig work.

These aren’t bad founders. These aren’t poorly run companies. They’re teaching digital marketing, basic Python, some tech skill or the other; the same curriculum as dozens of other edtech companies. The exact same skills every graduate puts on their CV. The same capabilities that make it impossible for employers to differentiate candidates.

We’re training for abundance, not scarcity. And the market is telling us this, very loudly, through both the unit economics and the funding winter.

What the Data Actually Shows

New research from Wharton and Accenture just quantified what I’ve been watching play out across the ecosystem. They analysed millions of job postings and worker profiles to map which skills are oversupplied and which are in critical shortage. Though US-based, these findings should terrify every African edtech founder:

Massive oversupply: Communication, teamwork, problem-solving, leadership, basic digital literacy, all the “safe” skills that every training program teaches because students ask for them and they’re easy to standardise.

Critical undersupply: Domain-specific technical depth, regulatory expertise, scientific methods, analytical precision, execution-level operational skills, all the “hard” skills that are difficult to teach and impossible to scale.

Interestingly, the oversupplied skills correlate with lower wages. The undersupplied skills command annual premiums. Basically, we’re systematically training students in the lowest-value, most abundant capabilities in the market, then wondering why employers won’t hire them and why you have to spend 60% of revenue on student acquisition.

Now add AI to this equation. The same research tracked which skills are gaining or losing value as AI adoption accelerates. Routine content writing? Declining demand. Basic analysis? Declining. Entry-level coding? Declining. Those are the exact skills that fill most edtech curricula. We’re not just training for abundance, we’re training for obsolescence.

What is the way forward?

I’m not saying these aren’t important too, given the massive gap in education across the continent, but African edtech can’t scale this way. We’re building a model where the product has no differentiation value. The graduates compete with millions of others who have identical generic capabilities. The “placement rate” depends entirely on the student’s prior hustle, not your curriculum. Employers can’t tell your graduates apart from anyone else with “digital marketing skills” on their CV.

Here’s what I think edtech could be: Pick one specific, high-value domain where supply-demand gaps are widest. Build deep expertise in that one thing. Make your graduates invaluable in that narrow space rather than adequate in a broad one. Train for the judgment layer AI can’t replicate, teach the students how to use AI tools to handle execution, then make the contextual decisions AI can’t.

Stop measuring completion rates and NPS scores. Start measuring placement rates within 90 days, starting salaries compared to market average, and wage progression year over year. If you can’t move those metrics, your curriculum is wrong.

Accept that you won’t train a million students. The edtech companies solving real skill gaps train hundreds or low thousands per year. Deeply. In specific, high-value domains. They’re also profitable, with strong unit economics and defensible advantages. The ones chasing millions of students are training undifferentiated learners in commodity skills, burning cash on ads, and wondering why VCs don’t find them interesting.

Let’s Debate This

Can African edtech scale while training for surplus skills, or is that model fundamentally broken?

Is there a version of “broad digital skills training” that actually works, or am I missing something?

If you were building edtech today, starting from labour market data rather than student demand, what would you build?

I especially want to hear from edtech founders with battle scars, investors who see a path I’m not seeing, and employers who can explain what they actually need versus what they tell edtech companies.

Drop your take in the comments. Especially if you think I’m dead wrong.

The African tech ecosystem needs skilled workers. We won’t get there by training people for increasingly non-existent jobs.